<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by bellman on Medium]]></title>
        <description><![CDATA[Stories by bellman on Medium]]></description>
        <link>https://medium.com/@bellman-h?source=rss-99e8e6896144------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*M2XjlmuSeann2tqTIfxlvQ.jpeg</url>
            <title>Stories by bellman on Medium</title>
            <link>https://medium.com/@bellman-h?source=rss-99e8e6896144------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 14 Nov 2026 12:14:53 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@bellman-h/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Managing Real-Life Portfolio based on Multi-Agent LLMs (real trading results of my main trading…]]></title>
            <link>https://bellman-h.medium.com/managing-real-life-portfolio-based-on-multi-agent-llms-real-trading-results-of-my-main-trading-3a70cd631fe8?source=rss-99e8e6896144------2</link>
            <guid isPermaLink="false">https://medium.com/p/3a70cd631fe8</guid>
            <dc:creator><![CDATA[bellman]]></dc:creator>
            <pubDate>Wed, 10 Dec 2025 09:43:33 GMT</pubDate>
            <atom:updated>2025-12-10T09:43:33.436Z</atom:updated>
            <content:encoded><![CDATA[<h3>Managing Real-Life Portfolio based on Multi-Agent LLMs (real trading results of my main trading book)</h3><h3>개요</h3><p>Mid-Frequency Alpha는 점점 시장 경쟁이 심해지며 sharpe한 개별 알파보다도, 다양한 표현식을 지속적으로 리서치 하여 잘 앙상블하는 것이 필승 전략이 되었다. 따라서 관련 하우스, 특히 대중에게도 잘 알려져 있는 WorldQuant 같은 하우스에서는 리서쳐를 규모있게, 여러 대륙에 뿌려 보다 다양한 팩터들을 많이 찾으려고 노력하고 있다.</p><p>사람들이 생각하는 것과는 약간 다르게, 시장의 holy grail 따위를 찾는 것이 아니라 시장의 무수히 많은 스펙트럼 중에, 가장 많은 면적을 커버하는게 보통 이쪽 싸움이라고 생각하면 조금 이해가 쉬울 것 같다.</p><p>이러한 변화는 곧, 트레이딩 알파들이 더 빠르게 saturation되고 시장은 더욱 noisy (혹은 efficient) 해진다고도 볼 수 있다. 그렇기에 높은 샤프비율을 달성하기 위해서 투여해야 하는 인프라와 자본의 레벨이 점점 비대해지고, 결국 큰 진입장벽이 생겨 이것이 오늘날까지도 미국의 거대 퀀트 헤지 펀드들의 강력한 해자 중 하나가 되고 있다.</p><p>그렇다면 그정도의 인프라와 자본을 투여할 능력이 없는 개인투자자로서는 그냥 접거나 그들 아래로 조인하는 수밖에 없는가? 라는 의문이 들지 않을 수 없다.</p><h3><strong>크립토 시장에서의 트레이딩 경험</strong></h3><p>필자는 크립토 시장에서 다양한 알고리즘 트레이딩 전략을 바탕으로 수년간 트레이딩을 해왔다. 크립토 ETP가 많이 활성화되기전인 22년도까지만 해도, 아주 간단한 시스템 트레이딩 전략들의 샤프비율도 아주 좋았다. 그러나 기관 투자자들의 진입이 가속화되고, 코인 시장 자체가 다각화됨에 따라 시장 구조가 복잡해지며 그런 전략들이 더이상 안 먹히기 시작했다.</p><p>그래서 그 이후로 찾은 전략들은 흔히 Pairs Trading이라고 분류되는 종류의, 상관관계 기반 Stat-Arb 전략이었다. 물론 단일 전략으로 운용한 것은 아니고, 이전 시스템이든 이러한 Stat-Arb 시스템이든 5–10개의 전략을 리서치 하여 분산해서 운용 하였다.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/645/0*-IzcsC0LYiO_OWVq.jpg" /><figcaption>Pairs trading</figcaption></figure><p>성과 자체는 나쁘지 않았으나, 급격한 매크로 상황 및 크립토 시장 상황 변동에 예상 범위를 크게 벗어나는 리스크를 맞아버려서 멘탈이 나가기도 했다. 물론 필자 뿐만 아니라, 수명-수십 명 규모의 운용 팀들도 손실을 많이 보는 시장이기는 했다.</p><h3>Alphas In Crypto</h3><p>크립토에서도 흔히 이야기하는 Mid-Frequency Stat-Arb를 실행하는 사람들이 꽤 있었지만, 한가지 느낀 점은 그들이 기관임에도 불구하고 (물론 제도권 자산에서 운용하는 팀들과 자금력 / 팀사이즈 차이가 있지만) 아주 큰 규모의 알파 풀을 운용하고 있지 않다는 것이었다. 물론 내가 혼자 하는 것보다야 알파 구성이 체계적이겠으나, 보통 많아야 20–30개 정도의 알파풀을 관리하는 것으로 보였다.</p><p>제도권에서는 전체 알파풀 단위는 수만개, 수십만개 단위로 가는 경우가 허다하고 실제 운용하는 전략도 수백-수천개 정도가 될 수 있기에 이런 부분에서 크립토 시장 자체는 아직 Mid-Frequency Stat-Arb 전략들의 룸이 많이 남아 있다고 느꼈다.</p><p>그리고 크립토 시장 자체가 원래는 Beta (~비트코인 수익률) 이 대부분을 설명하게 되는 시장이었으나 여러가지 섹터로 다변화되며 점차 시장 종목들 간의 correlation이 낮아지기 시작한 것이다. 이것도 위의 예상을 뒷받침하는 부분이다.</p><h3>Objective: 1000+ Uncorrelated. Alphas in Crypto</h3><p>Correlation이 낮은 1천개의 알파를 만들면, 적어도 크립토 시장에서는 무조건 괜찮은 수익률을 낼 수 있다고 생각했고, 1천개를 만드는 것 자체의 난이도도 제도권 시장보다야 훨씬 낮은 것으로 판단했다.</p><p>하지만 어느 세월에 내가 1천개를 다 깎는가?</p><h3>Solution: Automating Alpha Research</h3><p>답은 알파 리서치 자동화 뿐이었다.</p><p>필자는 이전부터 Genetic Algorithm, Bayesian Optimization, Combinatorial RL, Customized Transformers 등 Symbolic Regression에 기반한 다양한 Alpha Research Automation을 시도해 보았다. 성과가 썩 나쁘진 않았지만, 결국 Context 없이 수식을 끼워 맞추는 것에 불과하기에, Overfitting도 심했고,</p><p>가장 문제가 되었던 것은 알파 리서치를 자동화하겠다는 것은 “다각화&quot; 를 위한 것인데, 커스터마이징한 Symbolic Transformer의 경우 약간 이야기가 달랐지만 위의 알고리즘들의 경우 최적해가 한 점에 있다고 가정하는 수렴적인 알고리즘이기 때문에 구조적인 한계가 명확했다.</p><p>결국 제대로 튜닝했을 때는 어떤 한 점으로 수렴하게 되는데, 나는 무수히 많은 차원 속에서 다 정답이 될 수 있는 1000개의 점을 찍어야 이기는 게임을 하고 있었기 때문이다.</p><p>예전에도 LLM을 활용해 보려는 시도는 했으나, 아직 이런 큰 컨텍스트를 감당하기에는 2023년도까지만 해도 제대로 된 모델이 없어서 접어두고 Traditonal한 방법들을 어찌저찌 끌고가며 주로 운용 했다.</p><h3>LLM-Based Alpha Research</h3><p>그런데, LLM들이 기하급수적으로 발전하면서 약간 상황이 달라져버렸다. 이제 아주 복잡한 작업도 End-to-End로, 여러 에이전트가 협업하면서 처리해버리는 기이한 시대가 시작되어버린 것이었다.</p><p>돈을 꽤나 잃고 현타가 온 2025년 4월, 두달만에 프로토타입을 만들어 첫 실거래를 5/30일에 걸게 된다.</p><p>LLM 기반으로 알파 리서치를 어떻게 하는지는, 다 이야기하자면 너무나 기니까 내가 느낀 문제 — 해결을 쭉 적어주고 마지막에 다이어그램과 슬라이드를 첨부하겠다.</p><p>0. 백테스트 코드를 짜기가 싫어서, 전략을 말해주면 백테스트 코드를 짜주는 걸 만들었다 (여기까지는 사실 상용 코딩툴을 써도 똑같다)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/586/1*P_qV4f6htlsqSJQzpRlTfw.png" /><figcaption>실제 시스템 데모 (Backtest Agent)</figcaption></figure><ol><li><strong>근데 백테스트를 하기 전에 아이디어가 맞는지 이리저리 리서치를 해 봐야 하는데 리서치도 귀찮다 -&gt; 리서치 아이디어를 넣으면 Step-by-Step으로 리서치를 해주는 친구를 만들었다. (사실 여기가 제일 튜닝이랑 기괴한 LLM 술기가 많이 들어간 부분이다. 지금도 계속 뭔가 업데이트를 하고 있다.)</strong></li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Kg6LeRh57BQyWvN8WskUIA.png" /><figcaption>Step-By-Step으로 코드를 짜고, 리서치 결과를 내는 에이전트</figcaption></figure><p><strong>2. 리서치 — 백테스트 — 리서치 를 하는 이터레이션을 돌기가 너무 힘들고 피곤하다. -&gt; 그래서 리서치랑 백테스트 에이전트를 부리는 팀장 에이전트 같은 것을 만들었다.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/729/1*kPyU8v2keHXHi1DXXVMnTQ.png" /><figcaption>친절히 자기소개를 하는 Planner Agent</figcaption></figure><p><strong>3. 이제 팀장한테 던져주면 오토파일럿이 되니까 좋긴 한데, 이제 제안할 가설이 떨어졌다 -&gt; 자체적으로 과거 리서치와 웹을 분석하고 개선점 / 안해본 가설만 도출하는 인사이트 에이전트를 개발해서 조수로 붙여주었다.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/503/1*QeyEPwEpUHiUVPrWiw-4Ug.png" /><figcaption>대략 Claude Code의 그것과 비슷한…</figcaption></figure><p><strong>4. 말이 쉽지 만들고 보면 뭔가 계속 꼬이기 시작하는데 이거 프롬프트 튜닝도 너무 소모적이다 -&gt; 서로서로 피드백을 줘서 메모리를 발전시키는 메모리 시스템과, 피드백 탐색 + 회고하는 일종의 HR 에이전트를 만들었다.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/716/1*2SS0LkY7PA_Yh-qKMBeh4w.png" /><figcaption>자기회고도 한다. 절대 그런 Bias를 준 적이 없는데, 간단한 전략이 학술적인 전략보다 좋다는 내가 1년쯤 깨지고 깨달은 인사이트를 스스로 배워와서 엄청나게 놀랐다…</figcaption></figure><h3>Full Picture</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*YH5Q415curq-uRF2_5x-OA.png" /></figure><p>쓰고 보면 결국 하나의 회사를 만든 셈이다. 사람들이 보통 AI 에이전트를 하나를 만들고 A-Z를 하기를 바라는데, R&amp;R 구분 없이 일을 다 시켜버리면 아무리 유능한 사람도 퍼져버리는데 AI는 안 그러겠나?</p><p>결국 R&amp;R을 더욱 세밀하게 정의하고, 서로간의 소통을 아주 잘 정리해줘야 한다는 점에서 인간 팀을 관리하는 것과 에이전트 팀을 관리하는 것이 크게 다르지 않은 것 같다.</p><p>적어도 고소 / 감정 / 체력 싸움은 안 해도 되니까 어찌 보면 더 편할지도 모른다.</p><p>이런 이야기 하면, 멀티 에이전트 시스템을 잘못 만들면 대부분 일이 산으로 간다는 크리틱이 많은데, 나는 일이 산으로 가지 않는 실무 팀을 본 일이 손에 꼽는 것 같다. 사람들도 대부분 헤매고 산으로 가고 그러다가 되는거라서… AI에게 조금만 관대하게 대하자(?)</p><h4>Insight: Autopilot VS HITL(Human-In-The-Loop)</h4><p>요즘 나오는 서비스를 보면 (한달 단위로 뭔가 메타가 바뀌어서 2025/11 정도의 기준이다.) 거의 HITL, 어떻게 하면 인간과 AI가 협동을 잘 할까에 대한 고민을 가지고 만들어진 것 같다.</p><p>이게 안전하고, 구현도 쉽고, 비용도 쌀 거다. 그러나, 폭발적인 생산성은 내기 어렵다. 10인분까지는 해도, 1000인분을 하는 것은 아무리 좋은 툴을 들고와도 안 되는거다.</p><p>결국, HITL 서비스 유저는 산업혁명 시대 재봉틀을 들고 일하는 소년에 가깝고, 서비스 오너가 공장주가 되는 느낌이다. (물론 코스트 제대로 계산 못해서 망하는 공장주는 예나 지금이나 있다.) 그래서, 나같이 뭔가 사람들의 일상적인 서비스에 대한 공감능력이 떨어지는 사람은 Autopilot의 접근 방식을 따라야 할 것 같다.</p><p>AI가 백그라운드에서 계속 돌고, 나는 매니저로써 이상치에 대한 QC나 피드백만 하는 구조. 결재문서에 싸인만 하는 구조가 되어야 한다. 실무에 대한 중간 개입 없이 말이다. 물론 개발에 애로사항이 더 많아질 것이고, 토큰 사용량도 늘어날 것이다.</p><p>하지만 사람을 쓰는 (마케팅해서 고객을 획득하든 고용을 하든) 비용보다는 싸고 더 싸질 거다. 감정적인 소통 비용, 법무 / 세무 비용, 오피스 비용 등을 생각하면 더더욱 싸다. 지금도 말도 안되게 싸다.</p><p>그래서, 나는 소년공보다는 아무래도 공장주가 되고 싶으니까 Autopilot 기반의 뭔가를 앞으로도 계속 시도할 것이다.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jdARCAY5UCtjYzWMvSpXMA.png" /><figcaption>GPT로 만들기는 했으니 올리는데 별로 그렇게 마음에는 안 드는 일러스트</figcaption></figure><h3>Results</h3><p>결론은, 그렇게 초보 공장주가 되어 LLM 토큰 비용으로만 수천만원을 지불하고, 집에 스레드리퍼 / EPYC 워크스테이션이 한대씩 늘어가며 공장 설비가 점점 좋아지는 중이다.</p><h4>(labor) cost optimization</h4><p>생산물의 결과도 다행히 나쁘지 않았다. 아래는 수익성 있는 것으로 판별된 전략들의 Heatmap 인데, 그냥 상관관계가 낮으면 좋다 정도로 이해하면 된다. 이걸 사람을 고용해서 이정도를 만들려면, 적어도 10명 이상의 팀이 1년 이상을 일해야 됐다고 생각한다. 요즘 이런 업종(?) 의 인건비 1억도 우스운 시대이지만… 인당 평균 인건비 + 기타비용 1.5억이 들어간다고 치면 15억 정도를 써야 만들 수 있는 것 아닌가 싶다.</p><p>근데 생각보다 크립토를 아주 큰 규모로 (특히 Mid Freq 쪽은) 운용하는 팀은 많지는 않고, 있더라도 MM / Arb 쪽이라 이정도 알파풀을 운용하는 곳은 손에 꼽을 것이다. (사람 고용하고 시간 쓸 생각하면 수지가 잘 안나오지 않는가.)</p><p>LLM 비용, 클라우드, 장비비 해서 5천만원 이내로 비슷한 결과를 3개월 남짓한 시간에 만들어 낼 수 있는 것이다.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/490/1*riZHuN0KjkKnESOKj0tUWg.png" /></figure><h4>Real Trading Results</h4><p>백테스트는 쿨하지 않으니까 첨부하지 않고, 실거래 결과만 첨부해보겠다. 운용 규모는 조금 밝히기 곤란하지만, 현재 기준 필자의 Binance 선물 월평균거래대금은 50M USD 이상이며 VIP 2 등급을 유지하고 있다.</p><p>월간 거래 횟수는 15–30만회 정도이며, Binance USD-M 선물 중 유동성이 일정 수준 이상 나오는 종목은 전부 건드리고 있다. Live Trading의 Sharpe Ratio는 대략 3.5 내외로 나오는 것 같다. (알파풀을 일주일 단위로 업그레이드 하고 있어, 최근구간은 4 이상 나온다.)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*BHmnhPXCSyO4eMM9V4Peig.jpeg" /></figure><h3>Conclusion</h3><p>퀀트 리서쳐는 연봉이 아주 높은 편에 속하는 고 부가가치 인력들이다. 이런 인력도 대체할 가능성이 있다. 특히, 기존에 Alpha가 있다고는 알고 있지만 딥하게 리서치 하기는 수지타산이 맞지 않아 몸집이 큰 곳에서는 잘 건드리지 않는 Niche한 알파들을, LLM 기반으로 아주 적은 비용에 발굴하여 먹는 것이 되게 유효할 거라고 생각한다. 이건 사실 초개인화된 니치한 서비스들이 쏟아져 나오고 있는 맥락과도 비슷하다.</p><p>나도 언젠가는 직원 하나 없이 그 몸집이 큰 회사들 사이에 설 수 있기를 바라며 오늘도 열심히 프롬프트를 깎고 있다. 직원을 거의 없애서 (사실 자를 직원도 없으니 이것도 틀린 말이긴 하다만) 영업이익률을 기존 산업 평균의 5배로 만들면, 매출은 1/10이어도, 내가 인센티브로 가져갈 수 있는 금액은 대략 비슷할 것이기 때문이다.</p><p>그게 앞으로의 모든 분야에서 일어날 변화 아닐까?</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3a70cd631fe8" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Undervalued AI Infrastructure Software Stocks to Watch]]></title>
            <link>https://bellman-h.medium.com/undervalued-ai-infrastructure-software-stocks-to-watch-b5b2a3b9ac43?source=rss-99e8e6896144------2</link>
            <guid isPermaLink="false">https://medium.com/p/b5b2a3b9ac43</guid>
            <dc:creator><![CDATA[bellman]]></dc:creator>
            <pubDate>Mon, 14 Apr 2025 03:02:13 GMT</pubDate>
            <atom:updated>2025-04-14T03:02:13.456Z</atom:updated>
            <content:encoded><![CDATA[<h3>Undervalued — Software, Overvalued — Hardware in AI</h3><p>Artificial Intelligence has sparked a gold rush in tech investing, with hardware leaders like NVIDIA grabbing headlines for powering AI model training. Yet, beyond the chips lies a rich ecosystem of <strong>AI infrastructure software</strong> — from developer platforms and databases to MLOps and APIs — that is arguably undervalued compared to the red-hot hardware sector. This report highlights publicly traded companies (primarily U.S.-listed, with a few international candidates) that provide the <strong>software “plumbing” for AI</strong> and may offer mid- to long-term growth potential. We cover diverse sub-sectors (DevOps, databases, ML orchestration, vector search, voice/text APIs, security tools for AI workflows, etc.), focusing on <strong>overlooked or underappreciated stocks</strong> — including some high-risk, high-reward small/mid-caps — that could benefit as AI adoption accelerates. Below, we analyze top opportunities in each category, followed by a comparative table of key metrics.</p><h3>AI Development &amp; MLOps Platforms</h3><p>These companies provide platforms to build, deploy, or manage AI/ML applications — essentially the operating systems for AI projects — and have been <strong>gaining traction while remaining undervalued</strong> relative to hardware plays.</p><ul><li><strong>Palantir Technologies (PLTR)</strong> — A leader in <strong>enterprise AI platforms</strong> and data analytics. Palantir’s Foundry and new Artificial Intelligence Platform (AIP) help governments and corporations harness AI for decision-making. The stock has rallied massively (over <strong>590% year-on-year as of early 2025</strong> (<a href="https://www.indmoney.com/blog/us-stocks/palantir-technologies-m-cap-rise-how-it-became-one-of-world-s-most-valuable-ai-firms-in-a-year#:~:text=Palantir%27s%20594%25%20m,year%20rise">Palantir’s 594% m-cap surge: How it became one of the world’s most …</a>)) on rising demand, yet some analysts still see upside. Wedbush Securities calls Palantir a <em>“top name to own in 2025,”</em> expecting it to <strong>benefit from a “tidal wave” of federal AI spending</strong> — even as overall government IT budgets tighten (<a href="https://www.investopedia.com/wedbush-calls-palantir-a-top-stock-to-own-in-2025-11689615#:~:text=However%2C%20Wedbush%20analysts%20told%20clients,government%20agencies%2C%E2%80%9D%20the%20analysts%20added">Wedbush Calls Palantir a Top Stock to Own in 2025</a>). Palantir’s heavy government exposure (40% of revenue) turned a risk into a strength, as efficiency initiatives drive agencies to invest in AI solutions (<a href="https://www.investopedia.com/wedbush-calls-palantir-a-top-stock-to-own-in-2025-11689615#:~:text=Shares%20of%20the%20Palantir%20lost,revenue%20in%20the%20fourth%20quarter">Wedbush Calls Palantir a Top Stock to Own in 2025</a>). The company is now consistently profitable on an adjusted basis and guided for ~$3.75B in 2025 sales (<a href="https://www.forbes.com/sites/dereksaul/2025/02/04/palantir-stock-soars-as-much-as-28-to-all-time-high-becoming-200-billion-company/#:~:text=Forbes%20www,5%20billion">Palantir Stock Surges After Earnings — Crosses $200 Billion … — Forbes</a>). While its valuation is no longer cheap (forward P/E &gt; 160 (<a href="https://finance.yahoo.com/quote/PLTR/#:~:text=Palantir%20Technologies%20Inc,Price%2FSales%20%28ttm">Palantir Technologies Inc. (PLTR) Stock Price, News, Quote &amp; History</a>)), Palantir’s <strong>strategic position as an AI infrastructure leader</strong> in both defense and commercial sectors could justify further growth.</li><li><strong>C3.ai, Inc. (AI)</strong> — An enterprise <strong>MLOps and AI application platform</strong> provider. C3.ai offers a suite of pre-built AI applications and a platform for developing custom AI solutions across industries. The stock was a 2020 high-flyer that crashed and has been beaten down (about <strong>10% down in early 2025</strong> after a rough start (<a href="https://www.nasdaq.com/articles/artificial-intelligence-ai-stock-already-down-10-2025-wall-street-analysts-believe-it#:~:text=C3,much%20more%20in%2012%20months">This Artificial Intelligence (AI) Stock Is Already Down 10% in 2025, but Wall Street Analysts Believe That It Could Soar Impressively | Nasdaq</a>)), trading around ~11× sales (<a href="https://www.nasdaq.com/articles/artificial-intelligence-ai-stock-already-down-10-2025-wall-street-analysts-believe-it#:~:text=The%20stock%27s%20valuation%20and%20potential,could%20be%20a%20solid%20investment">This Artificial Intelligence (AI) Stock Is Already Down 10% in 2025, but Wall Street Analysts Believe That It Could Soar Impressively | Nasdaq</a>). This lull may be an opportunity: C3’s revenue growth <em>re-accelerated</em> to <strong>~25% year-over-year</strong> in late 2024 (<a href="https://www.nasdaq.com/articles/artificial-intelligence-ai-stock-already-down-10-2025-wall-street-analysts-believe-it#:~:text=After%20all%2C%20C3,of%20the%20preceding%20fiscal%20year">This Artificial Intelligence (AI) Stock Is Already Down 10% in 2025, but Wall Street Analysts Believe That It Could Soar Impressively | Nasdaq</a>), and it raised its fiscal 2025 guidance to ~$388M (≈25% growth) (<a href="https://www.nasdaq.com/articles/artificial-intelligence-ai-stock-already-down-10-2025-wall-street-analysts-believe-it#:~:text=annually%20at%20the%20end%20of,the%20forecast%20period">This Artificial Intelligence (AI) Stock Is Already Down 10% in 2025, but Wall Street Analysts Believe That It Could Soar Impressively | Nasdaq</a>). The company has forged deep partnerships with cloud giants Microsoft, AWS, and Google to distribute its AI software (<a href="https://www.nasdaq.com/articles/artificial-intelligence-ai-stock-already-down-10-2025-wall-street-analysts-believe-it#:~:text=C3,of%20its%20generative%20AI%20offerings">This Artificial Intelligence (AI) Stock Is Already Down 10% in 2025, but Wall Street Analysts Believe That It Could Soar Impressively | Nasdaq</a>), including a strategic alliance with Microsoft to expand generative AI offerings (<a href="https://www.nasdaq.com/articles/artificial-intelligence-ai-stock-already-down-10-2025-wall-street-analysts-believe-it#:~:text=Alphabet%27s%20Google%20Cloud%2C%20Microsoft%20Azure%2C,of%20its%20generative%20AI%20offerings">This Artificial Intelligence (AI) Stock Is Already Down 10% in 2025, but Wall Street Analysts Believe That It Could Soar Impressively | Nasdaq</a>). C3.ai is also expanding in the public sector — U.S. government agencies are adopting its tools, a market expected to grow 20% annually through 2033 (<a href="https://www.nasdaq.com/articles/artificial-intelligence-ai-stock-already-down-10-2025-wall-street-analysts-believe-it#:~:text=the%20company%20announcing%20a%20strategic,of%20its%20generative%20AI%20offerings">This Artificial Intelligence (AI) Stock Is Already Down 10% in 2025, but Wall Street Analysts Believe That It Could Soar Impressively | Nasdaq</a>). With the <strong>AI software market projected to 10× (to $52B by 2028) (</strong><a href="https://www.nasdaq.com/articles/artificial-intelligence-ai-stock-already-down-10-2025-wall-street-analysts-believe-it#:~:text=annually%20at%20the%20end%20of,the%20forecast%20period"><strong>This Artificial Intelligence (AI) Stock Is Already Down 10% in 2025, but Wall Street Analysts Believe That It Could Soar Impressively | Nasdaq</strong></a><strong>)</strong>, C3.ai’s current ~$2.7B market cap (<a href="https://stockanalysis.com/stocks/ai/market-cap/#:~:text=C3,17.85%25%20in%20one%20year">C3.ai (AI) Market Cap &amp; Net Worth — Stock Analysis</a>) could undervalue its long-term upside if it continues executing. Investors should note C3.ai is not yet profitable and carries execution risk, but its strong balance sheet and <em>focused AI niche</em> make it a classic high-risk/high-reward play.</li><li><strong>Nebius Group N.V. (NBIS)</strong> — An <em>under-the-radar</em> <strong>AI cloud and platform provider</strong> that recently listed on Nasdaq. Nebius — a spin-off of Yandex’s cloud division — operates an AI-native cloud platform offering computing, storage, and machine learning tools (essentially a smaller scale AWS tailored for AI startups and SMBs). Few investors know this name, but Nebius may be <em>one of the largest tech companies you’ve “never heard of.”</em> It grew revenue <strong>462% in 2024</strong> and is targeting a hefty $750M–$1B in annual recurring revenue in the near future (<a href="https://seekingalpha.com/article/4772706-nebius-group-trades-at-a-steep-discount#:~:text=Nebius%20Group%20Trades%20At%20A,Learn%20why%20NBIS%20stock">Nebius Group Trades At A Steep Discount — Seeking Alpha</a>). Despite that growth, the stock trades at only <strong>~2.9× EV/ARR</strong>, a steep discount to peers (<a href="https://seekingalpha.com/article/4772706-nebius-group-trades-at-a-steep-discount#:~:text=Nebius%20Group%20Trades%20At%20A,Learn%20why%20NBIS%20stock">Nebius Group Trades At A Steep Discount — Seeking Alpha</a>). Its market cap hovers around $5 billion (<a href="https://finance.yahoo.com/quote/NBIS/#:~:text=History%20finance,Financial">Nebius Group N.V. (NBIS) Stock Price, News, Quote &amp; History</a>) after a quiet October 2024 IPO, and shares are roughly flat since listing (down ~19% YTD 2025 after an initial post-IPO uptick) (<a href="https://www.cmcmarkets.com/en/optox/nbis-stock-nebius-is-an-under-the-radar-ai-play#:~:text=NBIS%20Stock%3A%20Nebius%20Is%20an,in%20October%2C%20but%20has">NBIS Stock: Nebius Is an Under-the-Radar AI Play | CMC Markets</a>). Nebius’s strategic edge is its <strong>cost-effective cloud infrastructure for AI</strong> workloads — a compelling niche as demand from AI startups explodes. Of course, as a smaller entrant it faces intense competition from hyperscalers, and its roots in Yandex (a Russian firm) may pose geopolitical risk. But for investors seeking an <strong>“off-the-radar” AI infrastructure play</strong>, Nebius offers explosive growth at a reasonable valuation.</li></ul><h3>Data Management and Database Companies</h3><p><strong>Data is the fuel of AI</strong>, and companies that manage, store, and retrieve data efficiently are critical in the AI stack. Unlike glamorous chipmakers, many data infrastructure providers trade at moderate multiples due to slowing growth or past hype burnout — even as they pivot to support AI workloads (e.g. adding vector search, AI integrations). This could spell opportunity.</p><ul><li><strong>MongoDB, Inc. (MDB)</strong> — A leading modern database platform (NoSQL document database) widely used by developers and increasingly in AI applications. MongoDB’s Atlas cloud service now offers <strong>native vector search</strong> capabilities, allowing AI developers to store and query embedding vectors for semantic search (<a href="https://www.mongodb.com/en-us/products/platform/atlas-vector-search#:~:text=Vector%20Search,featured%20vector%20database%20capabilities">Atlas Vector Search — MongoDB</a>). This positions MongoDB as a player in the emerging <em>vector database</em> space. The company’s growth has cooled from hyper-speed to a still-solid ~20% range — FY 2025 revenue was <strong>$2.01B (+19% YoY)</strong> (<a href="https://stockanalysis.com/stocks/mdb/revenue/#:~:text=MongoDB%20%28MDB%29%20Revenue%202016,in%20the%20quarter%20ending">MongoDB (MDB) Revenue 2016–2025 — Stock Analysis</a>) — and the stock suffered a sharp pullback (market cap down ~50% in the past year to ~$13B as of April 2025 (<a href="https://stockanalysis.com/stocks/mdb/market-cap/#:~:text=MongoDB%20,17B">MongoDB (MDB) Market Cap &amp; Net Worth — Stock Analysis</a>)). At ~6.5× forward sales, MongoDB looks comparatively cheap given its strong developer mindshare and expanding AI feature set. It continues to grow its cloud Atlas customer base (22% YoY total revenue rise last quarter (<a href="https://investors.mongodb.com/news-releases/news-release-details/mongodb-inc-announces-first-quarter-fiscal-2025-financial#:~:text=MongoDB%2C%20Inc,year">MongoDB, Inc. Announces First Quarter Fiscal 2025 Financial Results</a>)) and is nearing breakeven on a cash flow basis. As AI workloads require flexible, scalable databases for unstructured data, MongoDB’s <strong>document model with AI-focused add-ons</strong> could see renewed demand. The stock’s downturn suggests it’s been <em>overlooked amid the AI frenzy</em>, but any uptick in usage for AI projects (e.g. powering GenAI applications that need real-time document storage or vector similarity search) could rerate the shares.</li><li><strong>Snowflake Inc. (SNOW)</strong> — The cloud data warehousing company that became synonymous with Big Data hype. Snowflake provides a platform for storing and analyzing large datasets, and it has pivoted hard into the AI trend by enabling customers to build AI applications atop their data cloud. After astronomical early growth, Snowflake’s revenue is now expanding at a more modest clip (<strong>+28% YoY to $987M</strong> in the latest reported quarter) (<a href="https://www.marketbeat.com/originals/snowflake-boosts-growth-by-doubling-down-on-ai/#:~:text=Snowflake%20Boosts%20Growth%20by%20Doubling,FQ1%202025">Snowflake Boosts Growth by Doubling Down on AI — MarketBeat</a>). Yet this still represents robust expansion at scale — full-year FY2025 sales were ~$3.9B. The stock, which once traded at nosebleed valuations, has come down to earth around <strong>$48–50B market cap</strong> (<a href="https://finance.yahoo.com/quote/SNOW/#:~:text=Performance%20Overview%3A%20SNOW,Enterprise%20Value">Snowflake Inc. (SNOW) Stock Price, News, Quote &amp; History</a>) (~12–13× forward sales). Snowflake’s <strong>strategic positioning</strong> in AI is strong: its platform is used to feed data into AI models and to host the results, and Snowflake has introduced features to support data preparation for ML and even launched an app marketplace for AI/ML tools. The company is also doubling down on partnerships (e.g. with NVIDIA for accelerated computing and with OpenAI startups). With <strong>126% net revenue retention</strong> (<a href="https://investors.snowflake.com/financials/quarterly-results/default.aspx#:~:text=Relations%20investors,580%20customers%20with%20trailing">Financials — Quarterly Results — Snowflake — Investor Relations</a>), Snowflake is still expanding within its customer base. Investors have cooled on the name since the IPO frenzy, but that may be exactly the reason to look at it now — it’s a <strong>critical piece of AI data infrastructure</strong> with a more palatable valuation and plenty of runway as enterprises invest in AI-driven analytics.</li><li><strong>Confluent, Inc. (CFLT)</strong> — A smaller-cap data infrastructure play, Confluent is the leading provider of enterprise <strong>data streaming</strong> via Apache Kafka. Its platform enables real-time data pipelines — crucial for AI/ML systems that ingest continuous data (think event streams from apps, logs, IoT feeds for machine learning). Confluent’s growth in 2024 was ~24%, with $964M annual revenue (<a href="https://stockanalysis.com/stocks/cflt/revenue/#:~:text=Confluent%20%28CFLT%29%20Revenue%202019,December%2031%2C%202024%2C%20with">Confluent (CFLT) Revenue 2019–2024 — Stock Analysis</a>), and its cloud segment grew faster at +38%. At ~$10B market cap (<a href="https://macrotrends.net/stocks/charts/CFLT/confluent/market-cap#:~:text=Confluent%20Market%20Cap%202020,CFLT%20With%20Other%20Stocks">Confluent Market Cap 2020–2024 | CFLT — Macrotrends</a>) (~10× sales), Confluent isn’t a bargain-basement stock, but it is <strong>one of the only pure-play streaming data companies</strong>. As AI applications increasingly demand real-time data (e.g. live personalization algorithms, dynamic monitoring of ML models, etc.), Confluent’s technology becomes mission-critical. The stock is down from highs (it dipped as low as ~$6.7B cap in 2024 before rebounding) (<a href="https://www.gurufocus.com/term/mktcap/CFLT/Market-Cap/#:~:text=CFLT%20%28Confluent%29%20Market%20Cap%20,8">CFLT (Confluent) Market Cap — GuruFocus</a>), reflecting earlier concerns about tech spending, but the company is nearing profitability and improving operating margins. Confluent might not have the name recognition of other AI plays, but it fills an important niche in the AI infrastructure stack and could be an <em>acquisition target</em> in a consolidating industry. Investors looking for a reasonably valued data software stock with an AI angle should keep an eye on it.</li><li><strong>Elastic N.V. (ESTC)</strong> — The company behind Elasticsearch, a popular search and analytics engine, which is now integrating generative AI into its platform. Elastic enables search across logs, metrics, and textual data — capabilities that align with AI/ML ops (for instance, monitoring ML model outputs or enabling semantic search). After a rough patch, Elastic reported <strong>17% revenue growth</strong> in Q3 FY2025 and raised its outlook, crediting <strong>generative AI traction</strong> for accelerating its search business (<a href="https://www.fool.com/investing/2025/03/30/tech-sell-off-1-ai-stock-down-46-wish-bought-dip/#:~:text=Tech%20Sell,ago">Tech Sell-Off: 1 Spectacular Artificial Intelligence (AI) Stock Down 46 …</a>) (<a href="https://www.fool.com/investing/2025/03/30/tech-sell-off-1-ai-stock-down-46-wish-bought-dip/#:~:text=,ago">Tech Sell-Off: 1 Spectacular Artificial Intelligence (AI) Stock Down 46 …</a>). Notably, Elastic’s valuation has become attractive: about <em>4.2× forward revenue</em> according to a recent analysis (<a href="https://seekingalpha.com/article/4774044-elastic-rating-upgrade-on-aicloud-momentum-and-better-leading-growth-metrics#:~:text=Elastic%20,2x%20forward%20revenue">Elastic (ESTC): Rating Upgrade On AI/Cloud Momentum And Better …</a>), which is low for a software company with cloud momentum. The stock popped 15% on its strong earnings and AI commentary (<a href="https://seekingalpha.com/article/4763318-elastic-q3-earnings-investors-buying-into-ai-growth-story#:~:text=Elastic%20Q3%20Earnings%3A%20Why%20Investors,AI%20and%20improved%20profitability%2C">Elastic Q3 Earnings: Why Investors Are Buying Into The AI Growth …</a>) but still trades far below its 2021 highs. With ~43% gross margins and improving profitability (<a href="https://simplywall.st/stocks/hk/software/hkg-20/sensetime-group-shares#:~:text=SenseTime%20Group%20%28SEHK%3A20%29%20,Company%20Analysis">SenseTime Group (SEHK:20) — Stock Price, News &amp; Analysis</a>), Elastic is now a turnaround story. The <strong>built-in AI features</strong> (like vector search and an LLM-based assistant in Kibana) could reinvigorate its growth. While not a pure-play AI firm, Elastic’s core competency in scalable search/analytics makes it an enabling technology for AI systems, and its depressed valuation could mean substantial upside if the AI push gains steam (<a href="https://seekingalpha.com/article/4774044-elastic-rating-upgrade-on-aicloud-momentum-and-better-leading-growth-metrics#:~:text=Elastic%20,2x%20forward%20revenue">Elastic (ESTC): Rating Upgrade On AI/Cloud Momentum And Better …</a>).</li></ul><p><em>(The data management category is crowded — other notable mentions include </em><strong><em>Databricks</em></strong><em> and </em><strong><em>Pinecone</em></strong><em> in vector databases (both still private), and legacy giants like </em><strong><em>Oracle</em></strong><em> and </em><strong><em>IBM</em></strong><em> pivoting to AI-ready databases/clouds. However, we focus on the above public, mid-sized firms as compelling, more “overlooked” opportunities.)</em></p><h3>Developer Tools &amp; DevOps for AI</h3><p>Software development and deployment for AI introduce new workflow challenges — from managing code and models to automating infrastructure. Companies in DevOps and development tooling that have embraced AI features or cater to AI developers may see <strong>renewed growth</strong>. Many of these stocks slumped in 2022–23 and trade at reasonable valuations now.</p><ul><li><strong>GitLab Inc. (GTLB)</strong> — A DevSecOps platform (source code management and CI/CD) that is increasingly integrating AI to assist developers. GitLab has launched <strong>AI-powered coding features (GitLab Duo)</strong> and seen strong upsell of its higher-tier plans (<a href="https://www.itiger.com/news/2503613190#:~:text=When%20it%20comes%20to%20future,also%20been%20gaining%20strong%20traction">SoundHound vs. GitLab: Better Mid-Cap AI Stock Investment for 2025 — Tiger Brokers</a>). The company’s fundamentals are impressive: revenue grew <strong>29% YoY to $211M</strong> in the latest quarter (<a href="https://www.tipranks.com/news/company-announcements/gitlab-reports-strong-growth-in-fiscal-year-2025#:~:text=GitLab%20Reports%20Strong%20Growth%20in,million%2C%20alongside%20a%20notable">GitLab Reports Strong Growth in Fiscal Year 2025 — TipRanks.com</a>), with <strong>89% gross margins</strong> (<a href="https://www.itiger.com/news/2503613190#:~:text=While%20SoundHound%20has%20been%20growing,gap%20is%20currently%20very%20wide">SoundHound vs. GitLab: Better Mid-Cap AI Stock Investment for 2025 — Tiger Brokers</a>) and narrowing losses (non-GAAP operating margin improved to -7% in FY25 Q4 from -21% a year prior (<a href="https://www.tipranks.com/news/company-announcements/gitlab-reports-strong-growth-in-fiscal-year-2025#:~:text=GitLab%20Reports%20Strong%20Growth%20in,million%2C%20alongside%20a%20notable">GitLab Reports Strong Growth in Fiscal Year 2025 — TipRanks.com</a>)). Free cash flow turned positive, and GitLab expects ~24% growth this year (<a href="https://www.nasdaq.com/articles/gitlab-continues-see-strong-revenue-growth-powered-ai-stock-buy#:~:text=GitLab%20Continues%20to%20See%20Strong,is%20looking%20for%20adjusted">GitLab Continues to See Strong Revenue Growth Powered by AI …</a>). Despite this, the stock hasn’t been swept up in AI hype — it trades around ~$7B market cap (≈9× forward sales) (<a href="https://www.itiger.com/news/2503613190#:~:text=From%20a%20valuation%20standpoint%2C%20GitLab,times%202025%20analyst%20estimates">SoundHound vs. GitLab: Better Mid-Cap AI Stock Investment for 2025</a>) (<a href="https://www.itiger.com/news/2503613190#:~:text=Valuation%20and%20verdict">SoundHound vs. GitLab: Better Mid-Cap AI Stock Investment for 2025 — Tiger Brokers</a>), which is lower than many peers. In fact, <strong>GitLab looks cheaper than other mid-cap AI plays</strong>: for example, it commands ~9.9× 2025 sales vs. SoundHound’s 23× (<a href="https://www.itiger.com/news/2503613190#:~:text=Valuation%20and%20verdict">SoundHound vs. GitLab: Better Mid-Cap AI Stock Investment for 2025 — Tiger Brokers</a>). After an initial post-IPO dip, shares are recovering (up ~20% in the last 6 months) (<a href="https://www.nasdaq.com/articles/gtlb-stock-rises-21-6-months-will-rally-continue-2025#:~:text=GTLB%20Stock%20Rises%2021,8">GTLB Stock Rises 21% in 6 Months: Will the Rally Continue in 2025?</a>), but the valuation still appears reasonable for a 30% grower. With code hosting and CI/CD being central to AI software development (ML engineers need to version control models and pipelines too), GitLab’s <strong>entrenched position with developers</strong> plus its AI features could drive continued expansion. It’s one of Cathie Wood’s noted AI picks for 2025 as well (<a href="https://finance.yahoo.com/news/gitlab-inc-gtlb-among-cathie-152559531.html#:~:text=Gitlab%20Inc,AI%20stock%20picks%20in%202025">Gitlab Inc. (GTLB): Among Cathie Wood’s Top AI Stock Picks in 2025</a>).</li><li><strong>Progress Software (PRGS)</strong> — A lesser-known but established infrastructure software provider that has reinvented itself for the AI era. Progress offers a portfolio of products for application development, data connectivity, and recently <strong>AI-powered automation</strong>. It acquired a platform called ShareFile in 2023 to add AI-driven workflow automation and document management to its offerings (<a href="https://www.investing.com/analysis/progress-software-making-progress-driven-by-the-ai-revolution-200659234#:~:text=powered%20tools">Progress Software: Is the Stock Still Undervalued After Its AI-Powered Q1 Beat? | Investing.com</a>). After hitting a <strong>nine-month stock low</strong>, Progress delivered an “AI-powered” earnings beat in Q1 FY2025 — EPS of $1.31 (+25¢ vs estimates) and revenue up <strong>28.9% YoY</strong> to $238M (<a href="https://www.investing.com/analysis/progress-software-making-progress-driven-by-the-ai-revolution-200659234#:~:text=Progress%20Software%E2%80%99s%20stock%20had%20reached,62%20million">Progress Software: Is the Stock Still Undervalued After Its AI-Powered Q1 Beat? | Investing.com</a>) — thanks in part to the new acquisition and AI momentum. Annual recurring revenue jumped 48% YoY (<a href="https://www.investing.com/analysis/progress-software-making-progress-driven-by-the-ai-revolution-200659234#:~:text=beating%20consensus%20analyst%20estimates%20of,62%20million">Progress Software: Is the Stock Still Undervalued After Its AI-Powered Q1 Beat? | Investing.com</a>). Management raised guidance and now projects ~$5.30 EPS for full-year 2025 (<a href="https://www.investing.com/analysis/progress-software-making-progress-driven-by-the-ai-revolution-200659234#:~:text=Progress%20Software%20issued%20upside%20FQ2,82%20million">Progress Software: Is the Stock Still Undervalued After Its AI-Powered Q1 Beat? | Investing.com</a>). Despite this growth, the stock trades at roughly <strong>10× forward earnings</strong>, an arguably low multiple (<a href="https://www.investing.com/analysis/progress-software-making-progress-driven-by-the-ai-revolution-200659234#:~:text=Progress%20Software%E2%80%99s%20stock%20had%20reached,62%20million">Progress Software: Is the Stock Still Undervalued After Its AI-Powered Q1 Beat? | Investing.com</a>) (<a href="https://stockanalysis.com/stocks/prgs/market-cap/#:~:text=Progress%20Software%20has%20a%20market,Full%20Definition">Progress Software (PRGS) Market Cap &amp; Net Worth — Stock Analysis</a>) for a company nearing $1B in annual sales with high 30% adjusted operating margins (<a href="https://www.investing.com/analysis/progress-software-making-progress-driven-by-the-ai-revolution-200659234#:~:text=beating%20consensus%20analyst%20estimates%20of,62%20million">Progress Software: Is the Stock Still Undervalued After Its AI-Powered Q1 Beat? | Investing.com</a>). Progress is not a pure AI company, but it is embedding AI across its product suite (e.g. chatbots for customer support, AI-assisted app development). With a market cap of only ~$2.5B (<a href="https://stockanalysis.com/stocks/prgs/market-cap/#:~:text=Progress%20Software%20,Full%20Definition">Progress Software (PRGS) Market Cap &amp; Net Worth — Stock Analysis</a>), solid profitability, and a renewed growth story, Progress Software appears <em>undervalued</em> — the market may be overlooking it because it’s an “old” software name, but its <strong>pragmatic AI strategy</strong> and steady execution could drive further upside.</li><li><strong>HashiCorp (HCP)</strong> — <em>[Acquired by IBM in 2024]</em> One notable DevOps player was HashiCorp, known for infrastructure-as-code tools like Terraform and Vault that automate cloud setup (widely used to provision AI training environments). HashiCorp had been a promising public company in this space, but <strong>IBM acquired it for $6.4B</strong> to bolster its hybrid cloud and AI capabilities (<a href="https://erp.today/ibm-acquires-hashicorp-enhancing-hybrid-cloud-and-ai-capabilities/#:~:text=IBM%20Acquires%20HashiCorp%2C%20Enhancing%20Hybrid,AI%20capabilities%20through%20the">IBM Acquires HashiCorp, Enhancing Hybrid Cloud and AI Capabilities</a>) (<a href="https://newsroom.ibm.com/2025-02-27-ibm-completes-acquisition-of-hashicorp,-creates-comprehensive,-end-to-end-hybrid-cloud-platform#:~:text=IBM%20Completes%20Acquisition%20of%20HashiCorp%2C,infrastructure%20that%20underpins%20hybrid">IBM Completes Acquisition of HashiCorp, Creates Comprehensive …</a>). This underscores the value of cloud automation in AI workflows. While HashiCorp is no longer independent, its peers — e.g. infrastructure automation firms or cloud management platforms — might present similar opportunities. (Investors interested in this angle could consider <strong>IBM (IBM)</strong> itself as an undervalued AI play, given its low P/E and focus on AI software after acquisitions like Red Hat and HashiCorp.)</li></ul><h3>AI-Powered APIs and Services (Speech &amp; Text)</h3><p>Many AI applications are delivered via APIs — for example, speech recognition, language understanding, and communication services. Companies offering these AI-driven APIs or platforms for building voice/text AI applications stand to benefit from the proliferation of AI in end-user products. Some of these stocks have been volatile but could be poised for long-term growth as AI becomes ubiquitous in software.</p><ul><li><strong>Twilio Inc. (TWLO)</strong> — The leading cloud communications API provider (for SMS, voice calls, chat) that is integrating AI to enhance customer engagement. Twilio’s platform is not an AI system itself, but it enables developers to build <strong>AI-powered communications</strong> — e.g. contact centers that use Twilio for messaging and incorporate AI bots for responses, or phone IVR systems with speech recognition. After a post-pandemic crash, Twilio’s stock doubled in the latter half of 2024 (<a href="https://finance.yahoo.com/news/twilio-99-six-months-stock-131200624.html#:~:text=Twilio%20Up%2099,over%20the%20past%20six%20months">Twilio Up 99% in Six Months: Will the Stock Carry Momentum in 2025?</a>) as the company refocused on profitability and saw renewed investor interest. Even after a ~99% jump in six months, some analysts still call it <strong>undervalued</strong> based on fundamentals (<a href="https://www.nasdaq.com/articles/twilio-99-six-months-will-stock-carry-momentum-2025#:~:text=Twilio%20Up%2099,conscious%20investors">Twilio Up 99% in Six Months: Will the Stock Carry Momentum in 2025?</a>). Twilio’s forward outlook is modest (guiding ~7–8% organic revenue growth in 2025 (<a href="https://www.fool.com/data-news/2025/02/13/twilio-beats-on-revenue-misses-on-eps/#:~:text=Twilio%20Beats%20on%20Revenue%2C%20Misses,GAAP">Twilio Beats on Revenue, Misses on EPS | The Motley Fool</a>)), but that could prove conservative if its AI initiatives gain traction. It launched an <strong>AI-powered customer data platform (Twilio Segment)</strong> and is working on AI assistants for contact center (Twilio Flex). The stock trades at around 2–3× sales — low for a software company — and the company could return to strong growth if usage of its APIs increases with AI-driven apps (for example, every chatbot that sends a text uses Twilio under the hood). Twilio is a bit of a turnaround/value story; it’s <em>mid-cap ($15B)</em> and not as high-risk as some smaller names, but it’s a key enabler of AI in communications and might be <strong>overlooked relative to trendier AI names</strong>.</li><li><strong>SoundHound AI, Inc. (SOUN)</strong> — A pure-play <strong>voice AI platform</strong> that provides speech recognition and natural language understanding technology. SoundHound’s Houndify platform enables more natural, conversational voice interactions — it’s used in voice assistants for cars, smart devices, restaurants (voice ordering), and more (<a href="https://www.itiger.com/news/2503613190#:~:text=SoundHound%20operates%20a%20voice%20AI,insurance%2C%20retail%2C%20and%20banking%20spaces">SoundHound vs. GitLab: Better Mid-Cap AI Stock Investment for 2025 — Tiger Brokers</a>) (<a href="https://www.itiger.com/news/2503613190#:~:text=understanding%20technology%2C%20which%20can%20help,insurance%2C%20retail%2C%20and%20banking%20spaces">SoundHound vs. GitLab: Better Mid-Cap AI Stock Investment for 2025 — Tiger Brokers</a>). This company exemplifies <em>high-risk, high-reward</em>: in 2024 its stock went on a wild ride — soaring over <strong>1000%</strong> at one point (<a href="https://www.nasdaq.com/articles/soundhound-ai-stock-1051-2024-where-could-it-be-end-2025#:~:text=SoundHound%20AI%20Stock%20Is%20Up,Based%20on%20its%20trailing">SoundHound AI Stock Is Up 1,051% in 2024. Where Could It Be at …</a>) — as investors latched onto its AI potential. After a subsequent pullback, SoundHound’s market cap is around $3–4B (<a href="https://finance.yahoo.com/quote/SOUN/key-statistics/#:~:text=,72B">SoundHound AI, Inc. (SOUN) Valuation Measures &amp; Financial …</a>). The fundamentals are improving dramatically: Q4 2024 revenue jumped <strong>101%</strong>, and full-year 2024 revenue was $84.7M (+85% YoY) (<a href="https://www.investors.com/news/technology/soun-stock-soundhound-ai-q4-2024-earnings/#:~:text=Targets%20www,to%20%2484.7%20million">SoundHound Stock Howls After Voice AI Firm Tops Q4 Targets</a>). SoundHound <em>raised</em> its 2025 revenue outlook to ~$167M at midpoint, which would be <strong>~97% growth</strong> (<a href="https://www.investors.com/news/technology/soun-stock-soundhound-ai-q4-2024-earnings/#:~:text=Targets%20www,to%20%2484.7%20million">SoundHound Stock Howls After Voice AI Firm Tops Q4 Targets</a>). Gross margins (~50%) are expected to rise as revenue scales (<a href="https://www.itiger.com/news/2503613190#:~:text=While%20SoundHound%20has%20been%20growing,gap%20is%20currently%20very%20wide">SoundHound vs. GitLab: Better Mid-Cap AI Stock Investment for 2025 — Tiger Brokers</a>). The company even made a small EBITDA profit in Q4, unusual for a firm at this early stage (<a href="https://investors.soundhound.com/news-releases/news-release-details/soundhound-ai-reports-record-fourth-quarter-revenue-101#:~:text=SoundHound%20AI%20Reports%20Record%20Fourth,For%20more%20information">SoundHound AI Reports Record Fourth Quarter Revenue, Up 101 …</a>). The big picture: SoundHound is aiming to be the go-to independent <strong>voice AI provider</strong> across industries, positioning itself as an alternative to big tech’s voice assistants. It recently acquired Amelia, an AI dialog platform, to expand into enterprise use cases (<a href="https://www.itiger.com/news/2503613190#:~:text=understanding%20technology%2C%20which%20can%20help,insurance%2C%20retail%2C%20and%20banking%20spaces">SoundHound vs. GitLab: Better Mid-Cap AI Stock Investment for 2025 — Tiger Brokers</a>). Competition is a concern — giants like Salesforce and Microsoft (with Nuance) are also tackling AI voice assistants (<a href="https://www.itiger.com/news/2503613190#:~:text=This%20is%20a%20huge%20opportunity,and%20other%20fields%20with%20Nuance">SoundHound vs. GitLab: Better Mid-Cap AI Stock Investment for 2025 — Tiger Brokers</a>) — but SoundHound’s specialized focus and head start in certain verticals (automotive, hospitality) give it a fighting chance. Given its small revenue base and substantial cash burn, this stock is not for the faint of heart, but if one believes voice interfaces will be a key part of the AI future, <strong>SoundHound could be significantly undervalued</strong> relative to its long-term market opportunity.</li><li><strong>LivePerson, Inc. (LPSN)</strong> — A mentionable player in AI-driven text communications. LivePerson provides chatbot and messaging solutions for customer service, leveraging conversational AI. The stock has been beaten down (market cap well under $1B) after growth stalled, but the company is pivoting to Generative AI for chatbots. It’s a risky turnaround candidate that might see a revival if its new GPT-based offerings gain clients. (Not a top pick, but worth noting as a <em>speculative small-cap</em> in AI text interfaces.)</li></ul><p><em>(Another notable API-centric company was </em><strong><em>Nuance Communications</em></strong><em>, a speech recognition pioneer, but it was acquired by Microsoft in 2022. Many large-cap tech firms like </em><strong><em>Microsoft</em></strong><em>, </em><strong><em>Google</em></strong><em>, and </em><strong><em>Meta</em></strong><em> provide AI APIs too — but those are well-known and not “undervalued” in the small/mid-cap sense. Our focus is on the independent players.)</em></p><h3>Security and Infrastructure Protection for AI Workflows</h3><p>As businesses deploy AI, securing AI systems and infrastructure becomes vital — from protecting data pipelines to using AI for threat detection. Cybersecurity firms with AI-driven products, or those securing cloud workloads, can be considered part of the AI infrastructure story. Some have seen less hype than AI hardware, creating potential value gaps.</p><ul><li><strong>SentinelOne, Inc. (S)</strong> — A cybersecurity company offering <strong>endpoint protection</strong> that is natively AI-powered. SentinelOne’s platform uses machine learning to detect malware and automate threat response on devices and cloud workloads. It was a market darling in 2021 that later plunged over 75% from its peak, and as of early 2025 the stock remains <strong>~18% below year-ago levels</strong> (<a href="https://stockanalysis.com/stocks/s/market-cap/#:~:text=SentinelOne%20,80B">SentinelOne (S) Market Cap &amp; Net Worth — Stock Analysis</a>). At ~$5.8B market cap (<a href="https://pitchbook.com/profiles/company/58281-31#:~:text=Earnings%20pitchbook,58">SentinelOne 2025 Company Profile: Stock Performance &amp; Earnings</a>), SentinelOne trades around 6–7× sales — comparatively modest for a company that just grew revenue <strong>32% to $821M in FY2025</strong> (<a href="https://stockanalysis.com/stocks/s/revenue/#:~:text=SentinelOne%20%28S%29%20Revenue%202020,52M%20in%20the%20quarter">SentinelOne (S) Revenue 2020–2025 — Stock Analysis</a>). Its annual recurring revenue (ARR) is $920M (+27% YoY) (<a href="https://investors.sentinelone.com/press-releases/news-details/2025/SentinelOne-Announces-Fourth-Quarter-and-Fiscal-Year-2025-Financial-Results/default.aspx#:~:text=,1%20million%20as">SentinelOne Announces Fourth Quarter and Fiscal Year 2025 …</a>), and while growth is slowing, the company is on path to profitability with improving gross margins. What makes SentinelOne compelling is its <strong>strong positioning in AI-driven security</strong> — as generative AI opens new threat vectors (and also helps hackers), enterprises will need advanced tools to protect systems. SentinelOne is expanding beyond endpoint into cloud security and security data analytics (it even launched an AI-powered security data lake, integrating SecOps with big data). Analysts have noted the stock’s valuation is now relatively low and expect a “rerating” as the company proves its durability (<a href="https://seekingalpha.com/article/4763956-sentinelone-stock-why-major-rerating-likely-coming#:~:text=SentinelOne%3A%20Why%20A%20Major%20Rerating,SIEM%20adoption%2C%20cloud">SentinelOne: Why A Major Rerating Is Likely Coming — Seeking Alpha</a>) (<a href="https://seekingalpha.com/article/4760736-sentinelone-why-i-am-buying-more-now#:~:text=Alpha%20seekingalpha,and%20a%20compelling%20buying">SentinelOne: Why I Am Buying More Now (NYSE:S) | Seeking Alpha</a>). Indeed, SentinelOne’s forward enterprise-value-to-revenue is a fraction of rival CrowdStrike’s, despite similar innovation in AI. The risk here is high competition and the need to manage sales growth vs. losses, but if one is bullish that <strong>“AI cybersecurity” will be a major theme</strong>, SentinelOne at these levels could be a <em>rewarding pick</em>.</li><li><strong>Fortinet, Inc. (FTNT)</strong> — A larger, established cybersecurity vendor that deserves mention for its <strong>AI and security infrastructure</strong> convergence. Fortinet provides firewall appliances and network security software, and it has long embedded AI/ML in its threat detection (its FortiAI systems help detect anomalies in network traffic). Unlike many pure AI plays, Fortinet is profitable with steady growth (~20% revenue growth, ~$5B annual sales). The stock trades at a reasonable multiple (~30× earnings) and has been overshadowed by flashier AI names. Yet, as companies build out AI data centers and cloud environments, Fortinet’s security fabric (which secures on-prem and cloud networks) will be in demand — they even specifically target <strong>AI data center security</strong>. Fortinet may not offer explosive upside, but it’s a <strong>“picks and shovels” play for securing AI infrastructure</strong>, with a track record and stability. After a pullback from 2023 highs, its valuation looks more attractive, making it a solid <em>medium-risk candidate</em> to ride the AI wave in a more conservative way.</li><li><strong>CrowdStrike Holdings (CRWD)</strong> — Another prominent AI-driven security company (endpoint and cloud workload protection). CrowdStrike is larger (~$30B cap) and still growing ~35% YoY, using AI to identify threats across millions of endpoints. It’s not exactly under-the-radar — the stock is well-known — but it hasn’t seen the frenzy of an NVIDIA. Trading around 12× forward sales, CrowdStrike is more expensive than SentinelOne or Fortinet, but it’s considered a leader in applying AI to cybersecurity. For investors more comfortable with established players, CrowdStrike is a <em>strong strategic pick</em> (though arguably not “undervalued” in the strict sense). Its inclusion here is to note that <strong>AI software isn’t just about model training — it also secures the enterprise</strong>, and CrowdStrike exemplifies that.</li></ul><h3>International Contenders (Asia)</h3><p>While U.S. markets host most of the investable AI software plays, some overseas companies in Japan and Hong Kong offer exposure to AI infrastructure themes — often at lower valuations due to lesser coverage or local market factors. One notable example:</p><ul><li><strong>SenseTime Group (0020.HK)</strong> — China’s largest AI software company, known for its computer vision algorithms and AI cloud platform (SenseCore). SenseTime is a <strong>publicly traded pure-play AI firm</strong> on the Hong Kong exchange. It specializes in facial recognition, image processing, and has developed its own large AI models (the <strong>SenseNova</strong> model family). In 2024, despite China’s economic headwinds, SenseTime’s revenue grew <strong>10.8%</strong> and losses narrowed by about one-third (<a href="https://www.tipranks.com/news/company-announcements/sensetime-reports-revenue-growth-and-ai-advancements-in-2024#:~:text=SenseTime%20Group%2C%20Inc,engagement%20and%20willingness%20to%20pay">SenseTime Reports Revenue Growth and AI Advancements in 2024 — TipRanks.com</a>). The company touts that its SenseNova model achieved <em>“industry-leading performance, surpassing OpenAI’s GPT-4 Turbo”</em> on certain benchmarks (<a href="https://www.tipranks.com/news/company-announcements/sensetime-reports-revenue-growth-and-ai-advancements-in-2024#:~:text=December%2031%2C%202024%2C%20reporting%20a,engagement%20and%20willingness%20to%20pay">SenseTime Reports Revenue Growth and AI Advancements in 2024 — TipRanks.com</a>), and it is pushing AI adoption in productivity and interactive tools in China. SenseTime’s stock, however, has languished — currently around HK$1.50 per share, it’s down roughly 60% from its IPO price (HK$3.85 in 2021) and trades at a fraction of U.S. AI peers’ valuations. Analysts expect <strong>27% annual revenue growth</strong> for SenseTime over the next 3 years, above the industry average (<a href="https://simplywall.st/stocks/hk/software/hkg-20/sensetime-group-shares/news/sensetime-group-first-half-2024-earnings-misses-expectations#:~:text=SenseTime%20Group%20First%20Half%202024,forecast%20for%20the%20Software">SenseTime Group First Half 2024 Earnings: Misses Expectations</a>), as China accelerates national AI initiatives. With a market cap near US$7 billion, SenseTime is valued at about 7× sales (much lower than comparable U.S. AI firms). The catch: it’s still deeply unprofitable and has been impacted by U.S. export restrictions (on GPU chips, due to its alleged involvement in surveillance). For those willing to take on geopolitical and regulatory risk, SenseTime offers <strong>high-risk exposure to AI infrastructure in Asia</strong>, potentially undervalued if it maintains its technological edge in a huge domestic market.</li></ul><p><em>(Japanese market note: Pure AI software plays are scarce in Japan’s public markets. Large IT/electronics firms like </em><strong><em>Fujitsu</em></strong><em>, </em><strong><em>NEC</em></strong><em>, and </em><strong><em>SoftBank</em></strong><em> are investing in AI (e.g. Fujitsu offers AI platforms and HPC for AI), and they trade at lower multiples than U.S. tech giants. However, their AI segments are just part of broad businesses. Risk-tolerant investors might watch for any AI-focused spinoffs or smaller firms on the TSE. For now, the clearest international “AI infrastructure” stock is SenseTime in HK.)</em></p><h3>Comparison of Key Metrics and Valuations</h3><p>The following table summarizes key financial metrics, recent stock performance, valuation multiples, and strategic roles for the highlighted companies:</p><p>Company (Ticker) Market Cap (USD) 2024 Revenue Growth (YoY) Recent Stock Performance (Approx. 1-Yr) Valuation Multiples Strategic Position in AI Ecosystem <strong>Palantir (PLTR)</strong> ~$208B (<a href="https://public.com/stocks/pltr/market-cap#:~:text=As%20of%20Apr%2010%2C%202025%2C,while%20in">Palantir Technologies (PLTR) Market Cap Today</a>) +18% (est. 2025E rev. growth ~25%) <strong>+590%</strong> (YoY as of Feb 2025) (<a href="https://www.indmoney.com/blog/us-stocks/palantir-technologies-m-cap-rise-how-it-became-one-of-world-s-most-valuable-ai-firms-in-a-year#:~:text=Palantir%27s%20594%25%20m,year%20rise">Palantir’s 594% m-cap surge: How it became one of the world’s most …</a>) ~55× P/S; 160× forward P/E (<a href="https://finance.yahoo.com/quote/PLTR/#:~:text=Palantir%20Technologies%20Inc,Price%2FSales%20%28ttm">Palantir Technologies Inc. (PLTR) Stock Price, News, Quote &amp; History</a>) End-to-end AI/ML software platform for gov’t &amp; enterprise; leverages big data and AI, poised to gain from surging federal AI spending (<a href="https://www.investopedia.com/wedbush-calls-palantir-a-top-stock-to-own-in-2025-11689615#:~:text=However%2C%20Wedbush%20analysts%20told%20clients,government%20agencies%2C%E2%80%9D%20the%20analysts%20added">Wedbush Calls Palantir a Top Stock to Own in 2025</a>). <strong>C3.ai (AI)</strong> ~$2.7B (<a href="https://stockanalysis.com/stocks/ai/market-cap/#:~:text=Revenue,17.85%25%20in%20one%20year">C3.ai (AI) Market Cap &amp; Net Worth — Stock Analysis</a>) +26% (FY25 Q3 YoY) (<a href="https://stockanalysis.com/stocks/ai/revenue/#:~:text=C3.ai%20%28AI%29%20Revenue%202019,92M">C3.ai (AI) Revenue 2019–2025 — Stock Analysis</a>) <strong>-18%</strong> (YoY cap decline) (<a href="https://stockanalysis.com/stocks/ai/market-cap/#:~:text=C3,17.85%25%20in%20one%20year">C3.ai (AI) Market Cap &amp; Net Worth — Stock Analysis</a>) ~11× P/S ([This Artificial Intelligence (AI) Stock Is Already Down 10% in 2025, but Wall Street Analysts Believe That It Could Soar Impressively Nasdaq](<a href="https://www.nasdaq.com/articles/artificial-intelligence-ai-stock-already-down-10-2025-wall-street-analysts-believe-it#:~:text=The%20stock%27s%20valuation%20and%20potential,could%20be%20a%20solid%20investment">https://www.nasdaq.com/articles/artificial-intelligence-ai-stock-already-down-10-2025-wall-street-analysts-believe-it#:~:text=The%20stock%27s%20valuation%20and%20potential,could%20be%20a%20solid%20investment</a>)) (no P/E, not profitable) <strong>Nebius (NBIS)</strong> ~$5B (<a href="https://finance.yahoo.com/quote/NBIS/#:~:text=History%20finance,Financial">Nebius Group N.V. (NBIS) Stock Price, News, Quote &amp; History</a>) <strong>+462%</strong> (2024 revenue) (<a href="https://seekingalpha.com/article/4772706-nebius-group-trades-at-a-steep-discount#:~:text=Nebius%20Group%20Trades%20At%20A,Learn%20why%20NBIS%20stock">Nebius Group Trades At A Steep Discount — Seeking Alpha</a>) n/a (IPO Oct 2024; <em>-19%</em> YTD ([NBIS Stock: Nebius Is an Under-the-Radar AI Play CMC Markets](<a href="https://www.cmcmarkets.com/en/optox/nbis-stock-nebius-is-an-under-the-radar-ai-play#:~:text=NBIS%20Stock%3A%20Nebius%20Is%20an,in%20October%2C%20but%20has">https://www.cmcmarkets.com/en/optox/nbis-stock-nebius-is-an-under-the-radar-ai-play#:~:text=NBIS%20Stock%3A%20Nebius%20Is%20an,in%20October%2C%20but%20has</a>))) ~2.9× EV/ARR (<a href="https://seekingalpha.com/article/4772706-nebius-group-trades-at-a-steep-discount#:~:text=Nebius%20Group%20Trades%20At%20A,Learn%20why%20NBIS%20stock">Nebius Group Trades At A Steep Discount — Seeking Alpha</a>) (EV ~$2.7B) <strong>MongoDB (MDB)</strong> ~$13–14B (<a href="https://stockanalysis.com/stocks/mdb/market-cap/#:~:text=MongoDB%20has%20a%20market%20cap,17B">MongoDB (MDB) Market Cap &amp; Net Worth — Stock Analysis</a>) +19% (FY2025 rev) (<a href="https://stockanalysis.com/stocks/mdb/revenue/#:~:text=MongoDB%20%28MDB%29%20Revenue%202016,in%20the%20quarter%20ending">MongoDB (MDB) Revenue 2016–2025 — Stock Analysis</a>) <strong>-50%</strong> (YoY market cap) (<a href="https://stockanalysis.com/stocks/mdb/market-cap/#:~:text=MongoDB%20,17B">MongoDB (MDB) Market Cap &amp; Net Worth — Stock Analysis</a>) ~6–7× P/S (TTM) Leading NoSQL database with new vector search features (<a href="https://www.mongodb.com/en-us/products/platform/atlas-vector-search#:~:text=Vector%20Search,featured%20vector%20database%20capabilities">Atlas Vector Search — MongoDB</a>); core infrastructure for AI apps (unstructured data storage), stock depressed after hype unwound. <strong>Snowflake (SNOW)</strong> ~$48B (<a href="https://finance.yahoo.com/quote/SNOW/#:~:text=Performance%20Overview%3A%20SNOW,Enterprise%20Value">Snowflake Inc. (SNOW) Stock Price, News, Quote &amp; History</a>) +28% (FQ4 2025 YoY) (<a href="https://www.marketbeat.com/originals/snowflake-boosts-growth-by-doubling-down-on-ai/#:~:text=Snowflake%20Boosts%20Growth%20by%20Doubling,FQ1%202025">Snowflake Boosts Growth by Doubling Down on AI — MarketBeat</a>) ~Flat (roughly 0% to -5% YoY) ~12–14× P/S (forward) Cloud data warehouse for AI/analytics; integral for enterprise AI data pipelines, high net retention (126%) (<a href="https://investors.snowflake.com/financials/quarterly-results/default.aspx#:~:text=Relations%20investors,580%20customers%20with%20trailing">Financials — Quarterly Results — Snowflake — Investor Relations</a>), valuation down from peak. <strong>GitLab (GTLB)</strong> ~$6.9B (<a href="https://finance.yahoo.com/quote/GTLB/#:~:text=Mkt%20Cap%2019,87B">GitLab Inc. (GTLB) Stock Price, News, Quote &amp; History</a>) +29% (FY25 Q4 YoY) (<a href="https://www.tipranks.com/news/company-announcements/gitlab-reports-strong-growth-in-fiscal-year-2025#:~:text=GitLab%20Reports%20Strong%20Growth%20in,million%2C%20alongside%20a%20notable">GitLab Reports Strong Growth in Fiscal Year 2025 — TipRanks.com</a>) <strong>+20%</strong> (last 6 months) (<a href="https://www.nasdaq.com/articles/gtlb-stock-rises-21-6-months-will-rally-continue-2025#:~:text=GTLB%20Stock%20Rises%2021,8">GTLB Stock Rises 21% in 6 Months: Will the Rally Continue in 2025?</a>) ~9× P/S (forward) (<a href="https://www.itiger.com/news/2503613190#:~:text=Valuation%20and%20verdict">SoundHound vs. GitLab: Better Mid-Cap AI Stock Investment for 2025 — Tiger Brokers</a>) DevOps platform with AI-assisted coding and dev security (<a href="https://www.itiger.com/news/2503613190#:~:text=GitLab%2C%20on%20the%20other%20hand%2C,offering%20suggestions%20and%20through%20automation">SoundHound vs. GitLab: Better Mid-Cap AI Stock Investment for 2025 — Tiger Brokers</a>) (<a href="https://www.itiger.com/news/2503613190#:~:text=When%20it%20comes%20to%20future,also%20been%20gaining%20strong%20traction">SoundHound vs. GitLab: Better Mid-Cap AI Stock Investment for 2025 — Tiger Brokers</a>); high gross margins (89%) (<a href="https://www.itiger.com/news/2503613190#:~:text=While%20SoundHound%20has%20been%20growing,gap%20is%20currently%20very%20wide">SoundHound vs. GitLab: Better Mid-Cap AI Stock Investment for 2025 — Tiger Brokers</a>), benefiting from software demand and AI upsells. <strong>Progress Software (PRGS)</strong> ~$2.5B (<a href="https://stockanalysis.com/stocks/prgs/market-cap/#:~:text=Progress%20Software%20,Full%20Definition">Progress Software (PRGS) Market Cap &amp; Net Worth — Stock Analysis</a>) +29% (Q1 2025 YoY) ([Progress Software: Is the Stock Still Undervalued After Its AI-Powered Q1 Beat? Investing.com](<a href="https://www.investing.com/analysis/progress-software-making-progress-driven-by-the-ai-revolution-200659234#:~:text=Progress%20Software%E2%80%99s%20stock%20had%20reached,62%20million">https://www.investing.com/analysis/progress-software-making-progress-driven-by-the-ai-revolution-200659234#:~:text=Progress%20Software%E2%80%99s%20stock%20had%20reached,62%20million</a>)) ~Flat (off highs; rebounded after Q1 beat) ~10× forward P/E ([Progress Software: Is the Stock Still Undervalued After Its AI-Powered Q1 Beat? <strong>SoundHound AI (SOUN)</strong> ~$3.3–4.3B (<a href="https://finance.yahoo.com/quote/SOUN/key-statistics/#:~:text=,72B">SoundHound AI, Inc. (SOUN) Valuation Measures &amp; Financial …</a>) +85% (2024 rev) (<a href="https://www.investors.com/news/technology/soun-stock-soundhound-ai-q4-2024-earnings/#:~:text=Targets%20www,to%20%2484.7%20million">SoundHound Stock Howls After Voice AI Firm Tops Q4 Targets</a>) -&gt; <strong>+97%</strong> (2025E) (<a href="https://www.investors.com/news/technology/soun-stock-soundhound-ai-q4-2024-earnings/#:~:text=Targets%20www,to%20%2484.7%20million">SoundHound Stock Howls After Voice AI Firm Tops Q4 Targets</a>) <strong>≈+400%</strong> (YoY; extremely volatile) (<a href="https://www.nasdaq.com/articles/soundhound-ai-stock-1051-2024-where-could-it-be-end-2025#:~:text=SoundHound%20AI%20Stock%20Is%20Up,Based%20on%20its%20trailing">SoundHound AI Stock Is Up 1,051% in 2024. Where Could It Be at …</a>) ~20–25× P/S (forward) Voice AI platform (speech-to-meaning tech) for cars, IoT, etc. (<a href="https://www.itiger.com/news/2503613190#:~:text=SoundHound%20operates%20a%20voice%20AI,insurance%2C%20retail%2C%20and%20banking%20spaces">SoundHound vs. GitLab: Better Mid-Cap AI Stock Investment for 2025 — Tiger Brokers</a>); small revenue base but doubling growth, aiming to be a top voice AI provider (high risk/high reward). <strong>SentinelOne (S)</strong> ~$5.8B (<a href="https://pitchbook.com/profiles/company/58281-31#:~:text=Earnings%20pitchbook,58">SentinelOne 2025 Company Profile: Stock Performance &amp; Earnings</a>) +32% (FY2025 rev) (<a href="https://stockanalysis.com/stocks/s/revenue/#:~:text=SentinelOne%20%28S%29%20Revenue%202020,52M%20in%20the%20quarter">SentinelOne (S) Revenue 2020–2025 — Stock Analysis</a>) <strong>-18%</strong> (YoY) (<a href="https://stockanalysis.com/stocks/s/market-cap/#:~:text=SentinelOne%20,80B">SentinelOne (S) Market Cap &amp; Net Worth — Stock Analysis</a>) ~7× P/S (TTM) AI-driven cybersecurity (endpoint &amp; cloud) with $920M ARR (<a href="https://investors.sentinelone.com/press-releases/news-details/2025/SentinelOne-Announces-Fourth-Quarter-and-Fiscal-Year-2025-Financial-Results/default.aspx#:~:text=,1%20million%20as">SentinelOne Announces Fourth Quarter and Fiscal Year 2025 …</a>); trading at discount to rival CrowdStrike, potential rerating as AI security demand grows. <strong>SenseTime (0020.HK)</strong> ~US$6–7B (HK$50.9B) ([SenseTime Group Market capitalization 2025 KYG8062L1041 20 …](<a href="https://eulerpool.com/en/stock/SenseTime-Group-Stock-KYG8062L1041/Marketcapitalization#:~:text=SenseTime%20Group%20Market%20capitalization%202025,increase%20from%20the">https://eulerpool.com/en/stock/SenseTime-Group-Stock-KYG8062L1041/Marketcapitalization#:~:text=SenseTime%20Group%20Market%20capitalization%202025,increase%20from%20the</a>)) +10.8% (2024 rev) (<a href="https://www.tipranks.com/news/company-announcements/sensetime-reports-revenue-growth-and-ai-advancements-in-2024#:~:text=SenseTime%20Group%2C%20Inc,engagement%20and%20willingness%20to%20pay">SenseTime Reports Revenue Growth and AI Advancements in 2024 — TipRanks.com</a>) <strong>-60%</strong> (from 2021 IPO price)</p><p><strong>Sources:</strong> Company filings and press releases; financial data from Yahoo Finance, Macrotrends, and StockAnalysis; news and analysis from Motley Fool, Investor’s Business Daily, Nasdaq, and others (<a href="https://public.com/stocks/pltr/market-cap#:~:text=As%20of%20Apr%2010%2C%202025%2C,while%20in">Palantir Technologies (PLTR) Market Cap Today</a>) (<a href="https://stockanalysis.com/stocks/ai/revenue/#:~:text=C3.ai%20%28AI%29%20Revenue%202019,92M">C3.ai (AI) Revenue 2019–2025 — Stock Analysis</a>) (<a href="https://stockanalysis.com/stocks/mdb/market-cap/#:~:text=MongoDB%20,17B">MongoDB (MDB) Market Cap &amp; Net Worth — Stock Analysis</a>) (<a href="https://www.marketbeat.com/originals/snowflake-boosts-growth-by-doubling-down-on-ai/#:~:text=Snowflake%20Boosts%20Growth%20by%20Doubling,FQ1%202025">Snowflake Boosts Growth by Doubling Down on AI — MarketBeat</a>) (<a href="https://www.tipranks.com/news/company-announcements/gitlab-reports-strong-growth-in-fiscal-year-2025#:~:text=GitLab%20Reports%20Strong%20Growth%20in,million%2C%20alongside%20a%20notable">GitLab Reports Strong Growth in Fiscal Year 2025 — TipRanks.com</a>) (<a href="https://www.investing.com/analysis/progress-software-making-progress-driven-by-the-ai-revolution-200659234#:~:text=Progress%20Software%E2%80%99s%20stock%20had%20reached,62%20million">Progress Software: Is the Stock Still Undervalued After Its AI-Powered Q1 Beat? | Investing.com</a>) (<a href="https://www.investors.com/news/technology/soun-stock-soundhound-ai-q4-2024-earnings/#:~:text=Targets%20www,to%20%2484.7%20million">SoundHound Stock Howls After Voice AI Firm Tops Q4 Targets</a>) (<a href="https://stockanalysis.com/stocks/s/revenue/#:~:text=SentinelOne%20%28S%29%20Revenue%202020,52M%20in%20the%20quarter">SentinelOne (S) Revenue 2020–2025 — Stock Analysis</a>) (<a href="https://www.tipranks.com/news/company-announcements/sensetime-reports-revenue-growth-and-ai-advancements-in-2024#:~:text=SenseTime%20Group%2C%20Inc,engagement%20and%20willingness%20to%20pay">SenseTime Reports Revenue Growth and AI Advancements in 2024 — TipRanks.com</a>).</p><h3>Outlook and Strategic Positioning</h3><p>In summary, while AI hardware companies like NVIDIA have captured most of the market’s attention (and rich valuations), the <strong>infrastructure software layer of AI offers a breadth of investment opportunities</strong> that may be undervalued by comparison. The companies profiled here span a broad range of AI enablers — from data platforms and model deployment tools to communication APIs and security — reflecting the reality that <strong>AI is not one monolithic market, but an ecosystem</strong>.</p><p>Several themes emerge from this analysis:</p><ul><li><strong>Resilient Growth, Reasonable Multiples:</strong> Many of these software firms are still growing revenues at double-digit (even double or triple-digit) rates, yet their valuation multiples (P/S, EV/Revenue) have contracted significantly from peak levels. This presents a contrast to the hardware side where some valuations expanded. For example, MongoDB and Elastic saw their stocks halve even as they continue to grow ~15–20% and pivot toward AI use cases (<a href="https://stockanalysis.com/stocks/mdb/market-cap/#:~:text=MongoDB%20,17B">MongoDB (MDB) Market Cap &amp; Net Worth — Stock Analysis</a>) (<a href="https://seekingalpha.com/article/4774044-elastic-rating-upgrade-on-aicloud-momentum-and-better-leading-growth-metrics#:~:text=Elastic%20,2x%20forward%20revenue">Elastic (ESTC): Rating Upgrade On AI/Cloud Momentum And Better …</a>). This suggests the market may have swung from over-exuberance to over-skepticism, creating potential <em>bargains</em> for long-term investors.</li><li><strong>Critical Role in AI Value Chain:</strong> Each of these companies occupies a <strong>strategic link in the AI value chain</strong>. Whether it’s Palantir’s end-to-end AI application platform for enterprises, Snowflake’s data cloud that feeds AI models, or GitLab’s tools that streamline AI model development, these are enabling technologies. As businesses large and small race to adopt AI, they will need robust infrastructure — to store huge datasets, train models, deploy them into products, and secure them. The selected companies are well-placed in their niches (e.g. Confluent for real-time data streaming, C3.ai for turnkey AI apps, SoundHound for voice interfaces). If the AI tide keeps rising, these “picks and shovels” providers could see <strong>sustained demand growth</strong>, even if they lack the headline-grabbing appeal of a chipmaker or a consumer AI app.</li><li><strong>Recent Catalysts and Positive Developments:</strong> Many of the firms have had encouraging recent news:</li><li><em>Earnings Beats/Guidance:</em> Progress Software’s big beat and raised guidance showed investors that AI integration can boost even a legacy software business (<a href="https://www.investing.com/analysis/progress-software-making-progress-driven-by-the-ai-revolution-200659234#:~:text=Progress%20Software%E2%80%99s%20stock%20had%20reached,62%20million">Progress Software: Is the Stock Still Undervalued After Its AI-Powered Q1 Beat? | Investing.com</a>) (<a href="https://www.investing.com/analysis/progress-software-making-progress-driven-by-the-ai-revolution-200659234#:~:text=Progress%20Software%20issued%20upside%20FQ2,82%20million">Progress Software: Is the Stock Still Undervalued After Its AI-Powered Q1 Beat? | Investing.com</a>). GitLab and Elastic also delivered better results with an AI narrative (Elastic signing $1M+ deals for generative AI search (<a href="https://www.tipranks.com/news/company-announcements/elastic-n-v-reports-strong-growth-and-ai-momentum#:~:text=Elastic%20N,The%20company%20now">Elastic N.V. Reports Strong Growth and AI Momentum — TipRanks.com</a>)).</li><li><em>Strategic Alliances:</em> C3.ai’s new alliances with Microsoft and others (<a href="https://www.nasdaq.com/articles/artificial-intelligence-ai-stock-already-down-10-2025-wall-street-analysts-believe-it#:~:text=C3,of%20its%20generative%20AI%20offerings">This Artificial Intelligence (AI) Stock Is Already Down 10% in 2025, but Wall Street Analysts Believe That It Could Soar Impressively | Nasdaq</a>), Palantir’s deepened ties with governments, and Nebius’s likely partnerships (perhaps with ex-Yandex clients or emerging market tech firms) all point to expanding ecosystems.</li><li><em>Product Innovations:</em> Many launched AI-specific features — e.g., MongoDB’s vector search, GitLab’s AI pair programmer, Twilio’s AI customer engagement tools, and SenseTime’s record-setting SenseNova model (<a href="https://www.tipranks.com/news/company-announcements/sensetime-reports-revenue-growth-and-ai-advancements-in-2024#:~:text=December%2031%2C%202024%2C%20reporting%20a,engagement%20and%20willingness%20to%20pay">SenseTime Reports Revenue Growth and AI Advancements in 2024 — TipRanks.com</a>). These innovations keep them relevant as AI tech evolves, and can open up new revenue streams or upsell opportunities.</li><li><strong>Risk Factors:</strong> It’s important to note that “undervalued” doesn’t mean “without risk.” Common risks include <strong>competition from Big Tech</strong> (e.g. cloud giants encroaching on smaller software providers, or open-source projects reducing demand for paid solutions), execution risk in converting AI hype into sales, and for some, weak profitability or cash burn that needs to be managed (SoundHound, C3.ai, etc., rely on capital markets’ patience). Additionally, the <strong>macro environment</strong> — higher interest rates, IT budget tightening — can affect these stocks, as seen by declines in 2022–23. International plays like SenseTime carry geopolitical and regulatory risks that are higher than U.S. peers. Investors should therefore <strong>pick according to their risk tolerance</strong>: a basket of these names could be balanced between profitable stalwarts (e.g. Progress, Fortinet) and moonshots (e.g. SoundHound, Nebius) to mitigate risk.</li></ul><h3>Looking ahead</h3><p>The AI infrastructure software segment is likely to gain more recognition. As organizations move from experimenting with AI to operationalizing it at scale, they will invest in the “boring” back-end software — databases, MLOps, security, and so on. We are already seeing investor sentiment slowly shift; for instance, some of these stocks have started to rebound off their lows as earnings come in strong or as analysts flag their potential (Palantir’s rapid ascent and Wedbush’s bullish call (<a href="https://www.investopedia.com/wedbush-calls-palantir-a-top-stock-to-own-in-2025-11689615#:~:text=However%2C%20Wedbush%20analysts%20told%20clients,government%20agencies%2C%E2%80%9D%20the%20analysts%20added">Wedbush Calls Palantir a Top Stock to Own in 2025</a>), or Motley Fool dubbing Nebius a hidden AI gem). There may also be <strong>M&amp;A activity</strong>: larger tech or industrial players could acquire smaller infrastructure specialists to bolster their AI capabilities (IBM’s purchase of HashiCorp is a case in point (<a href="https://erp.today/ibm-acquires-hashicorp-enhancing-hybrid-cloud-and-ai-capabilities/#:~:text=IBM%20Acquires%20HashiCorp%2C%20Enhancing%20Hybrid,AI%20capabilities%20through%20the">IBM Acquires HashiCorp, Enhancing Hybrid Cloud and AI Capabilities</a>)).</p><h3>In conclusion</h3><p>savvy investors would do well to look beyond the flashy AI chipmakers and consider the software backbone of the AI revolution. The companies highlighted in this report offer exposure to various layers of AI infrastructure — often with <strong>lower valuations and less fanfare</strong> than their hardware counterparts, but with equally crucial roles to play. By combining due diligence on fundamentals with an understanding of each company’s niche in the AI ecosystem, one can identify which of these stocks align best with their portfolio strategy (be it growth, value, or speculative upside). As always, diversification is key; not all will be winners, but those that execute successfully could deliver substantial returns if the AI wave continues. With AI still in its early innings across the economy, the infrastructure software makers powering this shift are poised to share in the growth — and at today’s prices, they just might be <strong>diamonds in the rough</strong> in an otherwise crowded AI investment landscape.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b5b2a3b9ac43" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[차익거래자의 관점 VS 추세추종자의 관점]]></title>
            <link>https://bellman-h.medium.com/%EC%B0%A8%EC%9D%B5%EA%B1%B0%EB%9E%98%EC%9E%90%EC%9D%98-%EA%B4%80%EC%A0%90-vs-%EC%B6%94%EC%84%B8%EC%B6%94%EC%A2%85%EC%9E%90%EC%9D%98-%EA%B4%80%EC%A0%90-088b21c630d8?source=rss-99e8e6896144------2</link>
            <guid isPermaLink="false">https://medium.com/p/088b21c630d8</guid>
            <category><![CDATA[trading-ideas]]></category>
            <category><![CDATA[statistics]]></category>
            <category><![CDATA[arbitrage]]></category>
            <category><![CDATA[trading]]></category>
            <category><![CDATA[derivatives]]></category>
            <dc:creator><![CDATA[bellman]]></dc:creator>
            <pubDate>Sun, 09 Feb 2025 23:57:05 GMT</pubDate>
            <atom:updated>2025-02-09T23:58:48.733Z</atom:updated>
            <content:encoded><![CDATA[<h3>차익거래자의 관점 vs 추세추종자의 관점</h3><p>나는 요즘 트레이딩 스타일에 변화를 주고 있다. 주고 있다는 말보다는 겪고 있다(?) 라는 표현이 조금 더 적절한지는 모르겠다.</p><p>원래 근본주의적인 추세추종 매매를 지향했고 결과 또한 나쁘지 않았으나, 요즘 하는 연구와 더불어, 그리고 나 개인의 포지션도 많이 바뀌어서 스케일러블하고 자주 튜닝 필요없는 전략보다는 절대수익, 낮은 카파, 많은 튜닝 필요한 전략이 필요하게 되었다.</p><p>그래서 요즘은 차익거래와 통계적 차익거래를 많이 보고 있다. 퀀트방에서 우리 멤버들과 이야기를 나누면 차익거래가 무엇이라고 정의되느냐, 통계적 차익거래는 그럼 무엇이라고 정의되느냐, 추세추종/retraction (평균 회귀) 전략은 그럼 다른 범주인가 하는 논란(?) 이 팽배하다. 이 주제로 오프라인에서 수 시간 동안 토론한 적도…</p><p>추세추종 매매야 완벽히 체득하는데야 10년쯤 더 걸리겠으나 이론은 정말 충분할 정도로 알고 있고, 차익거래/통계적차익거래는 걸음마 단계의 수익 모델을 배포해서 수익이 나는 정도의 상태이니 내가 감히 정의까지는 아니고 TA/추세 기반 시스템 트레이딩과 Arb/Stat-Arb 기반 트레이딩의 주요한 관점 차이 하나는 이야기 할 수 있을 것 같다.</p><p>바로 Fair Price (Spread) 모델의 여/부인것 같다. Fair Price (Spread) 모델이란 시장에서 상대적 적정가치를 측정하는 모델이라고 정의해보자. 모델의 아웃풋은 꼭 Price로 나오지 않아도 된다.</p><p>추세추종 혹은 되돌림을 거래하는 전략들을 생각해 보자. 결국 되돌림을 거래하는 전략들도 그냥 무조건 내린다고 지르는 건 자살행위고 다시 추세로 회귀한다는 관점에서 지르는 거니까, 사실은 Fair Price라는 개념이 잘 없고 있더라도 Fair Price가 많이 움직일 때 어디까지 움직일진 모르겠지만 자금관리 잘하면서 따라가면서 손익비가 좋은 매매를 하겠다는 이야기에</p><p>가깝다.</p><p>매매의 특성상 차익거래/통계적 차익거래자에게 물량을 사오는 경우가 꽤 많지 않을까? 라는 생각이 든다. 그런 매매는 보통 휩쏘나 거래비용 높은 매매로 마킹되는 것이고… 대신 블랙스완 이벤트에 엄청 높은 손익비로 다 잡아먹는다.</p><p>반대급부로 상대적이든 절대적이든 그 적정가치가 존재하고 그 적정가치에 가격이 맞추어 지게 되어 있다는 관점을 가지는 전략들이. Arb/Stat-Arb 전략인 것이다.</p><h3>동일 자산 간 차익거래</h3><p>이를테면 가장 간단한 거래소간 차익거래는 서로 다른 거래소에 상장된 같은 상품 사이의 가격 스프레드는 수렴한다 (0이 해당 스프레드의 Fair Price) 는 관점이다. 혹은 양방향으로 차익거래를 하지 않더라도 거래소 간 가격반영의 Lagging을 이용한 거래 또한 유명한 방법이다.</p><h3>이종 자산 간 차익거래</h3><p>조금 위험을 동반하는 Pairs Trading의 경우 두 자산 사이의 관계를 모델링하여 상관관계가 높은 경우에 해당 자산 사이의 Spread는 장기적으로 어떤 포인트에 수렴한다는 관점을 가지고 거래한다. 이걸 커스터마이즈한 포트폴리오에 대한 페어로 만들어 확률을 더 높이는 작업도 가능하다.</p><h3>단일 종목 안에서 Price Action과 현재 Quote 간 차익거래</h3><p>요즘 가장 재미있는 연구 포인트는, 어찌 보면 당연한 말이지만 굳이 이종자산이나 거래소를 끼지 않고 하나의 거래소에서 하나의 자산의 방향성 거래도 (추세추종에 절여져서 아직은 이런 방식의 거래가 익숙하다)</p><p>이를테면 거래량 – 변동성 사이의 Fair Price 모델을 세운다거나, taker volume * 부호의 누적합과 가격 변화 사이의 Fair Price 모델을 세운다거나 하는 것이다. 결국 Price를 만드는 Action의 누적으로 만든 Fair Price가 현재 Quote과 차이가 나면 차익거래 기회로 생각할 수 있다는 것이다.</p><p>이는 마치 예전 세력주 국장고수(?) 들이 이야기하는 거래량 높지만 가격변화는 그렇게 안 일어나는 매집봉에 집중하라! 이런 말과 좀 비슷한 것 같기도 하고.</p><p>물론 그냥 쌩으로 베타노출을 가져가고 보는 류의 전략이기 때문에 나이브한 모델로는 어렵고 여기에 추세추종 필터 같은 걸 넣으면 꽤 볼만한 전략이 나온다. 그래도 베타를 가져가는 전략인 만큼 Capa / 수익률 측면에서 위에 언급한 통상적인 전략(?) 들보다는 다소 유리하다고 생각을 한다. 항상 하나를 내주면 하나는 가져올 수 있는 것이다.</p><h3>옵션 시장에서의 차익거래</h3><p>옵션 시장으로 가면 이런 류의 거래 기회가 훨씬 유연하고 많다. 물론 이쪽은 Execution이 거의 지옥에 가깝고 머릿속에서 그릭들 생각하는게 생각보다 복잡하다는 이슈는 있지만 말이다. 예를 들면 Skewness (풋옵션의 내재변동성과 콜옵션의 내재변동성 분포 차이) 차익거래, 행사가격별 내재변동성 분포가 정상수준에서 벗어났을 때의 행사가격 차익거래 등이다. 이것도 일종의 상대적 Fair Price를 정의하고 그것에 맞추는 거래인 것이다.</p><h3>마치며</h3><p>아무튼 오늘은 내 트레이딩 지평을 좀 확장하면서 드는 생각들을 좀 이야기해봤다. 트레이딩</p><p>철학과 스타일도 일종의 철학과 종교라고 생각하는데 그렇게 생각하면 요즘은 일종의 개종 중이려나 싶다.</p><p>생각보다 한번 어떤 스타일이 머리와 손에 익어버리면 그 프레임워크에서 벗어나는 것이 어떤 큰 이벤트가 있지 않는 이상 사람이 게으르고 관성적인지라 어려운 것 같다. 그래도 몇 단계 점프할 기회라는 생각이 본능적으로 들어서 열심히 해보는 중이다. 글이 비슷한 길을 가시는 분들의 도움이 되었으면 한다</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=088b21c630d8" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[그 많은 개미는 어디갔을까 (Feat. 0DTE, Weekly Options)]]></title>
            <link>https://bellman-h.medium.com/%EA%B7%B8-%EB%A7%8E%EC%9D%80-%EA%B0%9C%EB%AF%B8%EB%8A%94-%EC%96%B4%EB%94%94%EA%B0%94%EC%9D%84%EA%B9%8C-feat-0dte-weekly-options-63968b372ecc?source=rss-99e8e6896144------2</link>
            <guid isPermaLink="false">https://medium.com/p/63968b372ecc</guid>
            <category><![CDATA[0dte]]></category>
            <category><![CDATA[options]]></category>
            <category><![CDATA[crypto]]></category>
            <category><![CDATA[hft]]></category>
            <dc:creator><![CDATA[bellman]]></dc:creator>
            <pubDate>Mon, 23 Sep 2024 14:57:29 GMT</pubDate>
            <atom:updated>2024-09-23T15:00:27.340Z</atom:updated>
            <content:encoded><![CDATA[<h3>고객님, 어디가세요?</h3><p>퀀트 트레이딩을 하는 우리는 무엇을 Exploit 하고 있는가? 라는 질문에 응당 대답할 수 있어야 한다. 어떤 시장의 구조를 활용(악용) 하여서 돈을 벌고 있는지 대답이 안되는 전략은 잘 되기 어렵고, 잘 되더라도 오래 유지하기란 아주 어려운 전략이 될 것이다.</p><p>시장을 너무 뭉뚱그려서 생각할 수도 있는 것이겠지만 본질적으로 알파는 HFT에서 흔하게 이야기하는 Noise Trader, 즉 Wikipedia의 정의에 따르면,</p><blockquote>“그들이 도움이 된다고 생각하지만 실제로는 무작위 선택보다 더 나은 수익을 제공하지 않을 요소”를 기반으로 매수 또는 매도 결정을 내리는 주식 거래자</blockquote><p>들에게서부터 시장 초과수익률이 창출된다. 그들은 추세 추종자들이 먹는 추세도 만들고, Cross-Sectional 트레이딩이 주로 Exploit하는 종목간의 불균형적인 상대강도도 만들고, 마켓메이커들이 먹을 스프레드도 만들고…</p><p>물론 자본주의 체계 아래 GDP는 항상 오르는 것이고 금융시장은 전반적인 생산성을 더욱 올려주는 플러스썸이지만 베팅해서 나는 따고 너는 잃고의 관점에서 보자면 (실제로 Price Discovery 역할을 크게 하는 파생상품시장은 실제로 이런 구조이기도 하고) 퀀트트레이더들은 카지노 운영자이고 Noise Trader들은 카지노 고객에 가깝다.</p><p>그런데, 카지노 운영자들도 서로 경쟁하기 때문에 분명 망할 수 있다. 실제로 카지노도 망한다. 그래서 시장에 충분한 Noise Trader들이 없으면, 백날 리서치를 해봐야 살기가 팍팍한 것이다.</p><p><strong>고객이 발을 끊은 카지노에 아무리 재미있는 룰렛을 들여놓은듯 의미가 있겠는가?</strong></p><h4>요즘 크립토 시장, 국내주식시장 등등에 개미가 없다… (반쯤 데이터, 반쯤 느낌적 느낌)</h4><p><em>필자가 요즘 갑자기 이런 뜬구름잡는 생각을 하게 된 데에는 절대 요즘 전략이 재미없어서가 아니고…</em></p><p>주변에서 비슷한 업 하시는 분들이 힘들다는 소리를 많이 하셔서였다. 크립토 전략 모집해서 파는 브로커들도 요즘 성과좋은 전략이 많이 없다며… Cross-Sectional 하시는 분들 몇분 여쭤보면 예전에는 메이저 알트에 회전율 하루 몇백%도 거뜬했는데 요즘 그렇게 하면 슬리피지도 그렇고 턴오버당 수익률이 많이 낮아져서 어렵다고 한다.</p><p>물론, 이런 시장에서도 잘하시는 분들은 있겠으나 뭔가 힘들어진 것 같기는 하다. 눈에 붙이면 안경 코에 붙이면 코걸이지만 몇가지 데이터로도 뒷받침 할 수 있다.</p><p>예를 들면 BTC Volatility Index (BTC 옵션 내재변동성) 이 많이 줄었고 BTC의 실현변동성 또한 줄어들었다거나.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6qqIQvwOYfUdVm_cf1umlA.png" /></figure><p>$BTC는 신고가를 주도했지만, 이전에 장을 주도했던 메이저 알트 (?) 중 일부가 낙오되어 고점을 찍지 못하고 있으며 거래량까지 떨어졌다거나</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*l50tcxw0oXOSLZjWmo2wSg.png" /><figcaption>애증의 리플</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*w4_0t6aKnXoBkK9by-5e-A.png" /><figcaption>그리고 에이다</figcaption></figure><p>아마 알트를 주로 거래하시는 분들은 이런 메이저 알트들에 대해 많이 분석하고 비중을 실었을텐데 뭔가 추세도 힘도 거래량도 없어보이는 요즘이다. 카지노에서 고객님들이 떠났다.</p><h4>그 거래량은… 과연 매출이 나는 거래량일까?</h4><p>최근 바이낸스에서 Fee Tier 관련 정책이 새로 생겼다. 예전에는 Subaccount, Managed Sub Account (일종의 일임계좌) 등의 거래량을 전부 합쳐서 VIP티어를 산정해주다가, 이제는 각 계정별로 Fee Tier를 산정해주겠다는 이야기를 한 것이다.</p><p>이것의 결정 배경은 VIP 9 티어를 받은 다음 자기네들 오더북 없이 거래 수수료 장사를 해먹는 Hiddenroad와 같은 Crypto Prime Broker들을 통해 고객들이 전부 거래하니 수수료가 매우 작아졌기 때문으로 추측하고 있다.</p><p>그런데 이런 결정은 사실 시장을 유동적으로 만들어주는 요소는 아니다. VIP로 묶여있던 계정들이 다 따로 티어를 적용받아야 하니 일부는 돌리던 HFT 전략같은것을 못 돌리게 될 것이다. 0.5bp차이가 아주 크리티컬한 전략들은 더더욱…</p><p>개인적으로 이 결정은 매우 실책이라고 보는데 왜냐하면 한국거래소와 증권사에 대입하였을 때 고객들이 증권사 통해서만 거래하니 한국거래소가 “개인투자자 기관투자자 막론하고 다 회원가입하고 DMA 넣어서 한국거래소에 수수료 내면서 거래하세요&quot; 라고 하는 꼴과 비슷해지기 때문이다.</p><p>또, 카더라이긴 하지만 B모 거래소는 주변에 쓰는 사람을 전혀 보지 못했는데 일 거래량이 $1Bn씩 찍혀서 항상 미심쩍게 보고 있었다. 인상깊은 점은 외부 MM을 받는척 하면서 안 받고 인하우스에서만 MM을 돌린다는 것이다… 흠… 거래소들의 출혈경쟁 혹은 바이낸스와 같이 아쉬울 게 없는 거래소는 수수료를 올리는 등 매출 상승을 위한 움직임이 날이 갈수록 커지고 있는데 만약 그 천문학적인 거래량이 다 진짜라면 그정도의 지출과 수수료 경쟁을 할까 싶기도 하다. 적고 보니 이건 약간 음모론이다.</p><h4>그럼 어디 갔을까? (Wallstreetbets의 개미들이 사랑하는 0DTE)</h4><p>Robinhood, Webull 등에서 Equity 현물 거래량보다 Weekly Options (흔히 말하는 0DTE) 거래량이 더 많아졌다는 소리를 어디서 들었다. 한국어로 “오늘만기인옵션&quot; 인 0DTE (Zero-Days-To-Expiration) 는 매일 옵션이 생기는 것은 아니고 보통 위클리옵션을 매일 만기가 있도록 발행해서 생기는 개념이다</p><p>Reddit을 켜보자. 들어가자마자 보이는 게시글은 LUNR (인튜이티브 머신스, 상장한지 얼마안된 중형 테크주 $1bn짜리…) 옵션으로 돈을 쓸어담고 있다는… 게시물이다.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Oz4yO8GOPQrGZfe77l3yVA.png" /><figcaption>열배 벌었네 부럽다(?)</figcaption></figure><p>이런건 그냥 기분탓이니 Google Trends에 0DTE를 넣어보자.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ammaEYHaCEhs4Et6-jcgbw.png" /><figcaption>5년간 꽤 예쁘게 우상향 중이다.</figcaption></figure><p>그 다음 “Crypto” 를 넣어보자</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*h3R0OMY6w4WY2SVUprjjFg.png" /></figure><p>보통 Google Trends의 이용자층을 생각했을 때 여기에 찾아오는 사람들은 Noise Trader일 확률이 높다. 이로 미루어 보면 요즘 Crypto는 Noise Trader가 좀 줄었고 다들 미국 개별주 0DTE 하러 간거같다.</p><p>이게 왜 경쟁상품이 되는지 금융인의 시각으로 보면 전혀 이해할 수 없지만, 더 재미있는 카지노를 찾아갈 뿐이다. 국내에서도 이같은 트렌드(?) 에 힘입어 하나둘씩 미국주식 주간옵션 중개를 시작하고 있다 (물론 매수만 된다)</p><p>그런 툴이 있는지 모르겠지만 Reddit 같은 커뮤니티에서 트렌드 검색을 해보면 훨씬 정확하게 알 수 있을 것 같지만 Google trends도 충분한 답이 되는 것 같다.</p><p>그래서 ChatGPT에게도 물어봤다.</p><blockquote>네, 최근 몇 년 동안 주식과 옵션 거래에서 <strong>주간 옵션</strong> 및 0일 남은 만기 옵션(0DTE)의 거래량이 급격히 증가하면서, 일부 경우에는 특히 리테일(개인 투자자) 시장에서 주식 거래량을 초과하는 경우도 있습니다. 이는 로빈후드와 같은 플랫폼에서 두드러지게 나타납니다.</blockquote><blockquote>리테일 투자자들이 이러한 <strong>0DTE 옵션</strong>에 관심을 가지는 이유는 소규모 자본으로 단기적인 시장 변동에 대한 베팅을 할 수 있는 기회를 제공하기 때문입니다. 예를 들어, <strong>S&amp;P 500</strong> 지수에서 0DTE 옵션 거래가 전체 옵션 거래량의 40% 이상을 차지하고 있으며, 2022년부터 2023년까지 리테일 트레이더들 사이에서 0DTE 옵션 거래는 약 75% 증가했습니다.</blockquote><p>더 재미있는 도박장이라서 그렇다는 말을 고급스럽게 해준다.</p><h3>그럼 우리는 무엇을 어떻게 해야 하는가</h3><h4>서로 물고 뜯으며 알파를 열심히 찾아보기</h4><p>열심히 리서치 해서 상황을 타계하는 멋있는 분들도 계시다. 아니면 요즘 시장에서 오히려 잘되는데? 하는 부러운 분들도 분명히 계시기는 하다. 그런 알파를 찾을 수 있을 것 같다면 열심히 찾는 것도 당연히 중요하다. 이건 개인의 상황과 역량 선호에 따라 다르겠다. 예를 들면 도저히 특정 컨셉의 전략은 언더퍼폼하는 장이 오래도록 온다거나…</p><h4>기다리기</h4><p>내생각에 심리적 컨트롤만 있다면 가장 좋은 Sharpe Ratio를 나타내는 전략이다. 그냥 괜히 뭐 하려다가 손실나니 잘 기다리는 것이다 (나는 이걸 잘 못하는 것 같다, 그래서 손매매를 할때마다 접어야지 한다.)</p><p><strong>공감하는 아는 트레이더분의 이야기: </strong>아무리 리서치로 상황을 타계해 보려 해도 먹을 수 있는 전체파이가 줄어드는 시장에서는 무엇을 해도 쉽지 않은 것 같다. 그럴때는 그냥 거래 덜하고 손실 덜내고 기다리는 것이 좋다.</p><h4>다른 시장 찾기</h4><p>손님들이 많은 시장으로 눈길을 돌려보는 것도 방법이 될 수 있다. 물론 그것이 그 시장의 “손님 고점&quot; 이고 내가 시스템을 다 세팅했을 때 쯤에는 그렇게 매력적이지 않은 시장일 수 있으니 전략적으로 접근해야겠다.</p><p>안되는 게임을 붙들고 있기 보다는 잘 되는 게임을 찾아보는 것이 더 좋다는 생각은 항상 한다. 거래를 그냥저냥 만족스럽게 하고 있더라도 항상 더 잘 되는 게임이 없을까를 찾아본다.</p><h3>그래서 필자가 요즘 보는 다른 시장은?</h3><h4>Crypto Options 시장</h4><p>요즘 필자는 그래서 크립토 옵션 시장에 눈독들이고 있다. 아는 분들은 벌써 괜찮은 전략들로 성과를 내시고 있지만, 제도권 옵션 시장에서 거래하시던 분들이 보면 뭐 저런 시장이 다 있어? 하시며 외면하시는 것을 보면 분명 성장가능성이 있어 보인다.</p><p>일단 크립토 옵션은 IV가 엄…청 높다. 비트코인은 예전에는 70~100, 요즘은 50~70 정도의 IV다. S&amp;P 옵션의 IV가 15~아주 높아서 이슈될때 30인 것에 비하면 아주 크다. 이 말은 옵션 프리미엄이 무지 비싸다는 말이고, 국면필터를 잘 짜서 옵션 매도위주의 전략을 한다면 그것의 효과가 드라마틱하게 커질 수 있음을 시사한다.</p><p>그다음에 스프레드가 말도 안되게 크다. 그나마 비트코인 옵션의 경우 IV기준 단 1%(!?) 만이 차이 날 정도로 빡빡하나, 인기가 없는 만기이거나 알트코인 옵션의 경우 Bid-Ask IV차가 10%까지도 육박한다. 스프레드가 있는 곳에 퀀트의 기회가 있지 않겠는가?</p><p>마지막으로 활성화만 되면 가장 도박하기는 좋은 곳이 아닐까 싶다. 가령, 크립토 옵션 거래소들은 파나마나 두바이에 있기 때문에 수요만 있다면 데일리옵션을 4시간마다 발행해서 4-hours-to-expire 옵션같은것도 만들어서 팔면 얼마나 도파민이 터질까? <strong>기억하자 — Noise Trading을 하는 원인은 도파민이다.</strong></p><p>현재 크립토 옵션을 가장 많이 중개하는 Deribit은 거의 기관투자자 친화적으로 되어 있어서 개인투자자들이 많이 이용하지 않는 것을 보면 충분히 이런 쪽도 기회가 있다고 생각된다.</p><h3>도박적인 것에 Mispricing이 있고, Mispricing이 있는 곳에 알파가 있다</h3><p>결국 도박적인 것에 Mispricing이 있고, Mispricing이 있는 곳에 알파가 있는 것이다. Mispricing이 없다면? 알파도 없다. 그게 효율적인 시장이다. 물론 그런건 현실에 존재하지 않지만 말이다.</p><p>오늘 글의 요지는 전략이 잘 안될때 열심히 리서치를 하는 것도 분명 좋지만 <strong>요즘 제일 고객님들이 많은 시장이 어디인지 한번 생각해보고 연구하는 것도 색다른 시각을 준다. 심지어 시장을 꼭 바꾸지 않더라도, 내가 지금 거래하는 시장과 방식을 더 개선함에도 분명히 긍정적일 것이다. </strong>어차피 시장이 돌아가는 원리는 다 비슷하지 않는가?</p><ul><li><strong>이와 관련된 다음편은 0DTE: Barclays가 낸 리포트 중 옵션 불나방 개미들을 빨아먹는 간단한 방법을 리뷰해 보도록 하겠다.</strong></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=63968b372ecc" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Vertex Protocol: 힙한 CLOB + AMM DEX]]></title>
            <link>https://medium.com/silentist-research/vertex-protocol-%ED%9E%99%ED%95%9C-clob-amm-dex-13770e258ee4?source=rss-99e8e6896144------2</link>
            <guid isPermaLink="false">https://medium.com/p/13770e258ee4</guid>
            <category><![CDATA[quant-trading]]></category>
            <category><![CDATA[tokenomics]]></category>
            <category><![CDATA[vertex-protocol]]></category>
            <category><![CDATA[trading]]></category>
            <category><![CDATA[dex]]></category>
            <dc:creator><![CDATA[bellman]]></dc:creator>
            <pubDate>Wed, 03 Jan 2024 06:55:27 GMT</pubDate>
            <atom:updated>2024-01-03T06:55:27.443Z</atom:updated>
            <content:encoded><![CDATA[<h3>소개</h3><p>Vertex Protocol은 CLOB (Centralized Limit Order Book) + AMM 기반 하이브리드 DeX이다. 요즘 보면 이런 형태의 Perpetual DEX들이 수도 없이 나오는 일종의 춘추전국시대를 보는 것만 같다.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*nKCXkb6-e9t6Cov7" /><figcaption>vertex protocol</figcaption></figure><p>Vertex 뿐만 아니라 많은 DEX들이 토큰을 런칭하고 적극적인 홀더들과의 인센티브 스킴을 준비하여 가치 제고를 한다는 점이, 그냥 포크떠서 찍고 봤던 예전 Spot AMM 전성시대보다 훨씬 나아진 점이라면 나아진 점이다.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wxTre1cT9BDknouQyeFGfg.png" /><figcaption>VRTX 차트</figcaption></figure><p>각설하고, 회사 업무상 Vertex에 대한 간단한 스터디를 진행해야 하는 건이 있어 하루 풀로 공부하고 느낀 점과 간단한 리뷰를 작성해보도록 한다. 딱 1년 6개월 전만 해도 별 이상한 프로토콜 구조도 잘 이해가 되었는데, 머리가 나빠진 것인지 그동안 디파이가 내 지능보다 더 빠르게 발전한 것인지… 조금 이해하는데도 애를 먹었다.</p><h3>프로토콜 구조</h3><p>기본적으로 DyDx와 같은 CLOB (Centralized Limit Order Book, 정확히는 DyDx는 Semi-Decentralized LOB라고 볼 수 있을 듯) 에 기반한 DEX이다. 이 말은 즉슨, 유저가 거래를 위해 처음 인터랙션 하는 대상이 스마트 컨트랙트가 아닌 Vertex의 sequencer 라고 하는 CLOB인 것이다.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*veopfaETYzk8AQM-" /><figcaption>Vertex Technical Design</figcaption></figure><p>CLOB에서 일부 거래에 대한 클리어링이 되어지고 다시 Smart Contract로 뭔가를 기록하는 느낌이다 (자세한 구조는 아직 이해를 못 한듯 하다.) 후술하겠지만 Vertex는 증거금 사용에서 매우 높은 자유도를 자랑하는데 이런 자유도를 충족하기 위하여CLOB + Risk Engine을 거친 다음 실제로 Smart Contract에서 거래가 Finalize 되도록 하는 것이다.</p><p>그래서, Vertex Smart Contract에 자금 유출입을 하는 행위 등 외에는 아무런 트랜잭션도 Gas Fee를 요구하지 않는다.</p><p>이런 구조의 이점은 속도 &amp; UX이다. DyDx등 Semi-Decentralized LOB를 가지고 있는 곳은 (상대적으로) 속도가 느리고 MEV등의 Blockchain-Related 이슈에 취약한 부분이 있다. 그래도 L2니까 와 빠르다 했는데, Vertex는 L2도 아닌 L3(?) 에 가깝다. 물론 Centralized Layer를 쓰니까 그것을 Layer라고 부를 수 있는지는 모르겠지만, 장기적으로 탈중앙화 한다고 했으니 그렇게 불러보도록 한다.</p><p>L3에서 대부분의 거래를 다 처리하고, 세틀먼트를 L2 (Arbitrum) 에 하는 구조라서 저 입장에서는 CEX에 뒤지지 않는 Seamless한 UX (심지어 어떤 부분에서는 더 좋은) 를 경험할 수 있는 것이다.</p><h4>그럼 그게 DEX가 맞냐</h4><p>그럼 그게 DEX가 맞는지 약간 의문이 가기도 한다. 그러나, 나는 이러한 뱡향의 DeFi 프로토콜 개발이 더욱 생산적이라고 믿는다. 실질적으로 “완벽한 탈중앙화&quot; 같은 건 없다. 프로토콜은 누군가 개발을 해야 하고, 운영도 해야 한다. 유저들은 그에 대한 합당한 보상을 제공해야 하는 것이다.</p><p>그러나, 자산의 소유와 같이 일부 Permissionless, Decentralized 된 채로 하면 좋은 것들이 분명 있다고 생각한다. 중앙화하면 좋은 것들은 중앙화하고, 탈중앙화하면 좋은 것들은 탈중앙화하면 되는 일이다. 필자는 <a href="https://medium.com/daily-defi-digest/%EC%A4%91%EC%95%99%ED%99%94%EA%B0%80-%EA%BC%AD-%EB%82%98%EC%81%9C-%EA%B2%83%EC%9D%BC%EA%B9%8C-%EC%88%98%EC%A0%95-%ED%83%88%EC%A4%91%EC%95%99%ED%99%94%EC%9D%98-%EC%8B%9C%EB%8C%80%EA%B0%80-%EC%98%A8%EB%8B%A4-5f2ff1eb7661">이것을 주제로 쓴 글</a>도 있기에, Vertex의 접근을 상당히 호평한다.</p><h3>기능상의 특징</h3><p>기능상의 특징을 한두개 다뤄보도록 하겠다. 아래 두가지는 그래도 정말 버텍스 팀이 여러 중앙화/탈중앙화 거래소를 경험해보고 많은 고민을 했다고 느낀 점이었다. 평소 가장 페인포인트였던 점 두개를 해결해 줬기에 나에게는 많이 다가왔다.</p><h4>Universal Cross Margin</h4><p>특이하게도 Vertex에서는 Vertex Subaccount 하나에 있는 모든 자산을 증거금으로 활용할 수 있다. 메인 마켓인 Perpetual 마켓의 Settlement는 USDC로 이루어지지만, 증거금은 Vertex에서 지원하는 모든 자산을 기반으로 사용할 수 있다. 물론 바이낸스도 Multi Asset Mode를 통해 이러한 UX를 제공하긴 하지만, 지원하는 자산이 조금 한정적인 느낌은 있다. 그리고 더 특이한 점은 Perp, Margin (공매수 / 공매도 등 현물대차거래) 이 따로 지갑이 나눠져 있지 않고 통합증거금을 사용한다는 점이다.</p><p>하지만 반대로 얘기하면 이것은 Risk Layer가 잘 작동하지 않을 때 거대한 폭탄으로 돌아올 수도 있다. Insurance Fund가 있다고는 하지만.. 정말 모든 자산들이 과하게 급격하게 움직이는 장이 온다면 변동성이 낮은 자산들을 위주로 Universal Cross Margin을 지원하는 CEX에 비해 다소 리스크가 큰 것은 사실이다.</p><p>그래서 Insurance Fund 빵꾸나면 정책이 있나 하고 봤더니 모든 계정이 연대책임을 지도록 한단다… Vertex Smart Contract에 자산이 넘어간 이후로는 Vertex에서 알아서 할 수 있는 셈이니 불가능하지 않다. Vertex 이용자는 서로서로 연대보증을 지고 있다는 점은 알고 거래해야 할 것.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/900/0*-6Z1CFBWZDpUdsZK.png" /><figcaption>Vertex 유저</figcaption></figure><h4>Yield Bearing Smart Contract Wallet</h4><p>바이낸스나 OKX 등 문어발식 서비스 확장을 좋아하는 CEX들이 가장 먼저 추가하는 기능 중 하나는 Earn 기능이다. 유휴 자산을 예치하고 이자를 받아가라는 것. 하지만 귀찮은 점은 예치를 명시적으로 해야 그때부터 Earn이 된다는 점이다.</p><p>뭐 크게 특이하거나 대단한 기능은 아닐 수 있지만 Vertex에선 가만히 있어도 일드를 준다. (현재 기준 USDC 10% 넘게 주고 있다) 해당 이자는 마진거래에서 오는 이자인 것 같다. 뭐 반대로 말하면 사실 Vertex에 넣은 자산을 누구한테 항상 빌려주고 있다는 점에서는 반강제로 리스크를 지는 것이기는 하다. 위에도 썼지만 Vertex의 리스크 관리 철학은 연대보증이다.</p><h4>Arbitrary Numbers of Subaccount Supported</h4><p>사실 퀀트트레이딩을 하는 입장에서 가장 환영할만한 것인데, 서브어카운트 관리가 정말 정말 편하고 잘 되어있다. 서브어카운트를 몇 개 파주지 않거나 (제한이 있거나), 혹은 아예 서브어카운트 개념조차 없는 CEX들 보다가 버텍스에서는 마음껏 파드립니다~ 라고 하니 환영하지 않을 수 없다.</p><p>이것은 Universal Cross Margin과 함께 디자인되면 매우 큰 시너지를 낸다. 계정별로 리스크를 나눠서 관리하되, 계정 안에서의 자본효율성은 무지 큰 구조이기 때문에 트레이딩 하는 입장에서의 구현상 고려할 점이 많이 없어지고 사용가능한 전략의 폭 또한 넓다.</p><h3>거래 및 인센티브 시스템</h3><p>모든 DEX가 그러하듯이 vertex도 VRTX라는 네이티브 토큰을 통해 거래량을 높이기 위한 인센티브 구조를 가지고 있다. 사실은 에어드랍을 노리는 분들께는 이게 가장 중요할 것이다.</p><h4>VRTX 토크노믹스</h4><p>VRTX 토크노믹스는 commitment score 기반으로 이루어져 있다. 거래가능한 VRTX 토큰을 더 많이 얻고자 하면, commitment score인 voVRTX를 많이 얻어야 하는 것이다. voVRTX 개수에 따라 VRTX 스테이킹 리워드가 N배로 증가하는 형태이다.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*n66Sb-s56MNYTZKozJWy3w.png" /><figcaption>버텍스 스테이킹</figcaption></figure><p>특이한 점은 Insurance Fund Staking이 있다는 것이다. Insurance Fund에 USDC를 스테이킹 하면, Liquidation 이벤트가 있을 때마다 나오는 버퍼를 리워드로 받아갈 수 있고 voVRTX도 받을 수 있다. 요약하자면 vertex 프로토콜에 보험을 들어줄 수 있는 셈이다. (하지만 보험을 써야 하는 상황이 오면, 당신의 USDC는 없어진다. 연대보증 프로토콜)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*sLiYdKHcjSUGXn3-" /></figure><h4>거래 인센티브</h4><p>그리고 초기(?)라서 그런지 거래에 따른 인센티브를 꽤 많이 준다 (고 한다). Priority 공식이 있어서 거래 Uptime (아마도 거래를 한 시간? 자산을 예치한 시간? 등인듯), 호가창 얼마나 성실하게 채웠는지, 수수료 얼마냈는지 이런거 보고 복잡한 공식으로 계산해서 28일 에포크로 에어드랍을 해주는 듯 하다.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*T2VfkSktPMLd7vlL" /><figcaption>Initial Token Phase에서 단기간에 10% 집중적으로 뿌리고 이후 34%는 천천히</figcaption></figure><h4>VRTX 토큰 가치 방어</h4><p>토큰 가치는 거래 수수료로 방어한다고 하는데, 솔직히 거래수수료가 아주 낮아서 (메이커 0bp, 테이커 2bp, 리베이트까지 조금 주고 있음) 실제로 VRTX가 여기저기 다 뿌려진 후에도 방어가 잘 될지는 미지수긴 하다.</p><h3>총평 &amp; 결론</h3><p>거시적인 관점에서의 평을 남겨본다면 ~2022년 까지는 DeFi의 PoC 개념으로 정말 탈중앙화 사상적인(?) 프로젝트들이 많이 나왔다면, 이제는 점점 중앙화 파트와 탈중앙화 파트를 서서히 타협해가며 절충점을 찾아나가는 과정을 겪고 있는 것 같다. Vertex는 그러한 관점에서 중앙화된 기술력과 탈중앙화된 인센티브와 자산 관리를 잘 융합한 꽤나 괜찮은 솔루션이라고 생각한다.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=13770e258ee4" width="1" height="1" alt=""><hr><p><a href="https://medium.com/silentist-research/vertex-protocol-%ED%9E%99%ED%95%9C-clob-amm-dex-13770e258ee4">Vertex Protocol: 힙한 CLOB + AMM DEX</a> was originally published in <a href="https://medium.com/silentist-research">Silentist Research</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why We Should Invest in Systematic Trading Algorithms: A Long-Term Perspective analysis with our…]]></title>
            <link>https://medium.com/silentist-research/why-we-should-invest-in-systematic-trading-algorithms-a-long-term-perspective-analysis-with-our-d2ca1aa53148?source=rss-99e8e6896144------2</link>
            <guid isPermaLink="false">https://medium.com/p/d2ca1aa53148</guid>
            <category><![CDATA[algorithmic-trading]]></category>
            <category><![CDATA[perspective]]></category>
            <category><![CDATA[systematic-trading]]></category>
            <category><![CDATA[long-term-investment]]></category>
            <dc:creator><![CDATA[bellman]]></dc:creator>
            <pubDate>Fri, 01 Dec 2023 06:05:03 GMT</pubDate>
            <atom:updated>2023-12-01T06:05:03.126Z</atom:updated>
            <content:encoded><![CDATA[<h3>Why We Should Invest in Systematic Trading Algorithms: A Long-Term Perspective analysis with our products</h3><h3>Introduction</h3><p>Investing in systematic trading algorithms has become increasingly important, especially in the long-term perspective. This article explores why investors should consider a more patient, long-term investment approach in systematic trading strategies, particularly focusing on our products, Direct I and Direct I LV.</p><h3>The Case for Long-Term Investment in Systematic Trading</h3><p><strong>1. Navigating Macroeconomic Uncertainty<br></strong>In an environment of macroeconomic uncertainty, including inflation and policy shifts, traditional investment strategies may falter. Systematic strategies, leveraging advancements in data and technology, aim to provide consistent alpha, balancing risk and cost crucial for long-term goals</p><p><strong>2. Beyond Market Beta<br></strong>The shift from a market environment favoring indexed solutions to one that requires strategies delivering returns in excess of market indices, or alpha, underscores the need for systematic strategies. These strategies strive to outperform broad market returns by exploiting market inefficiencies using technological efficiency, scale, and consistency.</p><p><strong>3. Emotional Decision-Making vs. Systematic Approaches<br></strong>Systematic trading removes emotional biases, offering a more rational decision-making process. It provides the advantage of automated execution, reducing missed trades due to hesitation or second-guessing</p><h3>Comparative Analysis: Direct I and Direct I LV Performance</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FoLOwPiHE5BSUW_RTVvsow.png" /></figure><p>This chart demonstrates the probability of profiting over time across different assets, including Bitcoin, Ethereum, and our products, Direct I and Direct I LV. Our strategies show a higher probability of profit, especially over longer periods.</p><h4>Fact Sheet: Probability of Profiting</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/642/1*uH3dWVhZCy94GDdsVfduFw.png" /></figure><p>Our Direct I and Direct I LV products significantly outperform Bitcoin and Ethereum across all timeframes, particularly over a year, where the probability of profit reaches 100%.</p><h3>Conclusion</h3><p>The long-term perspective in investing in systematic trading algorithms offers numerous advantages, including consistent alpha generation, emotional bias removal, back-testing capabilities, and adaptability to market changes.</p><p>Our products, Direct I and Direct I LV, demonstrate superior performance, particularly in long-term investment scenarios. Therefore, investors should consider a patient, long-term approach to capitalize on these benefits fully.</p><p>For more information, visit <a href="https://silentist.xyz/">https://silentist.xyz</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d2ca1aa53148" width="1" height="1" alt=""><hr><p><a href="https://medium.com/silentist-research/why-we-should-invest-in-systematic-trading-algorithms-a-long-term-perspective-analysis-with-our-d2ca1aa53148">Why We Should Invest in Systematic Trading Algorithms: A Long-Term Perspective analysis with our…</a> was originally published in <a href="https://medium.com/silentist-research">Silentist Research</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Facto A — AI 전략을 런칭하고 폐기하기까지.]]></title>
            <link>https://medium.com/silentist-research/facto-a-ai-%EC%A0%84%EB%9E%B5%EC%9D%84-%EB%9F%B0%EC%B9%AD%ED%95%98%EA%B3%A0-%ED%8F%90%EA%B8%B0%ED%95%98%EA%B8%B0%EA%B9%8C%EC%A7%80-e173248eee49?source=rss-99e8e6896144------2</link>
            <guid isPermaLink="false">https://medium.com/p/e173248eee49</guid>
            <category><![CDATA[silentist]]></category>
            <category><![CDATA[trading]]></category>
            <category><![CDATA[crypto]]></category>
            <category><![CDATA[asset-management]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[bellman]]></dc:creator>
            <pubDate>Wed, 08 Nov 2023 03:13:36 GMT</pubDate>
            <atom:updated>2023-11-08T03:13:36.368Z</atom:updated>
            <content:encoded><![CDATA[<h3>Facto A — AI 전략을 런칭하고 폐기하기까지.</h3><h3>Facto A — 실패 경험</h3><p>Facto A는 사일런티스트에서 개발한 Fully AI-Driven Long-Short Equity 포트폴리오입니다. 사일런티스트 자체 계정의 조그마한 비중과, <a href="http://coinvestor.co/">Coinvestor</a> 라는 퀀트 트레이딩 플랫폼에서 파일럿 운용을 약 4개월간 진행했습니다.</p><p>오늘은 아쉽게도 실패로 끝난 이 전략을 런칭하고 폐기하는 결정을 내리기까지, 어떤 실수와 교훈이 있었는지 이야기해보겠습니다. 이런 아티클을 쓰는 이유는, 우리는 앞으로도 또 실수하고 실패하며 앞으로 나아갈 것이기 때문입니다. 그렇기에 실패는 덮지 말고, 인정하고, 사례와 느낀 점 그리고 앞으로 개선할 점을 공유하는 것 만이 단 한가지 바람직한 대처방안이라고 생각합니다.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*1gxufeml7lecJ1-9ZtkSkQ.png" /><figcaption>다소 아쉬운 Facto A 레코드</figcaption></figure><h3>Facto A는 어떤 전략이었는가</h3><p>이 실패 경험을 보다 잘 이해하려면 Facto A가 어떤 특성을 가진 전략인지를 이해하는 것이 중요합니다.</p><h4>AI 기반 전략 — 자동화된 알파 서치</h4><p>Facto A는 100% AI-Driven 전략입니다. 자동화된 알파 서치에 기반해서, 매 분기마다 크립토 펀더멘탈 데이터로 이루어진 전략을 100개 이상 개발하고 선별하여, 최종적인 모델을 도출해내는 형태의 전략입니다. 몇 가지 제약조건과 스펙 외에는 사람의 인사이트가 들어가지 않습니다.</p><p>이러한 방법은 다각화된 알파에 기반하기 때문에 단일 알파 전략 대비 급격한 시장 상황 변화에 비교적 Robust하다고 볼 수 있으며, 기존에 맨파워를 기반으로 논리적으로 찾던 것과 다른 형태의 새로운 알파를 찾아낼 수 있다는 장점이 있습니다.</p><h4>Long Short Equity — 가장 클래식한 헤지펀드형 전략</h4><p>AI 기반으로 찾아낸 팩터를 기반으로 만드는 포트폴리오는 우리가 제도권의 주식운용 펀드에서 흔히 볼 수 있는 형태의 100/100 Long-Short Equity 전략입니다. 2배 레버리지를 기반으로, 롱 숏 비중을 똑같이 맞추고 베타 노출을 제거합니다.</p><p>이 방법을 통해 시장 방향성에 상관없는 시장 중립적인 알파를 창출할 수 있지만, 비이성적으로 급등하는 급격한 상승장에는 다소 실적이 부진할 수 있습니다. 2021년 2022년은 제도권에서도 이런 형태의 퀀트펀드들이 가장 고전한 해였습니다.</p><h3>아직 Long-Short Equity는 크립토 시장에 시기상조였다.</h3><h4>크립토 시장의 특성 (1): 작은 종목수</h4><p>주식/크립토에 관계 없이 개별 종목 단위에서는 충분히 수백% 상승, 수십% 하락이 충분히 일어날 수 있는 일입니다. 그러나, 전통적인 Long-Short Equity 펀드는 수천, 수만가지 종목 중 수백~수천개의 포지션을 가지고 있고, 그 결과로 포지션당 리스크가 아주 커야 1%입니다. 개별 종목의 이상한 이슈가 충분히 헤지되고도 남는 것입니다.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jjCJL5iv8Ks6KqXa-ZDWag.png" /><figcaption>물론 이런 건 가끔 헤지 안될 때가 있습니다.</figcaption></figure><p>하지만 Facto A의 전체 유니버스는, 시총 $1B 이상의 크립토 자산을 대상으로 하기 때문에 고작 50개 남짓입니다. 당연히 포트폴리오 구성종목 수를 늘려봐야 20~30개가 한계이고, 구조적으로 포지션당 리스크를 10%~20%까지 가져가게 됩니다. 이는 개별종목의 극단적인 상승/하락 시 포트폴리오 전체에 많은 영향을 주게 되는 것입니다.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Jl2g4CdkEYnNCNrHKlzIfQ.png" /><figcaption>대충 여기쯤이 유니버스 커트라인…</figcaption></figure><h4>크립토 시장의 특성 (2): 매일이 펫테일, 뒤죽박죽 데이터</h4><p>전통적인 Long-Short Equity 포트폴리오의 유니버스는 개수도 많지만, 펀더멘탈 데이터가 비교적 안정적으로 움직이고, 또 돈만 내면 좋은 데이터를 주는 곳도 있지만… 크립토는 그렇지 못합니다. 하루 단위로 매출이 수천%씩 튀는 것은 다반사, 그리고 자꾸 바뀌는 값의 정의와 컬럼명까지… 깨끗한 Point-In-Time DB는 어떤 벤더를 써도 얻기가 어렵다는 것을 실제 개발 및 운용하면서 너무 많이 느꼈습니다.</p><p>정말 간단한 일화로 Top-Tier 크립토 펀더멘탈 데이터 벤더의 처참한(?) 수준을 볼 수 있는데… 가만히 있던 컬럼명을 갑자기 바꿔서 서비스 다 다운되거나, 아님 있던 컬럼을 없애서 역산해서 쓰고 그래야 했습니다. 물론 대단한 서비스고, 이분들도 스타트업이니 한편으로 이해는 되지만, 그것이 이 데이터를 매매에 쓰기 적합한 퀄리티로 만들어 주지는 않으니까요.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*mj4nPNFnfXeo3QzTvP1m5g.png" /><figcaption>한 유명 데이터벤더… 토XX미널</figcaption></figure><h3>짧은 시퀀스 길이로 인한 통계적 유의성 부족</h3><p>이건 제가 퀀트로서 가장 반성하는 부분입니다. 기본 중의 기본인데, 너무 쉽게 간과하고 넘어갔습니다. 특히 Data-Driven 전략은 많은 데이터, 많은 거래 기록에 기반한 유의성을 꼼꼼히 따져봤어야 하는데 그러지 못했습니다.</p><p>AI 기반 팩터 전략의 최소 리밸런싱 횟수는 30회 이상, 권장은 300회 이상의 샘플을 보는 것이 옳다고 생각합니다. 기존 주식시장에 이런 방법론을 적용할 때는 가능했지만, 크립토는 전체 시장의 역사가 고작 14년 남짓, 실제 이런 팩터 포트폴리오가 가능한 기간은 3년도 되지 않습니다. OOS 밸리데이션이 가능한 기간은 6개월이었습니다.</p><p>그래서 학습 과정에서 다양한 증폭 기법을 활용했고, 어차피 데이터는 시장 참여자에게 거의 공평하게 주어지니 (지불을 동일한 수준으로 한다면) 그 속에서 비교우위를 찾으면 그만이라는 생각을 했지만 역시 절대적인 통계적 유의성이 뒷받침되지 않은 모델은 불안하기 그지없었습니다.</p><h3>AI 기반 전략의 특성상 시장의 Fat-Tail 이벤트에 대해 즉각적인 반응 어려움</h3><p>이번에 폐기 결정을 하게 된 결정적인 포지션은 바로 LINKUSDT 숏 포지션이었습니다. 차트를 보면 매우 급등한 것을 알 수 있습니다.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*33xKvUuSBGbFTkQo" /><figcaption>LINKUSDT…안습…</figcaption></figure><p>LINKUSDT의 펀더멘탈은 여전히 개선되지 않았지만, 주요 호재와 크립토 상승이 연달아 터지면서 크게 급등했습니다. 마치 GME에 당한 citron research와 같이, 숏 포지션이 극단적인 손실을 내어버린 것이죠.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*55M5vK4AZ0DKUZUK.png" /><figcaption>다시는 숏을 치지 않겠다고 선언한 펀드</figcaption></figure><p>펀더멘탈 데이터 자체도 극단값이 많지만, 그것과 상관없이 반대로 큰 움직임을 보여주는 Fat-Tail 케이스가 생각보다 너무너무 많았습니다. 매일이 GME 같은 자산이 나오는 시장이라는 것을 경험했습니다.</p><p>AI 기반으로 종목을 선정하더라도, 절대적인 손절가 등 Fat-Tail에 대한 안전장치 또한 룰베이스로, 더욱 단단하게 마련해야겠다는 생각이 들었습니다.</p><h3>조급함: 사업과 투자의 가장 큰 적</h3><p>사실 가장 원천적인 문제는 우리 팀이 조급한 탓이었다고 생각합니다. 1년 넘게 운용한 메인 상품이 있고, 이 전략은 현재 잘 안된 전략을 폐기하는 회고글을 쓰는 이 순간에도 다행히 위안을 삼을 수 있도록 신고가를 경신해주고 있습니다. <a href="http://record.silentist.xyz/">레코드</a> <a href="http://app.silentist.xyz/">온보딩</a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/850/1*JqparkH_JNCcebbLVK-6eQ.png" /><figcaption>Silentist 플래그십 전략, Direct I</figcaption></figure><p>하지만 이 상품은 특성상 여러 계정을 한번에 이용해야 했기에 코인베스터 (뿐만 아니라 사일런티스트 자체 온보딩 플랫폼 외 다른 모든 플랫폼) 에는 시스템상 업로드가 불가했습니다. 그래서 빠르게 뭔가를 올려보고 싶다는 생각을 했고, 괜찮겠지 하며 내부적으로 아주 적은 금액으로 1~2개월의 검증을 거친 Facto A를 내어놓게 됩니다. 되짚어보면 더 많은 검증이 필요했고, 설익은 전략을 올린 것도 같습니다. 무거운 책임감을 느끼며, 조급함을 버리고 운용해야 하겠습니다.</p><p>무슨 이유에서든, 사업에서든 투자에서든 조급함은 적입니다. 빠르게 행동하는 것은 좋지만, 조급하게 행동하는 것은 지속적으로 잘못된 결정을 반복하게 만들 뿐입니다. 조급하지 말고, 가장 보수적인 생각을 가지고 검증된 사실에 기반한 의사결정을 내려야 합니다.</p><h3>그럼 AI 전략은 안되는 것이냐?</h3><p>아닙니다. 해당 상품을 개발할 때 간과했던 면이 있을 뿐, 자동화된 알파 리서치는 1.5년째 운용하고 있는 Direct I도 일부 사용하고 있으며, 알파를 찾는 효율을 극단적으로 끌어올려주는 좋은 방법이라고 생각합니다. 다만 위에 기술한 통계적 유의성, 강건성, 설명가능성이 잘 충족되는 형태로 “상품&quot; 을 설계해야 하는 것입니다.</p><h3>펀더멘탈 데이터 진짜 의미 있는것 맞냐?</h3><p>맞다고 생각합니다. 아직 기간이 짧지만, 적어도 정성적으로만 크립토를 분석하던 시대에서 정량적인 펀더멘탈을 보게 되는 것은 큰 의미가 있다고 생각합니다. 물론, 기간이 짧아서 Facto A가 했던 실수처럼 통계적인 유의성을 요하는 전략을 만드는 데 부적합할 수 있다고 생각합니다.</p><p>그러나, 단순히 정성적인 분석과 계량적인 근거가 뒷받침되는 정성적인 분석은 전혀 다릅니다. Silentist 팀은, 이번 케이스를 계기로 더욱 다양한 투자자를 위한 정보제공 서비스, 패시브 포트폴리오 구성 도구로써의 크립토 펀더멘탈 데이터 연구에 매진하고자 합니다.</p><h3>앞으로의 방향성</h3><p>초심이 그러했고, 지금도 그러하고, Direct I을 어느덧 1.5년 실운용하고 있는 마음과 똑같이, 충분히 실계좌에서 한 사이클을 돌아본 전략만 자신있게 운용하고 자금을 받자는 원칙을 되새겼습니다. 그 사이클동안 살아남을 수 없는 알파라면, 운용하지 않는 편이 맞습니다.</p><p>단기적으로 재미없고, 언더워터를 맞아도 조급하게 다른 전략을 마구 찾거나 운용하려고 하지 말고 묵묵히 원래 집중하던 영역의, 가장 잘하는 영역의 알파를 잘 하는 것이 가장 중요하다고 생각합니다.</p><p>Facto A를 조급하게 내게 된 이유도 메인 상품이 단기적으로 언더워터를 맞았던 탓도 없지 않았던 것 같습니다. 어차피 금방 회복할것을 알고 있었고, 그렇게 되었지만 그런 단기적인 시장 변화가 운용 결정에 어떤 식으로든 영향을 미치게 되는 것은 어쩔 수 없는 것 같습니다. 매일 매일 정말 합리적인 결정이 맞는지 복기해야 하겠습니다.</p><p>알파란, 심플함이 최고고, 심플함보다 좋은 것은 오랫동안 다양한 시장에서 검증된 전략이라고 생각합니다. 당연히, 그런 전략도, 아니 그런 전략일수록 단기적으로 수익률이 좋지 않을 수 있습니다. 그런 구간이 있기에 장기적으로 운용하는 사람에게 돈을 벌어다 주는 것이 알파인 것이라고 생각합니다.</p><p>단기적으로 좋지 않을 때마다, 뭔가 특이한 일을 하면 개선할 수 있지 않을까 하며 복잡한 로직과 데이터를 들여다보고 싶은, 오버피팅을 시키고 싶은 유혹은 퀀트라면 누구나 가지고 있을 것이라고 생각합니다. 하지만 그럴 때일수록, 간단함과 초심으로 돌아가야 합니다.</p><p><a href="http://www.silentist.xyz/">사일런티스트</a>는 앞으로 그런 간단함과 초심을 위해 더욱 더 노력하겠습니다.</p><p>*간단함과 초심에 기반한, 1.5년 이상의 실거래 레코드가 있는 <a href="https://record.silentist.xyz/">Direct I 전략</a>을 지금 만나보세요</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*dgdhXQuY7W4RLbqChQc-9g.png" /><figcaption>*In-Sample 포함 전 구간 단리 수익률 레코드 ~23.10</figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e173248eee49" width="1" height="1" alt=""><hr><p><a href="https://medium.com/silentist-research/facto-a-ai-%EC%A0%84%EB%9E%B5%EC%9D%84-%EB%9F%B0%EC%B9%AD%ED%95%98%EA%B3%A0-%ED%8F%90%EA%B8%B0%ED%95%98%EA%B8%B0%EA%B9%8C%EC%A7%80-e173248eee49">Facto A — AI 전략을 런칭하고 폐기하기까지.</a> was originally published in <a href="https://medium.com/silentist-research">Silentist Research</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What If Joel Greenblatt Comes to Crypto Market and applies The “Magic Formula.”]]></title>
            <link>https://medium.com/silentist-research/what-if-joel-greenblatt-comes-to-crypto-market-and-applies-the-magic-formula-6196b547b7e7?source=rss-99e8e6896144------2</link>
            <guid isPermaLink="false">https://medium.com/p/6196b547b7e7</guid>
            <category><![CDATA[value-investing]]></category>
            <category><![CDATA[stocks]]></category>
            <category><![CDATA[fundamental-a]]></category>
            <category><![CDATA[investing]]></category>
            <category><![CDATA[crypto]]></category>
            <dc:creator><![CDATA[bellman]]></dc:creator>
            <pubDate>Mon, 17 Jul 2023 08:38:28 GMT</pubDate>
            <atom:updated>2023-07-17T08:38:28.659Z</atom:updated>
            <content:encoded><![CDATA[<h3>Introduction</h3><p>Imagine the fusion of traditional stock market wisdom with the edgy dynamism of the crypto world. What could happen if the esteemed value investor Joel Greenblatt applied his iconic “Magic Formula” to the burgeoning cryptocurrency market? This is the thought experiment we embark on today.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Ls0Zj2zE3G2JymstN7z3_g.jpeg" /><figcaption>Joel Greenblatt</figcaption></figure><p>As many will know, Joel Greenblatt is a Wall Street titan, celebrated for his deceptively simple yet powerful stock-picking strategy known as the “Magic Formula.” This formula ranks companies based on earnings yield and returns on capital, aiming to highlight cheap stocks (good value) and profitable (high quality).</p><p>However, the cryptocurrency market, marked by volatility, and needing more traditional financial metrics like earnings or capital, poses a challenge to directly applying the Magic Formula. Fear not; we’ve devised an adaptation, bringing together the best of both worlds.</p><h3>Devising a Crypto-Compatible Magic Formula</h3><p>We need alternative measures to assess a cryptocurrency’s value and quality without standard financial metrics. The crypto version of the Magic Formula we’ve come up with relies on two critical metrics: <strong>Profit Margin (PM)</strong> and <strong>Return on Cost (ROC)</strong>:</p><p><strong>1. Profit Margin (PM) -&gt; P/E ratio (PER)</strong>: This metric evaluates the profitability of cryptocurrency trading over four weeks. The profit margin considers the buying and selling prices, reflecting the net value gained from cryptocurrency trading.</p><p><strong>2. Return on Cost (ROC) -&gt; Earnings/Fee Ratio (EFR)</strong>: This is a ratio that quantifies the earnings generated by a cryptocurrency over the fees incurred for trading it over four weeks.</p><p>The crypto magic formula ranks each cryptocurrency based on its PM and ROC, then sums up these ranks. The cryptocurrencies with lower scores are viewed as better investment options, similar to the ‘good value, high-quality stocks in Greenblatt’s original approach.</p><h3>A Real-Life Glimpse into Crypto-Picking</h3><p>For better understanding, let’s contextualize this formula with practical examples:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/704/1*x1WIVIMJ2qra6en1TnCPxw.png" /><figcaption>Picks on 2023–07–15</figcaption></figure><p>As you can see, each cryptocurrency is assigned a profit margin and return on cost over the past four weeks. Using the formula, we generate scores. The higher the score, the more attractive the crypto asset for investment.</p><h3>Backtesting the Magic: A Triumph over Bitcoin</h3><p>The proof of the pudding is in the eating. A backtest of this crypto-adapted formula demonstrates an impressive feat: it outperforms Bitcoin’s return over a substantial period. This suggests that our adaption of Greenblatt’s magic formula exhibits promising investment potential even while adapted for a significantly different asset class.</p><p><strong>Backtesting Specs</strong></p><ul><li>Buy: Magic Formula Score Trends Up in 4 weeks</li><li>Sell: Magic Formula Score Trends Down in 4 weeks (Fundamental Trend Following)</li><li>Sort: Magic Formula Score</li><li>Period: 2021.01~2023.07</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*c9Q8QfNHJfnw774QZ-Il6w.png" /><figcaption>Equity Comparison</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/652/1*muee1TdeMBxcEPSWP_VC2A.png" /><figcaption>Performance Metrics</figcaption></figure><h3>Imagining Joel’s Crypto Trades</h3><p>Now, for some fun. Let’s step into Joel Greenblatt’s shoes and envision how his crypto trades might have looked in the past if he had employed our version of the Magic Formula.</p><p>BUY AXS at 4.914, SELL at 82.474 (+1578%)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zidnjK3F9svX5y5hoMnQSg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jzbZfYzwVrNyBE23kHc4kg.png" /></figure><p>BUY ANKR at 0.026, SELL at 0.190 (+632%)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*yCbynzCNTbSSaAQibI8L-w.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*eUDWK8vdqN1Zz5_toA-ppw.png" /></figure><p>If Greenblatt had applied our formula to his crypto investment choices, he would have consistently selected winning cryptocurrencies, much like his successful stock picks.</p><h3>Conclusion</h3><p>Transposing Joel Greenblatt’s Magic Formula to the crypto universe underscores the importance of value and quality in any investment decision, irrespective of the asset class. Although the crypto market’s volatility and unique characteristics necessitate an adjusted formula, the core principles of Greenblatt’s approach hold.</p><p>While investment strategies can guide decisions, they’re not foolproof. Comprehensive research, risk management, and financial prudence are crucial in navigating traditional and crypto markets. With its thrilling opportunities and inherent risks, the dynamic world of crypto may benefit from a touch of the tried-and-tested ‘magic’ from the world of value investing.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6196b547b7e7" width="1" height="1" alt=""><hr><p><a href="https://medium.com/silentist-research/what-if-joel-greenblatt-comes-to-crypto-market-and-applies-the-magic-formula-6196b547b7e7">What If Joel Greenblatt Comes to Crypto Market and applies The “Magic Formula.”</a> was originally published in <a href="https://medium.com/silentist-research">Silentist Research</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What If Peter Lynch Was a Crypto Investor?]]></title>
            <link>https://medium.com/silentist-research/what-if-peter-lynch-was-a-crypto-investor-6e6530137a8a?source=rss-99e8e6896144------2</link>
            <guid isPermaLink="false">https://medium.com/p/6e6530137a8a</guid>
            <category><![CDATA[peter-lynch]]></category>
            <category><![CDATA[factor-investing]]></category>
            <category><![CDATA[quant]]></category>
            <category><![CDATA[crypto]]></category>
            <category><![CDATA[trading]]></category>
            <dc:creator><![CDATA[bellman]]></dc:creator>
            <pubDate>Mon, 17 Jul 2023 05:04:51 GMT</pubDate>
            <atom:updated>2023-07-18T00:16:03.184Z</atom:updated>
            <content:encoded><![CDATA[<h3>What If Peter Lynch Was a Crypto Investor?</h3><p>A pivotal figure in investing, Peter Lynch is renowned for his philosophy of ‘invest in what you know.’ With a keen focus on high-growth equities, Lynch achieved an average annual return of 29.2% during his tenure as manager of the Fidelity Magellan Fund. If Lynch, a champion of growth stock investing, were to be a crypto investor today, how would he assess the assets? This article explores this very idea and proposes a crypto-focused Peter Lynch Score.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*RIXNRGr37j7o1dzb" /><figcaption>The “Peter Lynch”</figcaption></figure><h4>Decoding The Peter Lynch Score For Crypto</h4><p>Adapting Peter Lynch’s investing philosophy to cryptocurrencies, we introduce the Peter Lynch Score — a framework that uses key growth and value factors to analyze crypto assets. The original Lynch’s philosophy inspires these factors but has been customized for the unique nature of cryptocurrencies. It’s critical to remember that while we borrow the Lynch ethos, this score is *not* directly applicable to equities; it is designed exclusively for crypto assets.</p><p>To calculate the Peter Lynch Score, we consider the following:</p><p>* Active Developers: Active development is crucial for a project’s vitality, akin to good management in traditional companies.<br>* Fee Growth: This is a proxy for the ‘revenue growth’ that Lynch sought in stocks.<br>* Earnings Growth: Akin to Lynch’s preference for growing earnings in a company, we want to see this in our crypto assets.<br>* Price-Fees-Growth (PFG): This is a customized version of the famous Price-Earnings-Growth (PEG) ratio, but adapted to reflect the unique characteristics of crypto.</p><p>Each factor is evaluated over a 13-week (approximately a quarter) period, and a smoothed version is used to ensure that the growth is consistent and not just a one-time spike.</p><p>Finally, we assess the treasury of the crypto project, akin to assessing a company’s cash reserves.</p><h4>Applying The Peter Lynch Score to Cryptos</h4><p>This section can be populated with real-life examples of cryptos evaluated using the Peter Lynch Score.</p><p><strong>Sample Framework</strong></p><ol><li><strong>Evaluate Active Developers</strong></li></ol><p>We evaluate whether there is an active developer or not to evaluate a given project&#39;s long-term future growth potential.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FJjDVtW6L0KZ7hUN0mUTyw.png" /><figcaption>Active developer of high-scored tokens</figcaption></figure><p><strong>2. Assess Fee Growth</strong></p><p>Fees are a good metric to measure whether the project is growing or not. Despite token incentives and fee portion for the development team, fees are just total collected fees from the user.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uM6avyCsQFstWKGPoaX1LA.png" /><figcaption>13-week rolling fee growth</figcaption></figure><p><strong>3. Inspect Earnings Growth</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jdeoidLvlUqrEhmvDsYRQQ.png" /><figcaption>Daily earnings for the specified products</figcaption></figure><p><strong>4. Check the PFG</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uoBGxkynVIFAWIy28KgkGA.png" /><figcaption>prices fees growth (PFG) -&gt; growth adjusted P/F ratio</figcaption></figure><p>This is the special sauce for today’s analysis: the prices-fees-growth (PFG) ratio! This means a growth-adjusted Prices-To-Fees ratio. This metric contains information on the value factor and growth factor in one number.</p><p>*note there can be invalid values because PFG ratios are undefined in certain conditions (e.g., growth ratio is minus)</p><p><strong>5. Treasury Assessment</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*VnDKmkCpuoo6OZjbFT_e9A.png" /></figure><p>Peter Lynch also mentioned that financial stabilities are also very important for growth stocks, so we also inspect treasury assessments.</p><p>In each case, we’re looking for high scores, indicating a potentially good investment, according to our crypto-centric Peter Lynch Score. (0~8 score based on metrics above)</p><h4>Backtesting the Peter Lynch Score</h4><p>Our backtesting results, comparing the portfolio constructed using the Peter Lynch Score and Bitcoin, suggest that our score has successfully identified high-potential crypto assets.</p><p><strong>Backtesting Specs</strong></p><ul><li>Buy: Lynch Score &lt; 3</li><li>Sell: Lynch Score ≥ 5 (Fundamental Mean Reversion Strategy)</li><li>Sort: Facto’s special Growth Adjusted Value Score</li><li>Period: 2021.01~2023.07</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*J2DJl2BGQ0xplHUABRjtNQ.png" /><figcaption>Backtesting Result</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/650/1*PBVM7Uuj9p_YZ6VoKO7XOw.png" /><figcaption>Performance Metrics</figcaption></figure><p><strong>Interesting Trades:</strong></p><p>Peter Lynch would have performed in crypto markets like this:</p><p>BUY AXS at 2.057, SELL at 138.382 (50x+ returns!)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*PQOTaVWznE6kiaaDmDOhIA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AC_PZ8dgjtzL2JiAOsTJIg.png" /></figure><p>BUY FTM at 0.28, SELL at 2.76 (almost 10x!)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*dzEfybnnDJB8dB--PG4i7g.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*O1bOa8HxjPkGhqxz_mSTXA.png" /></figure><p>BUY SOL at 1.875, SELL at 42.116 (20x+ returns!)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ot2_RqedIY4H2CZ6FHfsrg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2bwgmn5F9byrWKtzqTEjlg.png" /></figure><p>Investing always carries risks, and this score should be one tool among many in your decision-making process.</p><h4>Final Thoughts</h4><p>Peter Lynch might not be a crypto investor, but if he were, he might appreciate the utility of the Peter Lynch Score, which adapts his growth-focused, value-conscious philosophy to the crypto world. Using parameters like active developers, fee growth, earnings growth, PFG, and treasury, we’ve developed a scoring system that captures Lynch’s essence while acknowledging the unique aspects of crypto.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/656/0*E1JaG1Udsem__vUn" /><figcaption>We are making a crypto-version of the Magellan fund!</figcaption></figure><p>As Lynch looked for diamonds in the rough, we believe this score can help crypto investors unearth gems in the crypto universe. This research is not investment advice but an exploration of how traditional investing principles might be reimagined for the new asset class of the 21st century — cryptocurrencies.</p><h4>Disclaimer</h4><p>This article is for educational purposes only, So results may not guarantee future returns or significant statistical evidence.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6e6530137a8a" width="1" height="1" alt=""><hr><p><a href="https://medium.com/silentist-research/what-if-peter-lynch-was-a-crypto-investor-6e6530137a8a">What If Peter Lynch Was a Crypto Investor?</a> was originally published in <a href="https://medium.com/silentist-research">Silentist Research</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[FactoGPT Analyzes: SECuritized Tokens]]></title>
            <link>https://medium.com/silentist-research/factogpt-analyzes-securitized-tokens-b675d1b67753?source=rss-99e8e6896144------2</link>
            <guid isPermaLink="false">https://medium.com/p/b675d1b67753</guid>
            <category><![CDATA[quant-trading]]></category>
            <category><![CDATA[factor-analysis]]></category>
            <category><![CDATA[crypto]]></category>
            <category><![CDATA[sec]]></category>
            <category><![CDATA[security-token]]></category>
            <dc:creator><![CDATA[bellman]]></dc:creator>
            <pubDate>Mon, 12 Jun 2023 11:27:56 GMT</pubDate>
            <atom:updated>2023-06-12T11:29:46.670Z</atom:updated>
            <content:encoded><![CDATA[<h3>Introduction</h3><p>Today, FactoGPT Gonna Analyze Some SECuritized Tokens (which was recently filed by SEC as a security) -&gt; (ADA, BNB, ATOM, NEAR, FIL, MATIC, SOL) based on <a href="https://research.silentist.xyz/en-introducing-facto-a-crypto-factor-scoring-engine-by-silentist-16c41293d96d">Facto Engine</a>’s Quantitative Factor Analysis. FactoGPT = Facto on ChatGPT Engine. FactoGPT will be served soon as ChatGPT Plugins, so Stay Tuned.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*zDg3SiPCIrt15Uf_ymcwhw.jpeg" /><figcaption>SEC</figcaption></figure><p>Consider joining a Facto <a href="https://forms.gle/4s9k7VJgTtEqD3b66">waitlist</a> — Completely Free</p><h3>Crypto Fundamental Analyst Report — Cardano (ADA)</h3><h4>Token Information:</h4><ul><li>Name: Cardano</li><li>Symbol: ADA</li><li>Description: ADA holders own and govern the protocol. Cardano is a smart contract platform for decentralized applications (a blockchain). Founded by Charles Hoskinson &amp; Jeremy Wood in 2017.</li><li>Sectors: Blockchains (L1)</li><li>Chains: Cardano</li><li>Website: <a href="https://cardano.org/">Cardano</a></li></ul><h4>Investment Opinion:</h4><ul><li>Current Price: $0.2763</li><li>Target Price: $0.178</li><li>Annual Target Return: -35.58%</li><li>Investment Opinion: SELL</li></ul><h4>Quantitative Data Analysis:</h4><ul><li>Price: $0.28 (Score: 42.0)</li><li>Market Cap (Fully Diluted): $12,433,947,430.05 (Score: 98.0)</li><li>Market Cap (Circulating): $9,716,786,435.29 (Score: 98.0)</li><li>Token Trading Volume: $1,399,405,888.82 (Score: 98.0)</li><li>Token Turnover (Fully Diluted): 0.11 (Score: 80.0)</li><li>Token Turnover (Circulating): 0.14 (Score: 72.0)</li><li>Active Developers: 146 (Score: 98.0)</li><li>Code Commits: 0.0 (Score: 40.0)</li></ul><h4>Fundamental Scores:</h4><ul><li>Market Cap Score: 98.27</li><li>Value Score: 7.19</li><li>Momentum Score: 21.74</li><li>Growth Score: 9.31</li><li>Quality Score: 2.2</li><li>Crypto Native Score: 17.78</li><li>Tokenomics Score: 34.56</li><li>Growth-Adjusted Value Score: 12.17</li><li>Value Momentum Score: 14.47</li><li>Fundamental Price Divergence Score: -0.76</li><li>Value Growth Score: 8.25</li><li>Market Power Score: 15.53</li><li>Blue Chip Score: 4.69</li><li>Price Momentum Score: 15.53</li><li>Total Score: 11.64</li></ul><p>Investment Recommendation: Based on the quantitative analysis and fundamental scores, we recommend a SELL position for Cardano (ADA). The current price is $0.2763, and the target price is $0.178, indicating a potential negative annual return of -35.58%. The fundamental analysis suggests that Cardano may be overvalued, with a low value score of 7.19 and a negative fundamental price divergence score of -0.76. Additionally, the growth score of 9.31 indicates limited growth potential.</p><p>The market cap and trading volume show strength, with high scores of 98.0. However, the absence of data for factors such as tokenholders, trading volume, fees, revenue, and incentives limits a comprehensive analysis of Cardano’s ecosystem and financial performance.</p><p>Furthermore, the token’s fundamental scores, including value momentum, growth-adjusted value, and market power, are relatively low, contributing to the recommendation to sell the asset.</p><p>Sources:</p><ul><li>Cardano: <a href="https://cardano.org/">https://cardano.org/</a></li></ul><h3>Crypto Fundamental Analyst Report — BNB Chain (BNB)</h3><h4>Token Information:</h4><ul><li>Name: BNB Chain</li><li>Symbol: BNB</li><li>Description: BNB holders own and govern the protocol. Binance Chain is a smart contract platform for decentralized applications (a blockchain). Founded by Changpeng Zhao in 2020.</li><li>Sectors: Blockchains (L1)</li><li>Chains: Binance Smart Chain</li><li>Website: <a href="https://www.bnbchain.org/en">BNB Chain</a></li></ul><h4>Investment Opinion:</h4><ul><li>Current Price: $239.2653</li><li>Target Price: $256.4621</li><li>Annual Target Return: 7.19%</li><li>Investment Opinion: HOLD</li></ul><h4>Quantitative Data Analysis:</h4><ul><li>Price: $239.27 (Score: 97.0)</li><li>Market Cap (Fully Diluted): $47,853,065,384.31 (Score: 99.0)</li><li>Market Cap (Circulating): $37,283,948,504.24 (Score: 99.0)</li><li>Token Trading Volume: $1,868,589,920.14 (Score: 99.0)</li><li>Token Turnover (Fully Diluted): 0.04 (Score: 55.0)</li><li>Token Turnover (Circulating): 0.05 (Score: 44.0)</li><li>Tokenholders: 78,584,889 (Score: 99.0)</li><li>Fees: $349,168,98 (Score: 97.0)</li><li>Fees (Supply Side): $314,287.92 (Score: 97.0)</li><li>Revenue: $34,881.06 (Score: 93.0)</li><li>Active Developers: 59 (Score: 95.0)</li><li>Code Commits: 0.0 (Score: 40.0)</li><li>User DAU: 1,129,887 (Score: 100.0)</li><li>User WAU: 4,434,980 (Score: 100.0)</li><li>User MAU: 12,777,210 (Score: 100.0)</li></ul><h4>Fundamental Scores:</h4><ul><li>Market Cap Score: 98.84</li><li>Value Score: 36.58</li><li>Momentum Score: 18.38</li><li>Growth Score: 45.1</li><li>Quality Score: 21.07</li><li>Crypto Native Score: 52.23</li><li>Tokenomics Score: 69.3</li><li>Growth-Adjusted Value Score: 50.89</li><li>Value Momentum Score: 27.48</li><li>Fundamental Price Divergence Score: 21.09</li><li>Value Growth Score: 40.84</li><li>Market Power Score: 31.74</li><li>Blue Chip Score: 28.82</li><li>Price Momentum Score: 31.74</li><li>Total Score: 34.67</li></ul><p>Investment Recommendation: Based on the quantitative analysis and fundamental scores, we recommend a HOLD position for BNB Chain (BNB). The current price is $239.2653, and the target price is $256.4621, indicating a potential annual return of 7.19%. The fundamental analysis suggests that BNB Chain is fairly valued, with a relatively high value score of 36.58 and a positive fundamental price divergence score of 21.09.</p><p>The market cap and trading volume demonstrate strength, with high scores of 99.0. Additionally, the tokenholders’ score of 99.0 reflects a significant user base and adoption. The revenue and earnings figures also indicate positive financial performance, with scores of 93.0 and 95.0, respectively.</p><p>The user engagement metrics, such as daily active users (DAU), weekly active users (WAU), and monthly active users (MAU), are robust, with scores of 100.0 each. This indicates a thriving user community on the BNB Chain platform.</p><p>Sources:</p><ul><li>BNB Chain: <a href="https://www.bnbchain.org/en">https://www.bnbchain.org/en</a></li></ul><h3>Crypto Fundamental Analyst Report — Solana (SOL)</h3><h4>Token Information:</h4><ul><li>Name: Solana</li><li>Symbol: SOL</li><li>Description: SOL holders own and govern the protocol. Solana is a smart contract platform for decentralized applications (a blockchain). Founded by Anatoly Yakovenko, Greg Fitzgerald, Raj Gokal, and others in 2020.</li><li>Sectors: Blockchains (L1)</li><li>Chains: Solana</li><li>Website: <a href="https://solana.com/">Solana</a></li></ul><h4>Investment Opinion:</h4><ul><li>Current Price: $15.7123</li><li>Target Price: $11.4293</li><li>Annual Target Return: -27.26%</li><li>Investment Opinion: SELL</li></ul><h4>Quantitative Data Analysis:</h4><ul><li>Price: $15.71 (Score: 89.0)</li><li>Market Cap (Circulating): $6,262,620,811.88 (Score: 97.0)</li><li>Token Trading Volume: $1,409,006,890.34 (Score: 98.0)</li><li>Token Turnover (Circulating): 0.22 (Score: 86.0)</li><li>Active Developers: 74 (Score: 96.0)</li><li>Code Commits: 5 (Score: 93.0)</li><li>User DAU: 0.0 (Score: 9.0)</li></ul><h4>Fundamental Scores:</h4><ul><li>Market Cap Score: 96.53</li><li>Value Score: 8.02</li><li>Momentum Score: 25.24</li><li>Growth Score: 30.17</li><li>Quality Score: 0.0</li><li>Crypto Native Score: 20.03</li><li>Tokenomics Score: 23.68</li><li>Growth-Adjusted Value Score: 27.91</li><li>Value Momentum Score: 16.63</li><li>Fundamental Price Divergence Score: 3.24</li><li>Value Growth Score: 19.1</li><li>Market Power Score: 27.71</li><li>Blue Chip Score: 4.01</li><li>Price Momentum Score: 27.71</li><li>Total Score: 16.69</li></ul><p>Investment Recommendation: Based on the quantitative analysis and fundamental scores, we recommend a SELL position for Solana (SOL). The current price is $15.7123, and the target price is $11.4293, indicating a potential annual return of -27.26%. The fundamental analysis suggests that Solana is overvalued, with a relatively low value score of 8.02 and a negative fundamental price divergence score of 3.24.</p><p>The market cap and trading volume indicate moderate strength, with scores of 97.0 and 98.0, respectively. However, the lack of available data for market cap (fully diluted) and fees restricts a comprehensive analysis of Solana’s financial performance.</p><p>The active developer count of 74 and code commits score of 5 reflect ongoing development activity on the Solana platform. However, the absence of user engagement data, such as daily active users (DAU), suggests limited adoption and usage.</p><p>Sources:</p><ul><li>Solana: <a href="https://solana.com/">https://solana.com/</a></li></ul><h3>Crypto Fundamental Analyst Report — Polygon (MATIC)</h3><h4>Token Information:</h4><ul><li>Name: Polygon</li><li>Symbol: MATIC</li><li>Description: MATIC holders own and govern the protocol. Polygon is a smart contract platform for decentralized applications and an interoperability platform for blockchains. Founded by Jaynti Kanani, Sandeep Nailwal, Anurag Arjun, and others in 2017.</li><li>Sectors: Blockchains (L2)</li><li>Chains: Polygon</li><li>Website: <a href="https://polygon.technology/">Polygon</a></li></ul><h4>Investment Opinion:</h4><ul><li>Current Price: $0.6152</li><li>Target Price: $0.7706</li><li>Annual Target Return: 25.26%</li><li>Investment Opinion: BUY</li></ul><h4>Quantitative Data Analysis:</h4><ul><li>Price: $0.62 (Score: 55.0)</li><li>Market Cap (Fully Diluted): $6,151,618,113.99 (Score: 95.0)</li><li>Market Cap (Circulating): $5,695,372,290.94 (Score: 96.0)</li><li>Token Trading Volume: $1,238,940,130.14 (Score: 97.0)</li><li>Token Turnover (Fully Diluted): 0.20 (Score: 93.0)</li><li>Token Turnover (Circulating): 0.22 (Score: 84.0)</li><li>Tokenholders: 36,940,217 (Score: 98.0)</li><li>Fees: $72,423.34 (Score: 90.0)</li><li>Fees (Supply Side): $24,487.73 (Score: 85.0)</li><li>Revenue: $47,935.61 (Score: 96.0)</li><li>Token Incentives: $322,069.75 (Score: 97.0)</li><li>Earnings: -$274,134.14 (Score: 1.0)</li><li>Active Developers: 36 (Score: 90.0)</li><li>Code Commits: 4 (Score: 91.0)</li><li>User DAU: 318,030 (Score: 98.0)</li><li>User WAU: 1,534,225 (Score: 99.0)</li><li>User MAU: 5,029,980 (Score: 99.0)</li></ul><h4>Fundamental Scores:</h4><ul><li>Market Cap Score: 95.95</li><li>Value Score: 35.69</li><li>Momentum Score: 16.2</li><li>Growth Score: 50.77</li><li>Quality Score: 20.53</li><li>Crypto Native Score: 46.38</li><li>Tokenomics Score: 52.55</li><li>Growth-Adjusted Value Score: 64.8</li><li>Value Momentum Score: 25.94</li><li>Fundamental Price Divergence Score: 22.7</li><li>Value Growth Score: 43.23</li><li>Market Power Score: 33.48</li><li>Blue Chip Score: 28.11</li><li>Price Momentum Score: 33.48</li><li>Total Score: 33.91</li></ul><p>Investment Recommendation: Based on the quantitative analysis and fundamental scores, we recommend a BUY position for Polygon (MATIC). The current price is $0.6152, and the target price is $0.7706, indicating a potential annual return of 25.26%. The fundamental analysis suggests that Polygon is undervalued, with a relatively high value score of 35.69.</p><p>The market cap and token trading volume indicate strong market activity, with scores of 95.0 and 97.0, respectively. The large number of tokenholders (36,940,217) and active user base (318,030 DAU) demonstrate growing adoption and usage of the Polygon platform.</p><p>However, it is important to note that Polygon currently has negative earnings (-$274,134.14), which could be a potential risk factor. Further analysis of Polygon’s financial performance and revenue generation is recommended.</p><p>Sources:</p><ul><li>Polygon: <a href="https://polygon.technology/">https://polygon.technology/</a></li></ul><h3>Crypto Fundamental Analyst Report — Filecoin (FIL)</h3><h4>Token Information:</h4><ul><li>Name: Filecoin</li><li>Symbol: FIL</li><li>Description: FIL holders own and govern the protocol. Filecoin is a smart contract platform and a data storage network. Founded by Juan Benet in 2014.</li><li>Sectors: Blockchains (L1)</li><li>Chains: Filecoin</li><li>Website: <a href="https://filecoin.io/">Filecoin</a></li></ul><h4>Investment Opinion:</h4><ul><li>Current Price: $3.4911</li><li>Target Price: $2.8913</li><li>Annual Target Return: -17.18%</li><li>Investment Opinion: SELL</li></ul><h4>Quantitative Data Analysis:</h4><ul><li>Price: $3.49 (Score: 75.0)</li><li>Market Cap (Fully Diluted): Not available</li><li>Market Cap (Circulating): $1,499,134,043.78 (Score: 90.0)</li><li>Token Trading Volume: $474,402,019.19 (Score: 92.0)</li><li>Token Turnover (Fully Diluted): Not available</li><li>Token Turnover (Circulating): 0.32 (Score: 92.0)</li><li>Tokenholders: Not available</li><li>Fees: $0.0 (Score: 14.0)</li><li>Fees (Supply Side): $0.0 (Score: 21.0)</li><li>Revenue: $0.0 (Score: 21.0)</li><li>Token Incentives: $0.0 (Score: 26.0)</li><li>Earnings: $0.0 (Score: 41.0)</li><li>Active Developers: 73 (Score: 96.0)</li><li>Code Commits: 0 (Score: 40.0)</li><li>User DAU: Not available</li><li>User WAU: Not available</li><li>User MAU: Not available</li></ul><h4>Fundamental Scores:</h4><ul><li>Market Cap Score: 90.17</li><li>Value Score: 25.96</li><li>Momentum Score: 25.93</li><li>Growth Score: 20.65</li><li>Quality Score: 21.65</li><li>Crypto Native Score: 18.68</li><li>Tokenomics Score: 39.0</li><li>Growth-Adjusted Value Score: 35.08</li><li>Value Momentum Score: 25.95</li><li>Fundamental Price Divergence Score: 10.58</li><li>Value Growth Score: 23.31</li><li>Market Power Score: 23.29</li><li>Blue Chip Score: 23.81</li><li>Price Momentum Score: 23.29</li><li>Total Score: 22.58</li></ul><p>Investment Recommendation: Based on the quantitative analysis and fundamental scores, we recommend a SELL position for Filecoin (FIL). The current price is $3.4911, and the target price is $2.8913, indicating a potential annual return of -17.18%. The fundamental analysis suggests that Filecoin is overvalued, with a relatively low value score of 25.96.</p><p>The market cap and token trading volume indicate moderate market activity, with scores of 90.0 and 92.0, respectively. However, the lack of available data on fees, revenue, earnings, and user metrics raises concerns about the financial performance and adoption of the Filecoin platform.</p><p>Sources:</p><ul><li>Filecoin: <a href="https://filecoin.io/">https://filecoin.io/</a></li></ul><h3>Crypto Fundamental Analyst Report — NEAR Protocol (NEAR)</h3><h4>Token Information:</h4><ul><li>Name: NEAR Protocol</li><li>Symbol: NEAR</li><li>Description: NEAR holders own and govern the protocol. NEAR Protocol is a smart contract platform for decentralized applications. Founded by Alexander Skidanov and Ilya Polosukhin in 2017.</li><li>Sectors: Blockchains (L1)</li><li>Chains: NEAR Protocol</li><li>Website: <a href="https://near.org/">NEAR Protocol</a></li></ul><h4>Investment Opinion:</h4><ul><li>Current Price: $1.2045</li><li>Target Price: $0.8288</li><li>Annual Target Return: -31.19%</li><li>Investment Opinion: SELL</li></ul><h4>Quantitative Data Analysis:</h4><ul><li>Price: $1.2 (Score: 63.0)</li><li>Market Cap (Fully Diluted): Not available</li><li>Market Cap (Circulating): $1,102,940,384.63 (Score: 88.0)</li><li>Token Trading Volume: $137,288,944.69 (Score: 87.0)</li><li>Token Turnover (Fully Diluted): Not available</li><li>Token Turnover (Circulating): 0.12 (Score: 67.0)</li><li>Tokenholders: Not available</li><li>Fees: $0.0 (Score: 14.0)</li><li>Fees (Supply Side): $0.0 (Score: 21.0)</li><li>Revenue: $0.0 (Score: 21.0)</li><li>Token Incentives: Not available</li><li>Earnings: Not available</li><li>Active Developers: 72 (Score: 95.0)</li><li>Code Commits: 0 (Score: 40.0)</li><li>User DAU: 76,359 (Score: 95.0)</li><li>User WAU: Not available</li><li>User MAU: Not available</li></ul><h4>Fundamental Scores:</h4><ul><li>Market Cap Score: 88.44</li><li>Value Score: 18.77</li><li>Momentum Score: 18.94</li><li>Growth Score: 17.39</li><li>Quality Score: 21.46</li><li>Crypto Native Score: 16.67</li><li>Tokenomics Score: 40.29</li><li>Growth-Adjusted Value Score: 37.09</li><li>Value Momentum Score: 18.86</li><li>Fundamental Price Divergence Score: 9.67</li><li>Value Growth Score: 18.08</li><li>Market Power Score: 18.16</li><li>Blue Chip Score: 20.12</li><li>Price Momentum Score: 18.16</li><li>Total Score: 18.65</li></ul><p>Investment Recommendation: Based on the quantitative analysis and fundamental scores, we recommend a SELL position for NEAR Protocol (NEAR). The current price is $1.2045, and the target price is $0.8288, indicating a potential annual return of -31.19%. The fundamental analysis suggests that NEAR Protocol is overvalued, with a relatively low value score of 18.77.</p><p>The market cap and token trading volume indicate moderate market activity, with scores of 88.0 and 87.0, respectively. However, the lack of available data on fees, revenue, earnings, and token incentives raises concerns about the financial performance and adoption of the NEAR Protocol.</p><p>Sources:</p><ul><li>NEAR Protocol: <a href="https://near.org/">https://near.org/</a></li></ul><h3>Crypto Fundamental Analyst Report — Cosmos Hub (ATOM)</h3><h4>Token Information:</h4><ul><li>Name: Cosmos Hub</li><li>Symbol: ATOM</li><li>Description: ATOM holders own and govern the protocol. Cosmos Hub is an interoperability and security platform for blockchains and applications. Founded by Ethan Buchman and Jae Kwon in 2019.</li><li>Sectors: Blockchains (L1)</li><li>Chains: Cosmos</li><li>Website: <a href="https://cosmos.network/">Cosmos Hub</a></li></ul><h4>Investment Opinion:</h4><ul><li>Current Price: $8.2511</li><li>Target Price: $4.8429</li><li>Annual Target Return: -41.31%</li><li>Investment Opinion: SELL</li></ul><h4>Quantitative Data Analysis:</h4><ul><li>Price: $8.25 (Score: 85.0)</li><li>Market Cap (Fully Diluted): Not available</li><li>Market Cap (Circulating): $2,414,010,544.45 (Score: 93.0)</li><li>Token Trading Volume: $210,000,386.71 (Score: 89.0)</li><li>Token Turnover (Fully Diluted): Not available</li><li>Token Turnover (Circulating): 0.09 (Score: 56.0)</li><li>Tokenholders: Not available</li><li>Fees: Not available</li><li>Fees (Supply Side): Not available</li><li>Revenue: Not available</li><li>Token Incentives: Not available</li><li>Earnings: Not available</li><li>Active Developers: 173 (Score: 100.0)</li><li>Code Commits: 10 (Score: 96.0)</li><li>User DAU (Daily Active Users): 21,780 (Score: 91.0)</li><li>User WAU (Weekly Active Users): 120,188 (Score: 93.0)</li><li>User MAU (Monthly Active Users): 229,907 (Score: 92.0)</li></ul><h4>Fundamental Scores:</h4><ul><li>Market Cap Score: 93.06</li><li>Value Score: 8.32</li><li>Momentum Score: 17.23</li><li>Growth Score: 38.17</li><li>Quality Score: 0.0</li><li>Crypto Native Score: 16.31</li><li>Tokenomics Score: 40.29</li><li>Growth-Adjusted Value Score: 26.54</li><li>Value Momentum Score: 12.77</li><li>Fundamental Price Divergence Score: 7.31</li><li>Value Growth Score: 23.24</li><li>Market Power Score: 27.7</li><li>Blue Chip Score: 4.16</li><li>Price Momentum Score: 27.7</li><li>Total Score: 16.01</li></ul><p>Investment Recommendation: Based on the quantitative analysis and fundamental scores, we recommend a SELL position for Cosmos Hub (ATOM). The current price is $8.2511, and the target price is $4.8429, indicating a potential annual return of -41.31%. The fundamental analysis suggests that Cosmos Hub is overvalued, with a relatively low value score of 8.32.</p><p>The market cap and token trading volume indicate high market activity, with scores of 93.0 and 89.0, respectively. However, the lack of available data on fees, revenue, and earnings raises concerns about the financial performance and revenue generation of the Cosmos Hub ecosystem.</p><p>Sources:</p><ul><li>Cosmos Hub: <a href="https://cosmos.network/">https://cosmos.network/</a></li></ul><h3>Disclaimer</h3><p>Disclaimer: This report should not be considered as financial advice. Crypto investments carry inherent risks, and it is important to conduct thorough research and seek professional advice before making any investment decisions.</p><p>Note the data above can be wrong for few technological reasons of LLM, but we will figure it out. Outputs are almost straightly out from the LLM.</p><p>Thanks for watching!</p><p>Silentist Team.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b675d1b67753" width="1" height="1" alt=""><hr><p><a href="https://medium.com/silentist-research/factogpt-analyzes-securitized-tokens-b675d1b67753">FactoGPT Analyzes: SECuritized Tokens</a> was originally published in <a href="https://medium.com/silentist-research">Silentist Research</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>