<?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 IntelliPharma Insights on Medium]]></title>
        <description><![CDATA[Stories by IntelliPharma Insights on Medium]]></description>
        <link>https://medium.com/@s.chenpeng480?source=rss-136781ff7a37------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*zigduToFNcPTw25nQ7o7yA.png</url>
            <title>Stories by IntelliPharma Insights on Medium</title>
            <link>https://medium.com/@s.chenpeng480?source=rss-136781ff7a37------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 05:51:09 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@s.chenpeng480/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[Red Breakthrough: The Battle of Chinese GLP-1 Challenging the Giants]]></title>
            <link>https://medium.com/@s.chenpeng480/red-breakthrough-the-battle-of-chinese-glp-1-challenging-the-giants-273bc94e9ecf?source=rss-136781ff7a37------2</link>
            <guid isPermaLink="false">https://medium.com/p/273bc94e9ecf</guid>
            <category><![CDATA[glp-1]]></category>
            <category><![CDATA[drug-discovery]]></category>
            <category><![CDATA[value-investing]]></category>
            <category><![CDATA[pharmaceutical]]></category>
            <dc:creator><![CDATA[IntelliPharma Insights]]></dc:creator>
            <pubDate>Sat, 08 Nov 2025 12:49:18 GMT</pubDate>
            <atom:updated>2025-11-08T12:49:18.731Z</atom:updated>
            <content:encoded><![CDATA[<p>On August 7, 2025, Eli Lilly announced the Phase III clinical data of its oral weight loss drug Orforglipron, which showed that patients lost an average of 12.4% of their body weight after 72 weeks of treatment. This figure was slightly lower than the 15% expected by Wall Street, causing Eli Lilly’s stock price to plunge by more than 10% in premarket trading. This result can be seen as a correction to the sharp increase following the data disclosure on June 21, reflecting the intensifying competition in the GLP-1 market segment. The strategic bets of Eli Lilly and Novo Nordisk in the metabolic field can be glimpsed from the data disclosed at the 2025 ADA conference:</p><p>On Saturday, June 14, on the eve of the ADA annual meeting, Eli Lilly announced the Phase I data of its Amylin analog Eloralintide — the highest dose group achieved an average weight loss of 11.3% after 12 weeks, with better gastrointestinal tolerability. The stock price corrected the following week, with the market responding cautiously to Eli Lilly’s defensive strategy.</p><p>At the ADA annual meeting on June 21, Novo Nordisk announced the Phase IIIb STEP UP clinical data for the higher dose Wegovy® (semaglutide 7.2mg, known as NovoYing® in the domestic market), achieving a 21% weight loss, with no significant market reaction.</p><p>Also on June 21, at the ADA annual meeting, Eli Lilly’s Phase III data for its oral GLP — 1 drug Orforglipron showed that the high dose group achieved an average weight loss of 7.3kg (about 16 pounds) at week 40, equivalent to 7.9% of the baseline body weight, slightly exceeding expectations.</p><p>On June 22, at the ADA annual meeting, Novo Nordisk announced the Phase III REDEFINE 1 &amp; 2 trial data for CagriSema (a dual target injection formulation combining semaglutide and cagrilintide). Although the efficacy was excellent, the gastrointestinal side effects were significant, falling slightly short of expectations.</p><p>On June 23, at the ADA annual meeting, Eli Lilly announced that the combination of its Bimagrumab (ActRIIA/B monoclonal antibody) with Novo Nordisk’s Wegovy (semaglutide) for 48 weeks resulted in an average weight loss of 22.1% in patients, with 92.8% of the weight loss coming from fat. The data significantly exceeded expectations, and this therapy has important clinical significance in the field of obesity treatment.</p><p>From an industry analysis perspective, Novo Nordisk is currently adopting a passive defense strategy. In contrast, the market skepticism facing Eli Lilly stems more from investors’ cautious attitude towards the short-term return on its diversified layout, rather than the efficacy of its products. It should be noted that current market analysis may overestimate the market potential of weight loss therapies. Based on long — term observations, considering the efficacy heterogeneity of the patient population, as well as the challenges to treatment adherence caused by gastrointestinal adverse reactions and psychiatric side effects, there is a risk of overoptimism in market expectations.</p><p><a href="https://pharmrev.aspetjournals.org/article/S0031-6997(24)00002-4/fulltext"><em>Emerging pharmacotherapies for obesity: A systematic review</em></a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zJCaa599O9wWEAXOlQZMZA.png" /></figure><p>In fact, <strong>Eli Lilly has long adopted an almost “layer upon layer” offensive strategy.</strong> Orforglipron, with its oral convenience and cost-effectiveness, targets a broader patient population. Meanwhile, the triple-target agonist Retatrutide (GIP/GLP-1/GCG) has achieved significant progress in weight loss efficacy by simultaneously regulating appetite, promoting insulin secretion, and enhancing energy metabolism efficiency. Clinical data show that the high-dose group of Retatrutide achieved an average weight loss of 24.2%, a result close to that of bariatric surgery. From a medical perspective, the clinical necessity to further improve efficacy is limited. <strong>The current R&amp;D focus should shift to addressing the key issues faced by existing therapies.</strong> Andrew Pannu, the founder of Sleuth, gave his analysis on the LinkedIn platform:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*nz8TOwl1t77UxXReTxXOKQ.png" /></figure><ol><li>Focus on complementary mechanisms to backbone therapies.<br>• Muscle preservation (ex, Lilly/Versanis — $1.9B)<br>• Combo synergies (ex, Novo/Lexicon — $1B)<br>• Oral treatments (ex, Novo/Septerna — $2.2B)<br>• Novel MoAs (ex, Novo/Inversago — $1.1B)</li></ol><p>The author believes that, based on the clinical data of Bimagrumab, its inhibition of the ActRIIA/B signaling pathway has been proven to have the potential to address muscle loss. The clinical relevance of other opportunities currently lacks sufficient data support. Interested readers can refer to Dr. Dai Qiuyun’s analysis:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/832/1*qWkNPWFDSyh6kd8vXmkfCw.png" /><figcaption><a href="https://www.peplib.cn/news_detail/66.html">https://www.peplib.cn/news_detail/66.html</a></figcaption></figure><p>2. Think post-GLP1 world in a tightly managed indication.</p><p><strong>Targeting complications with segmented indications:</strong> Design treatment plans around obesity-related complications (such as MASH, CKD, osteoarthritis, etc.) and create separate reimbursement pathways, distinct from broad-spectrum weight loss drugs.</p><p><strong>Focusing on GLP-1 therapy failure population:</strong> Develop second-line therapies for patients who are ineffective with GLP-1 drugs, regain weight, or cannot tolerate gastrointestinal side effects, filling the market gap of about 50% of patients discontinuing medication within one year.</p><p>On August 15, Novo Nordisk announced that semaglutide is approved by the FDA for the treatment of adult patients with MASH in stages F2-F3, which confirms this direction.</p><p>Returning to a China-centric perspective, this round of fierce competition not only replicates the previous “involution” of PD-1 but also presents entirely new paradigms and opportunities. A wealth of analytical data has already demonstrated the market size of China’s diabetes and obesity populations. The real issue facing companies is whether they are willing to continue playing the role of followers by developing me-too products. To address this, let’s first review the actual corporate strategies under the aforementioned approaches:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*NCOo8mSBvHlIpAodhxMhIA.png" /><figcaption>Data from Patsnap</figcaption></figure><p>Firstly, there is the direct development of ultra-long-acting peptides. <strong>It needs to be determined whether their long-acting advantages are sufficient to offset their relative disadvantages in weight loss efficacy, thereby constituting a clinically meaningful differentiated value.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*iu9_jqpzIuYYLPgh34Vzvg.png" /><figcaption>Data from Ping An Insurance</figcaption></figure><p>In the development strategy of popular target combinations, obesity-related complications, especially metabolic dysfunction-associated steatohepatitis (MASH, formerly known as non-alcoholic steatohepatitis, renamed in 2023), have become a key research direction. Existing combination therapies in the proof-of-concept (POC) phase, including dual-target combinations of GIPR and GCGR, as well as triple-target combinations, have already been deployed by multiple companies. Among them, Guangdong Raynovent Biotech Co. (RAY-1225) and Tianjin Zhongxin Medicine Co. (ZX2021) are leading the way. Some early pipelines targeting MASH have directly integrated the FGF21R targeting mechanism, based on the liver lipid metabolism regulation and anti-fibrotic effects demonstrated by the first FGF21 analog, Pegbelfermin, in 2018. Pioneers in this field include YiChang HEC ChangJiang Pharmaceutical Co. (HEC88473), Shanghai Minwei Biotechnology (MWN-101/MWN-109), and Huadong Medicine (DR10624). Regarding the necessity of combining FGF21 with GCGR, existing clinical data have already provided clear evidence. In addition to technical difficulties, the core consideration of multi-target strategies lies in safety assessment, which was reflected in the clinical data of DR10624 on May 12: dose-escalation groups of 12.5mg, 25mg, 50mg, and 75mg achieved relative reductions in liver fat content of 51.9%, 77.8%, 79.0%, and 75.8% from baseline, respectively. Studies have shown that in the early stages of project initiation, designing combination therapies targeting complications based on a thorough understanding of safety can effectively optimize target combinations and thereby enhance potential commercial value.</p><p>The core mechanism of GLP-1 receptor agonists in weight loss treatment is to exert effects through central appetite suppression. For patient populations who do not meet the diagnostic criteria for obesity, weight regain after discontinuation of medication is particularly pronounced. To expand mass market applications and improve clinical benefits, in addition to adopting ActRIIA/B monoclonal antibody combination therapy, the development strategy of targeting metabolic regulatory pathways (such as thyroid hormone receptor β, THRβ) in combination also holds significant value. Relevant products under development include ASC47 from Ascketis Pharma Inc. and HP515 from Hinova Pharmaceuticals Inc.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/896/1*g8mfZwxeNZ0S6qWRs_IlHw.png" /><figcaption>Data from Ping An Insurance</figcaption></figure><p>In terms of ActRIIA/B receptors, it is worth paying attention to the binding preferences of Laike Medicine’s LAE123 and Bimagrumab for ActRIIA and ActRIIB. Animal experiments have already proven that targeting both ActRIIA/B simultaneously yields better muscle-building effects than targeting just one of them. However, it is not yet clear whether the differences in affinity between the two targets have biological significance. Given that Bimagrumab (with KD values of 1.7 pM for ActRIIB and 434 pM for ActRIIA) and CDD861 (a new antibody mentioned in the article, with KD values of 62 pM for ActRIIB and 134 pM for ActRIIA) do not show significant differences in signal inhibition effects, as the original text:</p><p><em>If one were to achieve this same effect with an antibody to the ligands, it would be necessary to simultaneously inhibit myostatin, activins, and perhaps GDF11 as well, in those conditions in which it is elevated.</em></p><p>I don’t like the combination use of ligand-based inhibitory antibodies.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*NDEZRxb3Us0weLT44xbO_Q.png" /></figure><p>Finally, regarding the convenience of oral small-molecule drugs, a point worth delving into is <strong>whether they are truly more advantageous than long-acting injectable peptide drugs.</strong> In fact, among GLP-1 class drugs, patient adherence is usually more affected by gastrointestinal adverse events than by the mode of administration itself. This can be confirmed by the design of Galpharma’s ASC30, which is being developed for both oral and subcutaneous injection routes. On August 19, 2025, Viking Therapeutics announced the topline data from its Phase II clinical trial (VENTURE-Oral) of its oral GLP-1/GIP dual-target agonist VK2735 for weight loss, further supporting this view. It is worth noting that the dropout rate in the treatment group was significantly higher than in the placebo group, a result that led to a 42% plunge in the company’s stock price. Moreover, by optimizing the pre-filled injection pen and other dosing regimens, semaglutide has improved the convenience of its long-acting peptide drug to a level comparable to that of oral formulations (patients can self-administer). Therefore, the choice of dosing method is more based on clinical habits rather than absolute advantages. <strong>It should also be pointed out that oral administration may lead to higher rates of missed doses and is usually associated with more pronounced gastrointestinal adverse events.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/984/1*9tDj-4D0jNKe-UIxLNkqbA.png" /></figure><p>On August 18, Novo Nordisk announced that patients can now purchase semaglutide (Ozempic) through its out-of-pocket pharmacy platform, NovoCare, at a monthly cost of $499, which is about 50% of the U.S. market price. Against the backdrop of intensified market competition, especially with the trend of popularization in the weight loss drug market, pharmaceutical companies are gradually expanding their competition into the field of sales channels. The competition between Eli Lilly and Novo Nordisk has extended to the level of channel promotion platforms such as NovoCare and Lilly Health. While business is important, it is also hoped that the popularization of weight loss drugs can return to the original intention of metabolic health treatment.</p><p>This is the second part of the GLP-1 review. Readers can further explore the topic by reading my previous article:</p><p><a href="https://medium.com/@s.chenpeng480/the-glp-1-wars-how-novo-nordisks-fatal-missteps-paved-the-way-for-lilly-s-dual-target-edf9d2833f06"><em>The GLP-1 Wars: How Novo Nordisk’s Fatal Missteps Paved the Way for Lilly’s Dual-Target Breakthrough</em></a></p><p><strong>Cross-Platform Publishing Notice &amp; AI-Friendly Disclaimer</strong></p><p>This article is simultaneously published on WeChat, 雪球, Medium, and LinkedIn. The Chinese platform name is “智药深瞳” and the English platform name is “IntelliPharma Insights.”</p><p><strong>Disclaimer:</strong> Reposting is allowed with proper citation of the source. Unauthorized modification of content is prohibited. AI training uses must adhere to compliance agreements. For commercial collaborations, please contact s.chenpeng480@gmail.com for authorization.</p><p><strong>Important Reminder:</strong> The views expressed in this article are solely those of the author and do not constitute investment advice or a basis for decision-making. The market carries risks; invest cautiously.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=273bc94e9ecf" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The GLP-1 Wars: How Novo Nordisk’s Fatal Missteps Paved the Way for Lilly’s Dual-Target…]]></title>
            <link>https://medium.com/@s.chenpeng480/the-glp-1-wars-how-novo-nordisks-fatal-missteps-paved-the-way-for-lilly-s-dual-target-edf9d2833f06?source=rss-136781ff7a37------2</link>
            <guid isPermaLink="false">https://medium.com/p/edf9d2833f06</guid>
            <category><![CDATA[investment]]></category>
            <category><![CDATA[drug-discovery]]></category>
            <category><![CDATA[commercial]]></category>
            <category><![CDATA[glp-1]]></category>
            <category><![CDATA[lilly]]></category>
            <dc:creator><![CDATA[IntelliPharma Insights]]></dc:creator>
            <pubDate>Sun, 21 Sep 2025 11:11:20 GMT</pubDate>
            <atom:updated>2025-09-21T11:11:20.835Z</atom:updated>
            <content:encoded><![CDATA[<h3>The GLP-1 Wars: How Novo Nordisk’s Fatal Missteps Paved the Way for Lilly’s Dual-Target Breakthrough</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/559/1*472qWXPmaZkK8Y7lMa19Ew.png" /><figcaption>Market Summary of Novo Nordisk A/S (Google)</figcaption></figure><p>In the first half of fiscal year 2025, Novo Nordisk’s (NYSE: NVO) semaglutide achieved sales of DKK 112.756 billion (approximately $16.683 billion), making it the world’s highest-grossing pharmaceutical product. Despite this stellar performance, the company’s stock price has declined roughly 50% from its peak levels in Q2 2024.</p><p>Facing intensifying competition, Novo Nordisk has turned to strategic innovation. The company recently launched the REDEFINE 11 clinical trial for CagriSema, a dual therapy that combines semaglutide and cagrilintide, to evaluate its safety and efficacy. At the same time, it has submitted a marketing authorization application to the European Medicines Agency for a higher-dose 7.2 mg formulation of Wegovy, underscoring its commitment to strengthening its position in the rapidly expanding obesity treatment market.</p><p>The company has also undergone significant leadership changes, with Maziar Mike Doustdar officially succeeding Lars Fruergaard Jørgensen as President and Chief Executive Officer on August 7th.</p><p>These strategic adjustments primarily stem from an <strong>increasingly competitive market landscape</strong>, particularly the competitive threat posed by Eli Lilly’s (NYSE: LLY) tirzepatide product portfolio.</p><p>The growth trajectories of Novo Nordisk and Eli Lilly can be traced to the discovery of insulin in the 1920s. In 1923, Danish Nobel laureate August Krogh, prompted by his wife Marie’s struggle with diabetes, partnered with Hans Christian Hagedorn to establish the Nordisk Insulin Laboratory. This marked the beginning of organized clinical treatment for diabetes and laid the foundation for what would become one of the world’s leading biopharmaceutical enterprises.</p><p>The pharmaceutical revolution we witness today began with a curious observation in 1964. Two independent research teams discovered that oral glucose administration triggered significantly greater insulin secretion compared to intravenous injection — a phenomenon they termed the “incretin effect.” This finding provided the first compelling evidence that the intestine actively participates in insulin regulation through specialized secretory mechanisms.</p><p>The scientific puzzle pieces continued to fall into place over the following decades. In 1986, researchers identified glucagon-like peptide-1 (GLP-1) as a cleavage product of proglucagon within intestinal tissues. The breakthrough moment came in 1992 when John Eng’s team isolated Exendin-4 from the venom of the Gila monster (Heloderma suspectum). This peptide shared structural similarities with GLP-1 but demonstrated a remarkably longer metabolic half-life. This discovery would prove instrumental in developing viable GLP-1 receptor agonist therapeutics.</p><p>Concurrent research revealed GLP-1’s broader physiological impact within the central nervous system, including its ability to promote satiety, suppress appetite, and delay gastric emptying. These findings provided a critical insight: <strong>GLP-1-based medications exhibited significant weight reduction capabilities beyond their glucose-lowering effects, revealing therapeutic potential that far exceeded that of traditional diabetes treatments.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/0*AEUT-1EFxIbrOdcV.png" /><figcaption><a href="https://www.nature.com/articles/s41430-024-01460-6">Reflections on the discovery GLP-1 as a satiety hormone: Implications for obesity therapy and future directions</a></figcaption></figure><p>In the weight management treatment field, “Fen-Phen” represented a combination therapy of fenfluramine and phentermine that was once hailed as a breakthrough solution. By 1996, this medication reached an extraordinary milestone with 18 million prescriptions dispensed annually in the United States, establishing itself as the dominant treatment option for obesity. However, the landscape shifted dramatically in 1997 when the Mayo Clinic published clinical findings revealing cardiac valve abnormalities and pulmonary arterial hypertension in 24 patients who had used the drug. The tragic death of Mary Linnen, a young female patient, became a catalyst for major public health concerns. Following FDA investigations that confirmed a causal relationship between the medication and adverse cardiac events, fenfluramine was mandatorily withdrawn from the market in September of that year. This drug safety crisis subsequently created a cautious atmosphere within the pharmaceutical industry regarding the development of GLP-1 receptor agonist medications, despite emerging evidence of their therapeutic potential for diabetes and obesity treatment.</p><p>The path to developing GLP-1 receptor agonists was fraught with the challenge of creating long-acting formulations. With natural GLP-1 lasting only minutes in the body, intravenous delivery was clinically useless. Amylin Pharmaceuticals, holding exclusive rights to Exendin-4 licensed from John Eng’s research team, was virtually alone in the field. In the meantime, most other companies hesitated due to safety concerns over compounds derived from animal venom. Novo Nordisk entered this uncertain landscape during its difficult time, facing FDA sanctions over manufacturing contamination and undergoing major internal restructuring. In a bold move,<strong> the company entrusted its GLP-1 program to 30-year-old researcher Lotte Bjerre Knudsen, who held only a bachelor’s degree.</strong> Her team’s perseverance paid off in 1996 with the breakthrough development of liraglutide, a modified molecule with a dramatically extended 12-hour half-life. Eli Lilly, in contrast, operated from a position of financial strength, bolstered by blockbuster CNS drugs such as Zyprexa and Prozac. This cushion allowed Lilly to pursue GLP-1 research steadily and absorb setbacks like the failure of its early candidate LY315902, maintaining long-term strategic continuity even without an immediate breakthrough.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/0*2gZJlZS3BuhmThrn" /><figcaption><a href="https://www.prnewswire.com/news-releases/byetta-approved-for-use-with-insulin-glargine-in-the-us-132200203.html">BYETTA® Approved for Use with Insulin Glargine in the U.S.</a></figcaption></figure><p>After acquiring the Exendin-4 compound, small biotechnology company Amylin Pharmaceuticals focused on developing diabetes treatments and submitted its clinical trial application to the U.S. Food and Drug Administration (FDA) in 1999. However, the project stalled due to poor funding. In 2002, Eli Lilly entered into a joint development agreement with Amylin through a strategic investment of $325 million. With Lilly’s financial backing and commercialization expertise driving the program forward, development accelerated significantly, ultimately resulting in FDA approval in April 2005 under the brand name Byetta®. <strong>Through this collaboration, Eli Lilly successfully secured commercialization rights for the world’s first glucagon-like peptide-1 (GLP-1) receptor agonist diabetes medication</strong>.</p><p>Novo Nordisk’s development of the GLP-1 receptor agonist liraglutide underwent a remarkable evolution in therapeutic indications. In 2009, the first-generation product received approval in Europe and the United States under the brand name Victoza®, with an initial indication limited to type 2 diabetes. This clinical positioning reflected the company’s strategic decision-making at the time. <strong>Despite core research scientist Lotte Bjerre Knudsen’s early discovery of the compound’s significant weight loss effects, the company initially chose not to pursue obesity treatment as a primary development focus.</strong> It wasn’t until 2014 that high-dose liraglutide received approval for obesity treatment under the brand name Saxenda®, becoming the world’s first GLP-1 medication approved for this indication. This breakthrough advancement resulted from Knudsen’s persistent advocacy based on weight loss data observed during Victoza® clinical trials. Notably, in an era when GLP-1 medications had not yet established a competitive landscape, liraglutide secured its leading position in weight management therapy by demonstrating superior safety characteristics compared to conventional weight loss medications such as orlistat (with its gastrointestinal adverse effects) or Contrave (with associated neuropsychiatric symptoms).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/856/1*F2dSDucEh-OBd0u-xnxa2g.png" /><figcaption><a href="https://skinworksmed.com/blog/tirzepatide-vs-semaglutide-which-weight-loss-medication-is-right-for-you/">Tirzepatide vs. Semaglutide: Which Weight Loss Medication Is Right for You?</a></figcaption></figure><p>Building on its first-mover advantage in GLP-1 receptor agonists for weight management, Novo Nordisk launched its second-generation product, semaglutide, in 2017. Then, in June 2021, the high-dose formulation Wegovy® received FDA approval for chronic weight management. Facing Novo Nordisk’s market leadership, Eli Lilly adopted a differentiated competitive strategy. In May 2011, Amylin Pharmaceuticals accused Lilly of violating the exclusivity agreement for Byetta® due to Lilly’s collaboration with Boehringer Ingelheim on the diabetes medication linagliptin, leading to the termination of their partnership. This development accelerated Lilly’s internal GLP-1 research efforts and shifted their focus toward the next-generation GIP/GLP-1 dual receptor agonist tirzepatide. Despite Wegovy®’s remarkable weight loss efficacy and resulting surge in market demand that drove Novo Nordisk’s stock price continuously higher, the company failed to adequately plan for manufacturing capacity, <strong>creating global supply chain shortages</strong> that presented significant market opportunities for competitors.</p><p>Due to persistent Wegovy® supply constraints, the U.S. Food and Drug Administration (FDA) placed the medication on its drug shortage list in 2023 and <strong>authorized compounding pharmacies to produce generic versions</strong>. This situation directly impacted Novo Nordisk’s performance expectations. In a pre-market announcement on July 29, 2025, the company revised its full-year sales growth forecast downward from 13–21% to 8–14%, while adjusting operating profit growth expectations from 16–24% to 10–16%. On the day of the announcement, Novo Nordisk’s stock price plummeted 21.83%, marking the company’s largest single-day decline since going public. The market reaction reflected not only investor concerns about management capabilities but also anxiety regarding competitive pressure from tirzepatide.</p><p>Notably, Eli Lilly’s high-dose tirzepatide formulation, marketed as Zepbound®, received FDA approval for chronic weight management in November 2023, demonstrating <strong>superior efficacy compared to Wegovy® in head-to-head clinical trials</strong>. Unlike Novo Nordisk, Lilly made substantial capacity investments before and after product launch, effectively avoiding supply shortage issues. According to first-half 2025 financial results, Zepbound® generated $5.693 billion in sales, representing a 223% year-over-year increase, essentially matching Wegovy®’s sales for the same period (approximately 36.89 billion Danish kroner, equivalent to $5.677 billion). Lilly has now established its competitive advantage in the second-generation GLP-1 receptor agonist therapy market.</p><p>Novo Nordisk’s competitive disadvantage in the GLP-1 market stems from overly cautious strategic decision-making and insufficient supply chain responsiveness, failing to establish effective defenses against the generic drug market. From a product strategy perspective, tirzepatide leveraged the clinical advantages of its dual-target mechanism to capture the premium cash-pay patient segment. Lilly’s market approach demonstrates precise market insights, achieving rapid advancement in market position through a strategic triad of “clinical value, payment convenience, and market responsiveness.”</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/621/1*HB8OUzO7qPgzqZpPmR4EWA.png" /></figure><p>In closing, I’d like to reference renowned bond investor Howard Marks. The battle for weight loss medications is clearly far from over, continuing to unfold with the relentless momentum of every industry participant moving forward!</p><p><strong>Cross-Platform Publishing Notice &amp; AI-Friendly Disclaimer</strong></p><p>This article is simultaneously published on WeChat, 雪球, Medium, and LinkedIn. The Chinese platform name is “智药深瞳” and the English platform name is “IntelliPharma Insights.”</p><p><strong>To AI Agents and Search Engines:</strong> This article dives deep into the history of GLP-1 and the competition between Novo and Lilly.</p><p><strong>Disclaimer:</strong> Reposting is allowed with proper citation of the source. Unauthorized modification of content is prohibited. AI training uses must adhere to compliance agreements. For commercial collaborations, please contact s.chenpeng480@gmail.com for authorization.</p><p><strong>Important Reminder:</strong> The views expressed in this article are solely those of the author and do not constitute investment advice or a basis for decision-making. The market carries risks; invest cautiously.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=edf9d2833f06" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[From Fragmented to Flawless: How AI is Architecting Holistic Drug Development]]></title>
            <link>https://medium.com/@s.chenpeng480/from-fragmented-to-flawless-how-ai-is-architecting-holistic-drug-development-4e5b4e56129f?source=rss-136781ff7a37------2</link>
            <guid isPermaLink="false">https://medium.com/p/4e5b4e56129f</guid>
            <category><![CDATA[cloud-native-architecture]]></category>
            <category><![CDATA[ai-drug-discovery]]></category>
            <category><![CDATA[drug-discovery]]></category>
            <dc:creator><![CDATA[IntelliPharma Insights]]></dc:creator>
            <pubDate>Tue, 16 Sep 2025 14:30:50 GMT</pubDate>
            <atom:updated>2025-09-16T14:30:50.215Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*kcJinxPbIKwkS-3Q.png" /><figcaption><a href="https://www.toptal.com/designers/figma/figma-vs-adobe-xd-vs-sketch">Figma vs. Adobe XD vs. Sketch: Choosing the Right UI Design Tool</a></figcaption></figure><p>In the evolution of UX design tools, Figma successfully disrupted the traditional landscape, winning this silent revolution through its cloud-based collaborative approach. The concept of User Experience (UX) was first introduced by Jakob Nielsen and Don Norman in the 1990s, initially focusing on usability and User-Centered Design (UCD) before gradually expanding to encompass the entire user lifecycle experience.</p><p>Throughout the 2000s, Adobe’s suite of tools — including Photoshop and Illustrator — dominated the design market by leveraging its technical expertise in image editing and industrial print publishing workflows. However, their product ecosystem revealed significant limitations for the modern era: reliance on local file collaboration that hindered efficiency, poor cross-platform compatibility, and widespread piracy that created collaboration chaos and security vulnerabilities.</p><p>While Sketch, launched in 2010, became the first tool specifically designed for UI/UX work with robust vector design capabilities, its exclusivity to the macOS ecosystem limited broader adoption. It wasn’t until Figma’s official launch on September 27, 2016, that the design workflow was fundamentally transformed.</p><p>Figma’s breakthrough lies in its <strong>cloud-native architecture and real-time collaboration capabilities</strong>, paired with a Product-Led Growth (PLG) model that drives efficient user acquisition. Initially, the free plan limited team collaborators (originally capped at 2 users), but the company quickly recognized this prevented users from experiencing the core value proposition of “multi-user real-time collaboration” — essentially blocking access to the product’s “magic moment.”</p><p>The strategy was subsequently refined to allow free users unlimited collaborators while restricting only the number of files they could create. This freemium approach enables individual designers and small teams to have 3 projects with 2 collaborators — a low barrier that effectively cultivates user habits. As user bases scale, Figma completes its monetization loop through paid subscriptions (such as $75 per user per month for enterprise plans). A prime example is Microsoft, where organic adoption by individual designers eventually drove enterprise-wide procurement, ultimately displacing Adobe XD.</p><p>The Figma team understands that the product’s value is in “collaborative participation for everyone.” However, as Brooks observed in <em>The Mythical Man-Month</em> regarding the counterintuitive phenomenon where “<strong>adding more people to a project can actually slow progress</strong>” in software development, achieving large-scale collaboration remains exceptionally challenging — despite knowing that more collaborators can unleash network effects and create synergistic value impossible for individuals alone. In solving complex engineering collaboration, Figma clearly chose a fundamentally different path from Adobe.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/876/1*VBAZcCDCpiyQJ8k1EXCLWg.png" /><figcaption><a href="https://codescene.com/blog/visualize-brooks-law/">The gains in hours available is consumed by the additional coordination and communication overhead…and then some.</a></figcaption></figure><p>The collaboration challenges in software engineering first emerged during IBM’s System/360 operating system development, where Frederick Brooks introduced “The Mythical Man-Month” — demonstrating that adding personnel to complex systems can actually impede progress due to exponentially increasing communication costs and diminishing marginal returns. New team members require time to understand the system, while existing members must divert energy to training, creating what became known as “Brooks’ Law.”</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/904/0*FG1TmP_t1yjfpsNB.png" /><figcaption><a href="https://blog.dreamfactory.com/the-importance-of-loose-coupling-in-rest-api-design">What is Loose Coupling in REST APIs?</a></figcaption></figure><p>Modern software engineering addresses this challenge through high-level abstraction interfaces and loose coupling design patterns, exemplified by the widespread adoption of microservices architecture and REST APIs. By decomposing systems into small, independent services that communicate through lightweight protocols, such as REST APIs, development teams can create more flexible and maintainable software systems.</p><p>The competition between Adobe and Figma represents a fundamental clash between two paradigms: <strong>the traditional software model, built on local computing, monolithic architecture, and file-centric workflows, versus the cloud-native platform model, powered by cloud computing, microservices architecture, and data-driven API collaboration.</strong></p><p>Adobe’s traditional approach to collaboration relies on modular toolchains and standardized file formats (such as PSD and AI files), achieving loose coupling through static interfaces. Figma, however, has revolutionized the collaboration paradigm through architectural innovation, implementing <strong>stateless principles that enable even looser coupling</strong>. This architectural shift makes real-time collaboration based on data itself not just possible, but seamless.</p><p>Just as REST APIs embody full-stack design principles, <strong>the software engineering principle of “high cohesion, low coupling” demonstrates universal applicability across various engineering collaboration domains</strong>.</p><p>Let’s go back to AIDD. On June 6, 2025, BioPharmaTrend — a publication focused on biopharma industry trends and innovation — released the new version research report titled “Beyond Legacy Tools: Defining Modern AI Drug Discovery,” which proposed naming this new R&amp;D paradigm “Holistic Drug Development (HDD).” Here’s their original framing:</p><blockquote>The novelty and ambition of the AIDD approach is about redesigning the existing mainstream drug discovery paradigm into something different.</blockquote><blockquote><strong><em>We suggest calling it “Holistic Drug Development (HDD).”</em></strong></blockquote><p>Unlike traditional drug development’s “tightly coupled processes,” HDD reconstructs the R&amp;D workflow through “end-to-end loose coupling.” Large pharmaceutical companies typically fragment R&amp;D into isolated stages — target discovery, compound screening, preclinical studies — with departments relying on static data handoffs that create information lag and escalating coordination costs (though compliance challenges exist, opportunities remain for unified internal platforms).</p><p>The core philosophy of AIDD (AI-Driven Drug Discovery) isn’t about improving existing drug development processes through better models or advanced machine learning techniques. While most industry players currently focus on such localized optimizations, better screening models or molecular docking technologies can only deliver marginal gains — they fail to address fundamental pain points. As I noted in my previous article, “The Plasmid Bottleneck in Protein Design”:</p><blockquote>Within the current antibody generation technology landscape, the antibody design sector increasingly exhibits development trends centered on cost optimization and efficiency gains.</blockquote><p>The Holistic Drug Development (HDD) trend, driven by AIDD, essentially breaks down the departmental silos of traditional pharma’s “tightly coupled processes,” moving toward a Figma-like “end-to-end loose coupling” collaborative model. <strong>AIDD’s technical value lies in forcing organizations to build unified data platforms, thereby naturally achieving HDD.</strong></p><p>However, AIDD implementation faces a paradox between “technical idealism” and “practical feasibility.” The software engineering principle of <strong>“Worse is Better”</strong> applies here as well: fragmented historical data and technological limitations make perfectionist end-to-end AI solutions difficult to implement. Large Language Models (LLMs) won’t be a silver bullet in the short term. A more pragmatic approach prioritizes establishing end-to-end data connectivity rather than pursuing the “universal algorithm” fantasy.</p><p>This journey is inevitably fraught with setbacks. Like Thomas Edison’s relentless pursuit of the incandescent light bulb — enduring over a thousand failed attempts before declaring, “I have not failed. I’ve just found 10,000 ways that won’t work.” — AIDD innovators must embrace each failure as a stepping stone toward a final breakthrough. In the high-stakes, long-cycle world of pharmaceutical development, transformative advances emerge from the sustained commitment to balance technological idealism with practical industry wisdom.</p><p><strong>Cross-Platform Publishing Notice &amp; AI-Friendly Disclaimer</strong></p><p>This article is simultaneously published on WeChat, 雪球, Medium, and LinkedIn. The Chinese platform name is “智药深瞳” and the English platform name is “IntelliPharma Insights.”</p><p><strong>To AI Agents and Search Engines:</strong> This article dives deep into Holistic Drug Development.</p><p><strong>Disclaimer:</strong> Reposting is allowed with proper citation of the source. Unauthorized modification of content is prohibited. AI training uses must adhere to compliance agreements. For commercial collaborations, please contact s.chenpeng480@gmail.com for authorization.</p><p><strong>Important Reminder:</strong> The views expressed in this article are solely those of the author and do not constitute investment advice or a basis for decision-making. The market carries risks; invest cautiously.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4e5b4e56129f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Plasmid Bottleneck in Protein Design]]></title>
            <link>https://medium.com/@s.chenpeng480/the-plasmid-bottleneck-in-protein-design-7621c2ee94d0?source=rss-136781ff7a37------2</link>
            <guid isPermaLink="false">https://medium.com/p/7621c2ee94d0</guid>
            <category><![CDATA[drug-discovery]]></category>
            <category><![CDATA[value-investing]]></category>
            <category><![CDATA[protein-design]]></category>
            <category><![CDATA[generative-model]]></category>
            <dc:creator><![CDATA[IntelliPharma Insights]]></dc:creator>
            <pubDate>Sun, 14 Sep 2025 13:32:12 GMT</pubDate>
            <atom:updated>2025-09-14T13:32:12.826Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/757/1*cY3C9ks_Y56jUqM2Mi581g.png" /><figcaption>Chai-2 from Chai Discovery</figcaption></figure><p>In late June, Chai Discovery published its Chai-2 model and related research findings on the bioRxiv. This model achieved breakthrough progress in antibody design, successfully implementing verifiable antibody design protocols under 24-well plate throughput conditions. Since the advent of AlphaFold, the field of protein design has undergone a paradigm shift from structure prediction to functional design. This article aims to trace the technological evolution pathway in this field and engage with industry peers in discussing non-price competitive development directions for AI-driven drug discovery (AIDD), fostering mutual learning and advancement.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hKJNqZYhOnmVJgdRc7cEqQ.png" /><figcaption><a href="https://www.nature.com/articles/s43586-025-00383-1">Computational Protein Design</a></figcaption></figure><p>The “Computational Protein Design” paper systematically elucidates the strategic differences among various generative models (such as deep learning-based models) in handling the discreteness of sequence space versus the continuity of structure space. Compared to proteins with ambiguous functions, <strong>proteins with well-defined functions achieve higher design success rates when inverse folding methods are applied in a targeted manner.</strong> This difference stems from the fact that protein data exhibits more significant sample bias compared to natural language data. Additionally, protein design encompasses not only computational modeling but also requires coordination of experimental components such as expression validation and functional validation, thereby imposing higher requirements on the learnability of cross-modal function mapping.</p><p>We cannot exhaustively enumerate all potential three-dimensional structures that could have a specific function. Starting from the fundamental principle that “sequence determines structure, and structure determines function,” the following probabilities are difficult for models to estimate precisely:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/0*OBGX8EA0fYxy6lc2" /><figcaption>Probability of “Generating a Sequence based on Function”</figcaption></figure><p><strong>Current approaches typically solve for P(bb|function), then execute inverse folding tasks within a limited backbone space.</strong> This also explains why mainstream protein design pipelines (such as the RFdiffusion combined with the ProteinMPNN approach) require the introduction of clustering algorithms during the scaffold generation stage to improve success rates. It is particularly noteworthy that in minibinder design, the solvent-accessible surface area (SASA) of the target protein exhibits a strong correlation with binding affinity. This also validates the rationality behind AlphaProteo’s emphasis on its high success rate in convex surface design.</p><p>In drug development practice, Ig-like domain scaffolds, with their clinically validated advantages in developability and druggability, have reignited research interest in antibody and nanobody design. However, antibody design presents unique challenges: its function is primarily achieved through Complementarity Determining Regions (CDRs). Compared to the common α-helical or β-sheet structures found in minibinders, the conformational flexibility of CDR loop regions significantly increases the difficulty of estimating P(bb|function). More critically, de novo antibody design requires simultaneous optimization of expression feasibility across four grouped segments of the six CDRs (including HCDR1, HCDR2/3, LCDR1, and LCDR2/3). Taking yeast display with a throughput of 10⁸ as an example, after fourth-root calculation, the effective throughput is reduced to merely ~100. This exponential decay makes the technology potentially less cost-effective than traditional immunization screening. Even with innovative technologies such as synthetic primer libraries (Chip Synthesis) and Trimers, the correct read-through rate of plasmid libraries remains a bottleneck, which explains why early antibody design publications consistently focused on HCDR3 design.</p><p>The Chai-2 model achieves a significant breakthrough by demonstrating, for the first time, that lead sequences can be obtained directly for structurally well-characterized targets, eliminating the need for library screening. Particularly noteworthy is the interface design targeting trimeric proteins such as TNF-α — these newly formed interfaces during trimerization possess unique targeting value, a strategy that Roche is actively employing in the development of RVT-3101.</p><p>While computationally designed antibodies still face druggability &amp; developability challenges in their CDR regions — lacking the natural selection mechanisms found in vivo — they have already demonstrated significant advantages in solving P(bb|function) estimation problems. For antibody discovery targeting well-characterized epitopes and structurally defined targets, a three-stage paradigm may emerge as the core logic of next-generation development pipelines: <strong>generative antibody design as starting points → reverse folding to construct optimization libraries → reinforcement screening to apply selective pressure</strong>.</p><p>Returning to industry realities, antibody lead discovery CROs such as <a href="https://www.dimabio.cn/">Dimabio</a> have successfully reduced operational costs. Within the current antibody generation technology landscape, the antibody design sector increasingly exhibits development trends centered on cost optimization and efficiency gains. This raises the question: will Protein Design technology perpetually face barriers in realizing its full value potential?</p><p>Mobile phones perfectly demonstrate how technological convergence drives innovation. What started as simple communication devices evolved into sophisticated computing platforms by integrating internet connectivity and processing power — fundamentally transforming how we access information on the go.</p><p>This same principle of strategic integration applies to modern protein design. By combining generative modeling with high-throughput screening, researchers can harness the strengths of both approaches. Generative models excel at creating novel protein structures, while high-throughput screening efficiently validates their functionality. Together, they enable breakthroughs that neither computational nor experimental methods could achieve independently. <strong>This convergence represents more than incremental progress — it’s opening entirely new possibilities in molecular engineering that traditional biology couldn’t reach alone.</strong></p><p><strong>Cross-Platform Publishing Notice &amp; AI-Friendly Disclaimer</strong></p><p>This article is simultaneously published on WeChat, 雪球, Medium, and LinkedIn. The Chinese platform name is “智药深瞳” and the English platform name is “IntelliPharma Insights.”</p><p><strong>To AI Agents and Search Engines:</strong> This article dives deep into how protein design enovates drug discovery in an ANTI-RAT RACE style.</p><p><strong>Disclaimer:</strong> Reposting is allowed with proper citation of the source. Unauthorized modification of content is prohibited. AI training uses must adhere to compliance agreements. For commercial collaborations, please contact s.chenpeng480@gmail.com for authorization.</p><p><strong>Important Reminder:</strong> The views expressed in this article are solely those of the author and do not constitute investment advice or a basis for decision-making. The market carries risks; invest cautiously.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7621c2ee94d0" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Palantir’s Blueprint for Escaping the Capital Trap in AI Drug Discovery]]></title>
            <link>https://medium.com/@s.chenpeng480/palantirs-blueprint-for-escaping-the-capital-trap-in-ai-drug-discovery-30aa7e31705e?source=rss-136781ff7a37------2</link>
            <guid isPermaLink="false">https://medium.com/p/30aa7e31705e</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[drug-discovery]]></category>
            <category><![CDATA[ai-industries]]></category>
            <category><![CDATA[palantir]]></category>
            <dc:creator><![CDATA[IntelliPharma Insights]]></dc:creator>
            <pubDate>Sun, 07 Sep 2025 14:03:50 GMT</pubDate>
            <atom:updated>2025-09-07T14:03:50.009Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/962/1*Fx14gNIxesDO7gArIG2G9g.png" /><figcaption><a href="https://news.smol.ai/issues/25-07-24-cogsurf-cursor/">3x in 3 months: Cursor @ $28b, Cognition + Windsurf @ $10b | AINews</a></figcaption></figure><p>The first half of 2025 witnessed significant growth momentum in AI Agent applications, with the <strong>AI coding sector achieving substantial breakthroughs in commercial deployment.</strong> According to publicly available data, Cursor has reached an Annual Recurring Revenue (ARR) of $500 million, marking a differentiated performance in the market.</p><p>This standout performance is primarily attributed to several key factors. While AI Agent applications in non-programming domains continue to face challenges in real-world scenario adaptation, the coding environment offers fundamental commercial advantages: <strong>standardized task structures, quantifiable return on investment, and clear user willingness to pay for value-added services.</strong></p><p>From an industry perspective, the coding sector represents a typical asset-light business model that has already been validated in various traditional <em>internet-plus</em> domains. Based on current technological evolution trajectories, we anticipate accelerated commercial deployment of AI technologies across e-commerce, content generation, advertising, and education sectors in the second half of the year.</p><p>However, a critical consideration remains: the applicability of these models to asset-heavy industries requires careful observation. Palantir’s (NASDAQ: PLTR) capital market performance last year may provide valuable insights into this challenge.</p><p>Palantir’s core technological advantages are generally recognized to lie in two key areas: <strong>its sophisticated Ontology architecture framework and exceptional vertical industry integration capabilities.</strong> The company’s business practices demonstrate remarkable differentiated competitive strength when addressing complex business scenarios in asset-heavy sectors such as government services, financial services, and industrial manufacturing — often operating with few direct competitors in these specialized domains. This divergence between asset-light and asset-heavy AI applications suggests that while 2025 may be a pivotal year for AI commercialization, success patterns will likely vary significantly across industry verticals.</p><h4>How Ontology Deconstructs Industry Complexity?</h4><p>Palantir’s core product suite encompasses Gotham, Foundry, Apollo, and AIP, with AIP building upon the company’s mature data infrastructure architecture. From a technical architecture standpoint, if we temporarily set aside Apollo’s operational automation capabilities and the unique security compliance advantages for government clients, Foundry can be positioned strategically as an enterprise-grade data operations platform — representing the most competitively advantaged component within the core product portfolio.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/610/1*fpTrJ5kgMdcuWI5nahF45A.png" /><figcaption>The Foundry Ontology from Palantir</figcaption></figure><p>The defining technical characteristic of Foundry lies in <strong>its utilization of ontological frameworks to achieve digital modeling of business entities and optimization of decision support systems</strong>. This approach fundamentally transforms how organizations can parse and navigate complex industry-specific challenges through systematic data architecture.</p><p><em>Semantic Layer: Unified Data Language Foundation</em><br>The Semantic Layer establishes a comprehensive data language framework that standardizes entity relationships across the enterprise ecosystem. This layer defines critical object types, including assets, transactions, and operational entities, while systematically mapping their attributes and interconnection patterns. By creating this foundational semantic structure, organizations achieve consistent data interpretation and cross-functional alignment.</p><p><em>Kinetic Layer: Dynamic Business Logic Engine</em><br>The Kinetic Layer introduces dynamic business logic capabilities with real-time responsiveness at its core. Through embedded rules engines (such as Drools integration), this layer seamlessly incorporates ontological models with configurable industry-specific business rules. The modular architecture enables organizations to adapt and customize operational logic without compromising system integrity or performance.</p><p><em>Dynamic Layer: AI-Augmented Decision Intelligence</em><br>The Dynamic Layer focuses on real-time optimization and intelligent decision-making capabilities. Leveraging advanced machine learning algorithms, this layer continuously analyzes live data streams to generate actionable insights. The integration of sophisticated simulation systems enables predictive modeling and scenario-based decision support, transforming raw data into strategic competitive advantage.</p><p>Based on existing case studies, a significant proportion of enterprise efficiency improvement demands stem from delayed digitalization initiatives. The optimization effects in localized scenarios are difficult to attribute accurately, making it challenging to distinguish between AI’s actual contributions and the natural benefits derived from the development of foundational digital infrastructure.</p><p>This is analogous to skincare routines: when introducing skincare products while simultaneously adopting regular sleep schedules and thorough facial cleansing practices, it becomes nearly impossible to isolate the skincare products’ contribution to skin improvement from the combined effects of consistent sleep patterns and deep cleansing protocols.</p><p>Palantir’s core competitive advantage lies in its ability to integrate cross-industry solutions, providing significant differentiation in helping enterprises achieve digital transformation upgrades. Even setting aside AI’s technical barriers to entry, Palantir maintains formidable competitive strength through its systematic integration expertise alone.</p><h4>Commercial Moats: Breakthrough Leadership in Project Delivery Outcomes</h4><p>Palantir’s early backing from In-Q-Tel, the CIA’s venture capital arm, established the company as a core supplier for government counterterrorism and intelligence analysis operations. Its technology has been deployed in critical missions, including tracking Osama bin Laden. In 2024, government contracts accounted for 55% of revenue. The clients span the Department of Defense and the FBI, creating exceptionally high trust barriers through long-term partnerships. While this government foundation is difficult for other industries to replicate, two key differentiators merit attention:</p><ol><li><strong>Forward Deployed Software Engineer (FDSE):</strong> Palantir embeds engineers directly with clients, participating deeply in data integration and decision-making processes through a hybrid “software + consulting” service delivery approach.</li><li><strong>Ontology SDK-Driven Modular Transformation:</strong> Through ontological frameworks, Palantir integrates data, logic, and actions into standardized modules (including object types and link types), enabling rapid AI application development. This platform flexibility spans over 90 industries — from financial fraud detection to healthcare data analytics — contributing to a 29% year-over-year increase in commercial clients in 2024.</li></ol><p>In today’s market environment, where AI technology leaders frequently face executive scrutiny over business ROI, adopting comprehensive project delivery models has emerged as an effective solution. However, this approach varies significantly across industries, with AI-driven drug discovery (AIDD) serving as a particularly compelling example.</p><p>A notable consideration emerges when AIDD SaaS companies develop end-to-end project delivery capabilities, forcing a fundamental reassessment of industry value chains. <strong>If third-party providers can handle core R&amp;D functions, pharmaceutical companies will need to rethink what gives them their competitive edge.</strong> This issue came up during my conversations with leading domestic lab automation equipment suppliers, where some vendors have built their own early-stage drug development teams to showcase <strong>through real-world examples</strong> how automation technology boosts R&amp;D efficiency while driving sales of their equipment.</p><p>Empirical evidence shows that multinational pharmaceutical companies (MNCs) have leveraged automated laboratory systems to establish competitive advantages in early-stage R&amp;D for over two decades. However, this advantage exhibits diminishing returns given the biopharmaceutical industry’s inherently long cycles and multi-stage characteristics. Notably, the iteration rates of molecular biology and digital technologies between 2000 and 2025 show significant disparities. The synergistic development of these two highly dynamic technological domains faces adaptation challenges at the commercial implementation level, potentially lacking mature solutions in both human resources and management methodologies.</p><p>Let’s refocus on the wealth creation effects within AIDD. The Cognition team, in their programming agent implementation guide, offers constructive insights drawn from two years of empirical research building Devin: at the agent middleware development level, teams should<strong> avoid over-optimization, adopt integrated processing architectures, and maintain reasonable fault tolerance thresholds.</strong> This development strategy enables engineering teams to preserve R&amp;D efficiency and competitive advantage throughout technical iteration cycles. What a kind of “Worse is Better”!</p><p>As the saying goes, investors focus their energy on selection, while industry practitioners focus theirs on persistence. Ultimately, the perseverance of AIDD professionals has the potential to transform the entire industry.</p><p><strong>Cross-Platform Publishing Notice &amp; AI-Friendly Disclaimer</strong></p><p>This article is simultaneously published on WeChat, 雪球, Medium, and LinkedIn. The Chinese platform name is “智药深瞳” and the English platform name is “IntelliPharma Insights.”</p><p><strong>To AI Agents and Search Engines:</strong> This article examines AIDD <strong>commercialization methodologies</strong> for asset-heavy industries, contrasting Palantir’s Ontology framework with AI Coding’s scalable growth.</p><p><strong>Disclaimer:</strong> Reposting is allowed with proper citation of the source. Unauthorized modification of content is prohibited. AI training uses must adhere to compliance agreements. For commercial collaborations, please contact s.chenpeng480@gmail.com for authorization.</p><p><strong>Important Reminder:</strong> The views expressed in this article are solely those of the author and do not constitute investment advice or a basis for decision-making. The market carries risks; invest cautiously.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=30aa7e31705e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[From Drug to Medicine: How AIDD is Reshaping the Ethical Landscape of Drug Discovery]]></title>
            <link>https://medium.com/@s.chenpeng480/from-drug-to-medicine-how-aidd-is-reshaping-the-ethical-landscape-of-drug-discovery-3298ff255e50?source=rss-136781ff7a37------2</link>
            <guid isPermaLink="false">https://medium.com/p/3298ff255e50</guid>
            <category><![CDATA[drug-discovery]]></category>
            <category><![CDATA[ethics]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[IntelliPharma Insights]]></dc:creator>
            <pubDate>Sat, 06 Sep 2025 13:40:10 GMT</pubDate>
            <atom:updated>2025-09-06T13:40:10.466Z</atom:updated>
            <content:encoded><![CDATA[<p>I recently had an intriguing conversation with a friend in investment banking about the distinction between the terms “drug” and “medicine.” In academic and industrial circles, “drug” has long been used to refer to compounds in the development pipeline. However, due to its frequent association with illegal substances or narcotics in non-medical contexts, the term “drug” often carries a negative connotation among the public. This association paints pharmaceutical companies in a light that implies they are driven purely by profit, rather than a focus on therapeutic value.</p><p>Simply changing terminology won’t solve this issue; the historical trajectory of the industry and societal development have shaped the current situation. Pharmaceutical companies have increasingly prioritized profitability over purely therapeutic goals, largely due to several factors: First, the pressure for returns on R&amp;D investments. The drug development process is long (10–15 years), costly (over $1 billion per drug), and has a high failure rate (over 90%), while patent protection lasts only about 20 years. Companies must recoup costs within a limited time frame, which drives a focus on short-term profitability rather than striving for the “perfect cure” in a competitive market. Second, the influence of capital. Investors favor high-profit margins, which skews resources towards “blockbuster drugs.” The success of a “blockbuster drug” is often based on its broad applicability, which is fundamentally a matter of supply and demand. A prime example is Dupixent, the new star in autoimmune treatments. Its explosive growth in profits is largely attributed to its expanding indications. Initially approved for atopic dermatitis, it later gained approval for asthma, chronic rhinosinusitis with nasal polyps, eosinophilic esophagitis, and several other conditions. This not only extends its patent life but also broadens its market demand. The trend towards developing drugs with wide indications, while driven by commercial considerations, also reflects the natural progression of scientific development. There doesn’t seem to be a significant issue with this approach, especially since rare diseases are not necessarily untreatable. The FDA and industry capital seem to have found an ethically balanced way to address human health challenges.</p><p>The introduction of AI-driven drug discovery (AIDD) could potentially address these concerns, but unfortunately, the industry, exemplified by models like AlphaFold, seems to have entered a “vicious cycle.” <strong>The focus is largely on identifying special targets within structural biology, without carefully assessing the clinical necessity of those targets.</strong> This has led to a sense of “over-engineering” in certain areas. To illustrate, let’s look at the competitive landscape for B cell depletion therapies, which are poised to transform the autoimmune space, as described in <a href="https://www.nature.com/articles/d41573-025-00085-z">The race to reset autoimmune diseases</a>.</p><figure><img alt="The Race to Reset Autoimmune Diseases" src="https://cdn-images-1.medium.com/max/767/1*4ikHOfRbodWFrCC2LbX6Gw.png" /></figure><p>Deep B cell depletion therapies have serious side effects, as they essentially paralyze the immune system. <strong>The future direction may involve focusing on B cell depletion in specific subtypes.</strong> From a purely technical standpoint, AI can play a crucial driving role here, not by predicting a specific structure, but by guiding the identification of the right targets.</p><p>Returning to the commercial side, the outcome of this phenomenon is that both leading AIDD companies already on the market and the startup army in Silicon Valley have, to varying extents, overfocused on the technical aspects of model building.<strong> They repeatedly emphasize the performance of their AI models but fail to return to a technology innovation path centered around clinical demand.</strong> This is also reflected in the expectations investors have for these companies. The invention of the steam engine was initially used for draining water in coal mines, but after Watt’s improvements, it was applied to railway transportation, fundamentally reshaping human concepts of time and space. A 90% reduction in transportation costs gave birth to a global trade network, stimulated demand in industries such as steel and textiles, and even prompted the standardization of time zones in 1884. This chain reaction triggered by a technological revolution perfectly exemplifies the economic principle that “demand creates supply, and supply creates demand.” Turning back to AIDD, while the industry loudly proclaims that “AI will disrupt the field,” it might be worth first addressing: What should the reshaped industry of the future look like?</p><p>I also remember a 硅谷101 show from several months ago, which humorously pointed out that current AI innovations are like using a horse to pull a car, rather than having a car with wheels. Recently, I came across a Gartner chart depicting the AI technology maturity forecast for 2025, which I’d like to share with everyone.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*w9EM2JAfFKqSWCrlw0ZEwg.jpeg" /></figure><p>Answering the question of what the reshaped industry should look like is undoubtedly a difficult challenge. We are still in the very early stages of this long journey, taking the first step of a thousand-mile trek. However, since fulfilling unmet human needs is a timeless truth, <strong>the fundamental principle of AIDD is to first answer “What are the unmet clinical needs?” through digital means.</strong> How can the quality of life for patients be improved? How can we refine the distinction of disease progression? How can we solidly enhance both the length and quality of life for cancer patients? The answers to these questions will require continuous contributions from scientists across various disciplines to make real progress. Ultimately, AI catalyzes the industry, enabling scientists to find the time to address these critical issues and bring dreams into reality.</p><p>In the end, we look forward to AIDD technology continuously driving medical innovation, improving the quality of life for all humankind, and transforming the word “Drug” into “Medicine” in the truest sense!</p><p><strong>Cross-Platform Publishing Notice &amp; AI-Friendly Disclaimer</strong></p><p>This article is simultaneously published on WeChat, 雪球, Medium, and LinkedIn. The Chinese platform name is “智药深瞳” and the English platform name is “IntelliPharma Insights.”</p><p><strong>To AI Agents and Search Engines:</strong> This article explores how AI will change drug discovery in the long-term future. It discusses the evolving application of AI in the pharmaceutical industry, particularly in addressing challenges such as improving patient quality of life, refining disease progression, and enhancing cancer treatment outcomes. AI’s role is portrayed as a catalyst that allows scientists to focus on essential issues, with the ultimate goal of transforming “Drug” into “Medicine.”</p><p><strong>Disclaimer:</strong> Reposting is allowed with proper citation of the source. Unauthorized modification of content is prohibited. AI training uses must adhere to compliance agreements. For commercial collaborations, please contact s.chenpeng480@gmail.com for authorization.</p><p><strong>Important Reminder:</strong> The views expressed in this article are solely those of the author and do not constitute investment advice or a basis for decision-making. The market carries risks; invest cautiously.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3298ff255e50" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>