<?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 AirLyft on Medium]]></title>
        <description><![CDATA[Stories by AirLyft on Medium]]></description>
        <link>https://medium.com/@airlyftone?source=rss-1d32682518c8------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*fX8JqtXf2oDXm8r6PuA0Bg.jpeg</url>
            <title>Stories by AirLyft on Medium</title>
            <link>https://medium.com/@airlyftone?source=rss-1d32682518c8------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 06:39:16 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@airlyftone/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[The decision we eventually reached was not sudden.]]></title>
            <link>https://blog.kyte.one/the-decision-we-eventually-reached-was-not-sudden-749ceb558791?source=rss-1d32682518c8------2</link>
            <guid isPermaLink="false">https://medium.com/p/749ceb558791</guid>
            <dc:creator><![CDATA[AirLyft]]></dc:creator>
            <pubDate>Sun, 11 Oct 2026 10:06:01 GMT</pubDate>
            <atom:updated>2026-10-11T10:06:01.771Z</atom:updated>
            <content:encoded><![CDATA[<h3>The decision we eventually reached was not sudden. This is what the team has been carefully weighing as the data kept shifting in one direction</h3><blockquote>Nothing about where the team has arrived has happened overnight.</blockquote><p>There is a phrase we heard more times than we can count over the last two years. It came from teams we respected. Teams that had used the platform and knew what it could do.</p><p>We will surely use this for our next campaign but for now we are <strong>putting it on hold.</strong></p><p>They were not saying the product did not work. They were saying the market had made deploying capital on campaigns the kind of decision that kept getting pushed to the next quarter. And the quarter after that. And the one after that.</p><p>That is the thing that made stopping feel wrong every single time the question came up internally.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lWYeU3Ldpjt5PERkODF54g.jpeg" /></figure><h3>The moments that made us believe it could still work</h3><p>There were specific periods where the evidence felt like it was pointing somewhere different.</p><p>When we worked with <strong>Sony Labs Soneium</strong> and <strong>Astar Network</strong>, two of the more credible names building serious infrastructure in the space, we had the clearest validation we had seen that the platform could operate at the level we had been trying to reach. The campaigns ran. The verification held. The analytics gave those project teams visibility into their users that they had not had before. It felt like the moment we had been building toward.</p><p>The market cooled before it could compound into something larger.</p><p>When we worked with <strong>Mythical Games</strong>, one of the most established names in web3 gaming, the proof was different in kind but equally clear. It showed the platform could extend from web3 native projects into the broader gaming world we had been trying to reach. The concept worked. The execution worked. The response was the same one we were hearing everywhere.</p><p>We will look into it but we are not deploying capital into campaigns right now.</p><p>That gap between the product working and the market being ready to pay for it consistently is the one that defined the last two years more than anything else.</p><h3>Everything the team tried</h3><p>We want to be honest about the list because we think you deserve to see it.</p><p>We changed pricing multiple times hoping the barrier was cost rather than conditions. We introduced different strategies trying to meet teams where their available budgets actually were. We reached out to existing partners looking for any path to renewed activity.</p><p>We tried to build a bridge between web2 and web3, identifying projects from traditional gaming and entertainment that wanted to establish a presence in web3 and bringing them to the platform as a new category of customer. We had conversations with gaming and esports companies across Brazil, Japan and China that went through multiple rounds of discussion and felt like they were building toward something real. And then they went quiet. Not a no. Just silence. Which is its own kind of answer. We conceptualised an automated loyalty engine, a system that would remove the manual effort of campaign design entirely, and took it to multiple investors looking for a lead. We did not find one.</p><p>The feedback we kept hearing was not about the product itself. It was about the market question underneath everything. The question investors kept asking was why users from web2 would come to web3 at scale, in a cycle where the category was contracting rather than expanding. We understood why they were asking it. We did not have a version of the answer that changed what they decided.</p><p>We used AI to build a probability model for outreach, scoring potential partners based on their likelihood of deploying a campaign and prioritising conversations accordingly. We ran internal brainstorming sessions every single week based on what those conversations were telling us, looking for the configuration that would change what the data kept showing.</p><p>We worked at wages below what the team could have earned elsewhere. We stopped hiring. We kept the infrastructure running. We kept looking.</p><p>The market data explains why those efforts kept running into the same wall. Blokhaus described the broader environment in their 2026 State of Crypto Marketing report directly. Airdrop farming had become professionalized to the point of meaninglessness. The protocols still using campaigns as their primary growth lever were inflating their metrics while losing the long game. [blokhaus.io/blog/state-of-crypto-marketing-2026.html]</p><p>A Keyrock analysis of 62 airdrops across six chains found that 88 percent of airdropped tokens declined in price, with most drops happening within the first fifteen days of launch. [hackernoon.com/88percent-of-airdrops-fail-and-nobody-in-web3-wants-to-admit-it] Memento Research, analysing 118 token generation events in 2025, found that 84.7 percent of launched tokens were trading below their token generation event price.</p><p>The projects that should have been our strongest customers were the ones most exhausted by what campaigns had become in the category. The teams that understood the value of genuine engagement were also the teams most aware of how badly farming behaviour had distorted every metric they were being measured on.</p><p>We kept trying because we believed the right configuration was somewhere in what the data was telling us. We kept looking hoping it would point somewhere different than it kept pointing.</p><h3>The thing that made it all worth it</h3><p>Through all of it, the moments that stayed with us were not the partnership announcements or the campaign numbers.</p><p>They were the moments when a team would see their analytics dashboard come alive and something would click. When the data turned green and it meant something real to the people looking at it. When a project that had been running campaigns and watching the numbers but not quite understanding what they were seeing would finally see which users were genuine and which were not, and what it would take to keep the real ones.</p><p>Their response in those moments was not just gratitude. It was relief.</p><p>There were teams from different parts of the world. Different ecosystems. Different stages of building. What connected them was the same moment of recognition across all of it. The platform had helped them understand something real about their users that they had not been able to see before. That problem got solved, in those moments, for those teams, wherever they were.</p><p>That is what the team knows it built. And it is the reason the data on the other side of the weighing has been the hardest thing to look at honestly.</p><h3>What this has led toward</h3><p>The conversations happened over months. The data was looked at more than once hoping it would point somewhere different. The question of what would need to be true for continuing to make sense was asked and answered and then asked again.</p><p>We will continue being honest with you in the notes ahead. We owe you that and we will give it to you that way.</p><p>The team</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=749ceb558791" width="1" height="1" alt=""><hr><p><a href="https://blog.kyte.one/the-decision-we-eventually-reached-was-not-sudden-749ceb558791">The decision we eventually reached was not sudden.</a> was originally published in <a href="https://blog.kyte.one/">Kyte.One</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Of the ninety nine+projects that closed in 2026, some were teams we had run campaigns alongside.]]></title>
            <link>https://blog.kyte.one/of-the-ninety-nine-projects-that-closed-in-2026-some-were-teams-we-had-run-campaigns-alongside-83ff491f68ec?source=rss-1d32682518c8------2</link>
            <guid isPermaLink="false">https://medium.com/p/83ff491f68ec</guid>
            <dc:creator><![CDATA[AirLyft]]></dc:creator>
            <pubDate>Sun, 04 Oct 2026 04:31:06 GMT</pubDate>
            <atom:updated>2026-10-04T04:31:06.170Z</atom:updated>
            <content:encoded><![CDATA[<h3>Of the ninety nine+projects that closed in 2026, some were teams we had run campaigns alongside. What the closure wave looked like from inside it</h3><p>We found out the way you usually find out in this space.</p><p>Through a Telegram channel that went quiet. A campaign dashboard that stopped showing activity. A team contact who stopped responding to messages. Then weeks later a post from the founder explaining what had happened. Or sometimes nothing at all. Just silence where a project had been.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*9Nnmn33x74E7_ROrff0Leg.jpeg" /></figure><p>We went through that sequence more than once in 2025 and 2026.</p><h3>What the scale of it looked like</h3><p>RootData published its 2026 Crypto Industry Dead Projects List in July 2026, tracking formal closures across the space. Their count reached 99 projects that had halted operations within the year, spanning every sector of the ecosystem. DeFi. NFTs. Wallets. Exchanges. Analytics platforms. Infrastructure services. [en.coin-turk.com/99-crypto-projects-shut-down-in-2026-rootdata-reports/]</p><p>The count was confirmed independently by multiple outlets tracking the same data. U.Today reported the same figure. [u.today/2026-crypto-graveyard-99-projects-officially-shut-down] Memeburn reported on the patterns behind the closures. [memeburn.com/99-crypto-projects-shut-down-in-2026-bear-market/]</p><p>Ninety nine formal closures in a single year. That count does not include the projects that simply went quiet. The ones that posted their last update in early 2025 and never posted again. Community channels still open. Websites still live. Team gone.</p><h3>What the ones we knew had in common</h3><p>Some of those ninety nine were teams we had worked alongside.</p><p>Not all of the projects in the closure wave were ones we had a relationship with. But some were. Teams that had run campaigns through our platform. Teams we had sent campaign ideas to before community events or around feature launches. Teams whose analytics we had helped them read and understand. Teams we had watched try to build something real in a category that was making it very difficult.</p><p>What their closure announcements had in common was an honesty that most of their earlier communications had not had.</p><p>When teams decided to stop they often said plainly what they had not been able to say while they were still going. That the economics had not worked. That the capital had not come. That the market had moved on before they could reach the stage where any of it might have resolved.</p><p>We recognised in their descriptions something about the conditions the entire category had been navigating.</p><p>The teams that had been closest to us in the work were the ones where the closure landed differently than a number on a dashboard.</p><p>We had been on calls with some of those teams as their campaigns were running, watching verification data come in alongside them in real time. We had helped them read their first analytics breakdown. Understand which regions their campaign participants were actually coming from. How long users were genuinely spending on the product versus clicking through to claim a reward and leaving. We had sent campaign ideas to some of them because we thought the ideas would help them build something more durable than a single reward event.</p><p>When those specific teams made the decision to stop it was not a client metric that changed. It was the end of a conversation we had been part of for a while.</p><h3>What it looked like from where we were standing</h3><p>For a platform that built its work around serving projects like those, each closure was a different kind of event from the outside.</p><p>From the outside it is a statistic. One of ninety nine. One more data point in a year that had too many.</p><p>From where we were standing it was something more specific.</p><p>There is a particular weight to looking at a campaign dashboard you have maintained for a team that has since stopped. The data is still there. The campaigns are still in the history. But there is no new activity. No brief coming in for the next campaign. No follow-up about what the last analytics showed about user behaviour.</p><p>Some of those relationships had gone on for years. We had watched those teams iterate on how they thought about their campaigns, get better at asking what their users were actually doing rather than just how many had shown up, start to understand the difference between participation numbers and something more meaningful. And then the conditions became too much and they stopped.</p><p>A proof point for the engagement problem we had spent years working on that would now be filed under things that did not work out rather than things that were still being built.</p><p>We are not saying this to make it more dramatic than it was. We are saying it because it is what the closure wave actually felt like to watch from inside the category rather than from a distance.</p><h3>A note on what comes next</h3><p>The ninety nine projects that closed in 2026 had each navigated versions of what we described in the last note. Treasury pressure. A funding environment that was not there. A client base that was contracting. The difference between them and the teams that are still here is not always as large as it might appear from the outside.</p><p>What the next notes will address is what the team has been working through as the same pressures accumulated. And what that has required us to face honestly.</p><p>The team</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=83ff491f68ec" width="1" height="1" alt=""><hr><p><a href="https://blog.kyte.one/of-the-ninety-nine-projects-that-closed-in-2026-some-were-teams-we-had-run-campaigns-alongside-83ff491f68ec">Of the ninety nine+projects that closed in 2026, some were teams we had run campaigns alongside.</a> was originally published in <a href="https://blog.kyte.one/">Kyte.One</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How more than two trillion dollars leaving the crypto market translated into reduced budgets, lost…]]></title>
            <link>https://blog.kyte.one/how-more-than-two-trillion-dollars-leaving-the-crypto-market-translated-into-reduced-budgets-lost-99a063c21685?source=rss-1d32682518c8------2</link>
            <guid isPermaLink="false">https://medium.com/p/99a063c21685</guid>
            <dc:creator><![CDATA[AirLyft]]></dc:creator>
            <pubDate>Sun, 27 Sep 2026 02:31:01 GMT</pubDate>
            <atom:updated>2026-09-27T02:31:01.937Z</atom:updated>
            <content:encoded><![CDATA[<h3>How more than two trillion dollars leaving the crypto market translated into reduced budgets, lost funding opportunities, and real treasury pressure for a small team like ours</h3><p>There is a specific kind of arithmetic that teams in this space try not to run too early. The kind where you look at what you have, what you owe, what you are spending, and what is coming in, and see whether the lines cross somewhere you can still reach.</p><p>At some point we ran it.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0smru0L0hvJfZmB4S6w5qA.jpeg" /></figure><p>We are not going to give you the specific numbers. What we can tell you is what the three pressures looked like individually and what they looked like when they arrived together.</p><h3>What two trillion dollars actually means for a small team</h3><p>CoinGecko’s Q2 2026 Industry Report documented what happened to the total crypto market between October 2025 and June 2026. The market capitalisation fell roughly 52 percent below its October 2025 peak. [coingecko.com/research/publications/2026-q2-crypto-report]</p><p>CoinPedia reported the specific numbers in June 2026. The total cryptocurrency market value dropped from a record $4.3 trillion in October 2025 to approximately $2 trillion. More than $2.3 trillion erased in eight months. [coinpedia.org/crypto-live-news/crypto-market-cap-falls-2–3-trillion-from-october-2025-peak/]</p><p>A small team does not absorb a number that large directly. But when your treasury, your funding prospects, and your client budgets all move with the same market, you absorb it indirectly. In every direction at once.</p><h3>Treasury</h3><p>We wrote in an earlier note about what the Terra collapse of May 2022 did to our treasury. A meaningful portion of what we had raised disappeared in that event. That was the acute version of the pressure.</p><p>The 2025 to 2026 contraction worked differently. It was not a cliff edge. It was a long slope. Month after month the value of what we held declined while the cost of what we needed to maintain stayed the same. Server infrastructure. Engineering time. The things that keep a platform running do not get cheaper because the market has turned.</p><p>There is a specific weight to watching what you have available decline month after month not because you are spending recklessly but because the value of what you are holding is moving in one direction while everything you need to pay for stays fixed. You make decisions about what to maintain and what to let go. You look at what the team needs and what the platform needs and you try to keep both going. There is no version of that period that was simple to navigate.</p><h3>Funding</h3><p>We had the conversations. We showed the product. We showed the growth in campaigns run, the verification technology, the analytics layer we had built in response to what projects were asking for.</p><p>We understood why the capital was not there.</p><p>Understanding the reason does not change what it means to hear it.</p><p>Industry data from April 2026 showed that crypto VC funding had fallen to its lowest monthly figure in two years. A 74 percent decline from the prior month. The capital had moved. To AI. To real world assets. To areas attracting the attention and the investment that this space had once commanded. The investors who would have written that cheque in 2023 were writing different cheques in 2025.</p><p>What we needed to fund the next stage of the platform was not available in the market when we needed it. That is not a complaint. It is the honest account of what the window looked like when we were trying to move through it.</p><h3>Client budgets</h3><p>The third pressure was the one we watched happen in real time.</p><p>When a project’s treasury is denominated in its own token and that token has lost 70 or 80 percent of its value, the budget for everything contracts at the same rate. Not by choice. Just by the arithmetic.</p><p>The teams we had worked with closely, the ones who had run multiple campaigns and were starting to understand what sustained engagement could build for them, were having different conversations with us by 2025. Not about expanding. About what they could still afford. Some made the decision to pause. Some went quiet. Some kept going but at a scale that made it clear they were operating differently than they had been.</p><p>We understood it. We had written about exactly this dynamic in earlier notes. Understanding why something is happening does not make it easier to absorb when it is happening to you.</p><h3>What all three at once actually feels like</h3><p>Treasury pressure. A funding environment that was not there. A client base operating on contracting budgets.</p><p>Each of these alone is something a small team can work around with enough time and enough creativity. All three arriving together, compounding on each other, in a market that was not moving toward resolution, is a different kind of problem.</p><p>Two trillion dollars. That is what left this space in eight months. For a small infrastructure team the arithmetic that comes out of that is the arithmetic we had been trying not to run for as long as we could.</p><p>The next few notes will be honest about what it showed us when we finally did.</p><p>The team</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=99a063c21685" width="1" height="1" alt=""><hr><p><a href="https://blog.kyte.one/how-more-than-two-trillion-dollars-leaving-the-crypto-market-translated-into-reduced-budgets-lost-99a063c21685">How more than two trillion dollars leaving the crypto market translated into reduced budgets, lost…</a> was originally published in <a href="https://blog.kyte.one/">Kyte.One</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The retention problem we built AirLyft to solve has not been solved by anyone.]]></title>
            <link>https://blog.kyte.one/the-retention-problem-we-built-airlyft-to-solve-has-not-been-solved-by-anyone-a8e7cc7e7a4b?source=rss-1d32682518c8------2</link>
            <guid isPermaLink="false">https://medium.com/p/a8e7cc7e7a4b</guid>
            <dc:creator><![CDATA[AirLyft]]></dc:creator>
            <pubDate>Mon, 21 Sep 2026 08:53:20 GMT</pubDate>
            <atom:updated>2026-09-21T08:53:20.062Z</atom:updated>
            <content:encoded><![CDATA[<blockquote>What the data from 2026 says about how far the category still has to go</blockquote><p>The first time we saw it clearly we thought it was one project’s specific problem.</p><p>A team had run a campaign through our platform. It had gone well by every measure campaigns are usually judged by. Thousands of wallets. Completed actions. Rewards distributed. The dashboard looked like success.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jOWFIuAWs93lt9N3jFDpfg.jpeg" /></figure><p>Three weeks later we checked in. The users who had participated were mostly gone. Not just from the campaign. From the product. The wallets that had showed up for the reward had no further activity.</p><p>We sat with that. We looked at the campaign design. We thought about what they could change for the next one. We changed a lot of things.</p><p>Then we saw it happen again. Same pattern. Different project. Different ecosystem. Different campaign mechanics. Same outcome. The users who came for the reward left when the reward was collected.</p><p>By the time we had seen it across dozens of campaigns, then across fifty, then across a hundred, we understood something the industry is still not comfortable saying plainly. The problem is that the category has not built what it needs to keep the people it reaches.</p><h3>What the data confirms</h3><p>HackerNoon published an analysis in June 2026 examining web3 wallet retention across the industry. Their finding was direct. Industry-wide, fewer than 20 percent of web3 wallet users return after 7 days. [hackernoon.com/80percent-of-your-wallet-users-are-gone-in-a-week-the-fix-isnt-ux]</p><p>CoinLaw’s research from February 2026 confirmed the same pattern independently. For many wallet-connected dApps, retention beyond 7 days remains below 20 percent. [coinlaw.io/web3-wallet-user-growth-statistics/]</p><p>Roughly 80 percent of users who connect to a web3 product are gone within a week. The campaign that brought them in accomplished the acquisition. It did not accomplish anything after that.</p><p>We had been watching this number before those studies were published. It is not new. It is not a bear market phenomenon. It is the category operating at a level below what it needs to be operating at if it is going to be something other than speculative.</p><h3>What building better analytics showed us</h3><p>Early in our work projects would come back and tell us they could see the participation numbers but not the story underneath them. So we built more.</p><p>We built session time parameter tracking so projects could understand how long users were actually spending on a campaign rather than just whether they had completed it. A user who spent thirty seconds on a campaign page and claimed the reward told a very different story from a user who spent ten minutes exploring the product. The time parameter showed which was which. What the data showed was often not what the teams had hoped it would show.</p><p>We built geographic breakdowns so project teams could see which regions their campaign participants were actually coming from. Whether the distribution matched where the project had real potential users. Whether the concentration patterns reflected genuine distributed interest or something more difficult to explain. For projects running campaigns in specific communities or markets this data was often the clearest picture they had ever had of whether they were reaching the right people.</p><p>What that data showed was not easier to look at than the headline participation number. It was harder.</p><p>We built better tools to understand the problem. The better tools showed us the problem more clearly. And seeing it more clearly did not make it smaller. It made it more specific and more honest. The gap between participation numbers and lasting engagement was real and it was one the category had not yet found a way to close.</p><h3>What five years of working on this problem actually showed us</h3><p>We built on chain verification because social screenshots told you nothing real about whether someone had used a product. We built an AI based content verification layer because manual review did not scale once campaigns reached any meaningful volume. We built a faucet for non EVM ecosystems because the onboarding friction was losing users before they ever had a chance to engage. We kept returning to the same question. What would it take to get someone from participating once to coming back.</p><p>The honest answer is that we never fully found it.</p><p>We got closer than the category average. We helped projects understand their real users better than a campaign dashboard could. We built tools that made verification more meaningful and genuine engagement harder to fake. We reached projects that started asking the right questions about their users rather than just their numbers.</p><p>But the number has not moved. Fewer than 20 percent returning after 7 days in 2020. Fewer than 20 percent returning after 7 days in 2026. Five years of work across an entire category and the retention problem that made us start building in the first place is still the retention problem.</p><p>We do not say this with bitterness. We say it because it is what the data shows. And because the people who have been reading this series deserve to have us name it plainly.</p><h3>Why this matters for what comes next</h3><p>The projects that made it through the 2026 contraction were the ones that had built genuine recurring relationships with users who had real reasons to stay. Real engagement. Real communities that did not disappear when the campaign ended. That is what the category needs. The 2026 data is clear about that.</p><p>What the data is equally clear about is how far the category still is from having it consistently. Most dApps. Most campaigns. Most projects trying to grow their user base. Still losing 80 percent of the people they reach within the first week.</p><p>We built AirLyft because we believed we could help close that gap. What five years of trying has shown us, and what the next few notes in this series will address more directly, is what happens when the problem you believed in stays stubbornly unsolved and the conditions around you make it harder and harder to keep working on it.</p><p>Eighty percent of users gone in a week. That number has not moved.</p><p>By the time you finish reading the next few notes you will understand what that has meant for us. And why we are writing them now.</p><p>The team</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a8e7cc7e7a4b" width="1" height="1" alt=""><hr><p><a href="https://blog.kyte.one/the-retention-problem-we-built-airlyft-to-solve-has-not-been-solved-by-anyone-a8e7cc7e7a4b">The retention problem we built AirLyft to solve has not been solved by anyone.</a> was originally published in <a href="https://blog.kyte.one/">Kyte.One</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What it costs to maintain something the market stopped valuing properly]]></title>
            <link>https://blog.kyte.one/what-it-costs-to-maintain-something-the-market-stopped-valuing-properly-e51b9922f3a2?source=rss-1d32682518c8------2</link>
            <guid isPermaLink="false">https://medium.com/p/e51b9922f3a2</guid>
            <dc:creator><![CDATA[AirLyft]]></dc:creator>
            <pubDate>Sun, 13 Sep 2026 04:31:03 GMT</pubDate>
            <atom:updated>2026-09-13T04:31:03.133Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Lsc1Oi1Rscdqkwylxtd6NQ.jpeg" /></figure><p>There is an equation that underlies every technology business. What it costs to build and maintain the product has to eventually be covered by what the market is willing to pay for it.</p><p>When those two numbers move too far apart in the wrong direction, the people maintaining the product absorb the difference.</p><p>That absorption has limits.</p><h3>What maintaining web3 infrastructure actually costs</h3><p>Research published in 2026 on web3 development costs made the maintenance arithmetic visible. Post-launch maintenance for web3 platforms averages between fifteen and twenty-five percent of the original build budget annually. Infrastructure monitoring costs between one thousand and five thousand dollars per month. Node infrastructure runs between two thousand and ten thousand dollars per month. Routine maintenance cycles, fixes, and ongoing operational costs can add twenty to forty percent to the total budget across the life of a project.</p><p>These are not one-time costs. They are the cost of keeping something running. Every month. Regardless of what the market is doing.</p><p>When you are operating across multiple chains, maintaining an on chain verification system that has to evolve as farming behavior evolves, running an AI based content verification layer that requires ongoing tuning, and supporting the integrations that project teams depend on, the maintenance costs are real and they compound over time.</p><h3>What the market stopped being willing to pay</h3><p>The grant culture we described in an earlier note trained the market to expect tools to be close to free. The token based pricing that collapsed with token markets made stable subscription revenue nearly impossible to generate at the rates that maintenance actually required. The survival mode that projects moved into made growth infrastructure the first line item to cut.</p><p>PANews described the pattern in its analysis of the 2026 closures. The common trait of the projects that survived, it noted, is that they address real demand and have clear revenue models. They do not live in narratives but in cash flows.</p><p>The projects that closed had the opposite problem. Polynomial processed twenty-seven million transactions and handled more than four billion dollars in cumulative trading volume. On paper that looks like a functioning product. Their total value locked peaked at around eight million dollars. The product was there. The market was not valuing it at what it cost to run.</p><h3>What absorbing the difference looks like</h3><p>When the gap between what something costs to maintain and what the market will pay for it persists long enough, the people maintaining it absorb the difference.</p><p>We have written in earlier notes about what that absorption looked like for the team behind this. The compensation decisions that made continuing possible. The hours that went into the work beyond what the economics were rewarding. The calls that happened at hours when they would have been easier not to take.</p><p>Paul Simroth, writing an analysis of web3 in 2026, offered the clearest framing of the category’s broader situation. Be skeptical of any trend that can only describe itself through its token launch. The equation of how long the teams building toward something more durable could absorb the difference between what maintenance cost and what the market would pay did not wait for the market to resolve that question.</p><h3>The honest account</h3><p>The category eventually recognised that genuine users matter, that sustainable revenue cannot be built on token incentives alone, and that campaigns need to be recurring relationships not one-off events. The engagement problem we had been working on is still real and still unsolved.</p><p>But the category arriving at that recognition is not the same as that recognition creating a sustainable market for the platforms that had been building toward it. Those are separate things. The market got to the right place. Whether it did so in time or in a way that made the economics work for a team like ours is a different question. One we will address more directly in the notes that follow.</p><h3>A note on what comes next</h3><p>In this note we shared what it actually costs to maintain web3 infrastructure, what the market stopped being willing to pay for it, and what the gap between those two numbers requires from the people who keep building anyway.</p><p>This is the most direct note we have written about the specific economics of what we have been navigating. The newsletter series we started set out to be honest with the people who have stayed with this project. We intend to keep being honest about what comes next as we work through it.</p><p>We will keep sharing as we figure out what comes next.</p><p>– The Team</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e51b9922f3a2" width="1" height="1" alt=""><hr><p><a href="https://blog.kyte.one/what-it-costs-to-maintain-something-the-market-stopped-valuing-properly-e51b9922f3a2">What it costs to maintain something the market stopped valuing properly</a> was originally published in <a href="https://blog.kyte.one/">Kyte.One</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How survival mode changed what web3 projects would and wouldn’t spend on]]></title>
            <link>https://blog.kyte.one/how-survival-mode-changed-what-web3-projects-would-and-wouldnt-spend-on-55f1c9a43c0a?source=rss-1d32682518c8------2</link>
            <guid isPermaLink="false">https://medium.com/p/55f1c9a43c0a</guid>
            <dc:creator><![CDATA[AirLyft]]></dc:creator>
            <pubDate>Sun, 06 Sep 2026 04:31:02 GMT</pubDate>
            <atom:updated>2026-09-06T04:31:02.601Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*156A0gVgxNFzUmbvdkGtsw.jpeg" /></figure><p>There is a specific way that a team’s decision-making changes when it moves from building to surviving.</p><p>The strategic questions about what to build next, what markets to enter, what tools to invest in for growth, all take a back seat to one question. How long can we keep going.</p><p>But here is what the data from 2025 and 2026 actually shows. The projects that cut everything and went silent did not survive better than the ones that had built genuine user relationships. The projects still operating are the ones that had built toward genuine users and sustainable revenue. And the market is only now reckoning with what it meant to have built on different foundations.</p><h3>What token-denominated operations actually produced</h3><p>CoinDesk described the financial reality of the 2026 contraction directly. Most of the projects now winding down, it noted, were never generating revenue in the traditional sense. They paid engineers in tokens, subsidised liquidity in tokens, and funded operations in tokens. When the vast majority of altcoins lost between 70 and 90 percent of their value the runway calculations that had seemed solid became wildly inaccurate overnight.</p><p>The direction was readable in the data for anyone following the category closely. The farming data had been showing for years that token incentivised engagement was producing dashboards full of numbers that did not represent real users. The treasury data was showing that token denominated operations were one market cycle away from crisis. The engagement data was showing that users who came for rewards left when the rewards changed.</p><p>The market built on those foundations anyway. And now it is dealing with the consequences.</p><h3>What got cut first and why that mattered</h3><p>When survival mode hit, the pattern of what got cut followed a consistent logic. Campaign platforms. Growth tools. Analytics services. Engagement infrastructure. The things projects invest in when they are building toward something larger.</p><p>The logic made sense in isolation. When focused on survival you stop spending on growth. But it missed something important.</p><p>The projects that made it through the contraction were not the ones that went silent and cut everything. They were the ones that had built genuine ongoing relationships with users who had real reasons to stay. Memeburn’s analysis was direct. The projects dying were overwhelmingly the ones that had relied on token incentives to attract users and token price appreciation as revenue. They had never built what sustained engagement actually requires.</p><p>CoinTribune captured the lesson plainly. Temporary rewards can still launch a platform. They are no longer enough to maintain it.</p><p>Cutting the tools that build genuine engagement in a survival moment was, in many cases, cutting exactly the wrong things.</p><h3>What the surviving projects were actually doing</h3><p>CoinDesk identified the common factor among the protocols that not only survived but grew through the bear market. They generate revenue in dollars not in their own token. They had real users. Genuine engagement. Communities built on more than rewards.</p><p>Hyperliquid crossed one billion dollars in cumulative fees by June 2026. Aave held more than twelve billion dollars in deposits and generated over one hundred million dollars in annualised borrow fees. These were not projects that built their communities through farming and airdrop campaigns. They built communities through genuine product utility and recurring user engagement.</p><p>That is what we had been trying to help projects do. Build the kind of community that does not disappear when the rewards run out. Understand who their real users were and not just their wallet counts. Run campaigns as recurring touchpoints that give users ongoing reasons to stay rather than as one-off events designed to inflate a number.</p><p>The category moved in that direction. The projects that survived were the ones that had built toward genuine users and sustainable revenue. The market eventually recognised what mattered.</p><p>But the market recognising something and that recognition creating a sustainable environment for the platforms building toward it are not necessarily the same thing. That distinction is what the next few notes will address more directly.</p><h3>What this means for platforms like ours</h3><p>The engagement problem was visible from the beginning and the market was building on foundations that would eventually require a reckoning. What no one could predict was when that reckoning would come or how the category would look when it did.</p><p>We were inside the same contraction. Not as observers. As one of the platforms that had spent years building what the surviving category needed, in a market that was focused on something else until the moment it had no other choice.</p><h3>A note on what comes next</h3><p>In this note we shared how survival mode changed what web3 projects would and would not spend on. The financial mechanics that made token denominated operations unsustainable. Why the projects that cut growth infrastructure did not necessarily make the right call. What the surviving projects were actually doing that proved which approach was right. And what all of this meant for platforms like ours that had been building toward exactly that.</p><p>Not a recovery narrative. Not a roadmap. Just an honest account of what the market chose, what the data showed about those choices, and where that left the platforms that had been building toward what the category eventually needed, in a market that had not yet arrived at the same place.</p><p>These notes are part of an ongoing series from the team. We will keep sharing as the work continues.</p><p>– The Team</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=55f1c9a43c0a" width="1" height="1" alt=""><hr><p><a href="https://blog.kyte.one/how-survival-mode-changed-what-web3-projects-would-and-wouldnt-spend-on-55f1c9a43c0a">How survival mode changed what web3 projects would and wouldn’t spend on</a> was originally published in <a href="https://blog.kyte.one/">Kyte.One</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What farming behaviour did to the economics of building campaign platforms]]></title>
            <link>https://blog.kyte.one/what-farming-behaviour-did-to-the-economics-of-building-campaign-platforms-ae803708dd90?source=rss-1d32682518c8------2</link>
            <guid isPermaLink="false">https://medium.com/p/ae803708dd90</guid>
            <dc:creator><![CDATA[AirLyft]]></dc:creator>
            <pubDate>Sun, 30 Aug 2026 04:31:02 GMT</pubDate>
            <atom:updated>2026-08-30T04:31:02.138Z</atom:updated>
            <content:encoded><![CDATA[<p>The conversation about farming in web3 has always focused on the projects that were farmed. The tokens that got drained. The airdrops that rewarded nobody who actually used the product.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*h5cKQo9nxn6DBOUXzq-Biw.jpeg" /></figure><p>What gets discussed less is what farming behaviour did to the platforms trying to build genuine campaign infrastructure. The economics there are different. And in some ways harder.</p><h3>What the data actually shows</h3><p>A 2025 analysis published on Medium described the economics of airdrop farming with unusual clarity. The customer acquisition cost for airdrops, it argued, is among the highest in the tech industry. Not in dollars but in token dilution. Projects were spending millions worth of tokens to acquire users who were not curious, aligned, or loyal. They were opportunistic. They corrupted product data. They left dashboards full of metrics that looked like momentum but represented none.</p><p>Eighty-eight percent of airdropped tokens lose value within three months of launch. That is not a market cycle problem. That is a fundamental problem with the model.</p><p>By 2026 the farming community had industrialised. A service offering anti detection tools for airdrop farmers noted that the majority of significant airdrops in 2025 and 2026 had implemented Sybil filtering. Which means the farming community had advanced to the point where professional guides existed for evading that filtering. Each new Sybil resistance tactic, observed one analyst, only prompts more sophisticated farming tools.</p><h3>The arms race and what it cost</h3><p>Passport, a digital identity infrastructure platform, described what farming behaviour had done to the category in 2026 with a sentence worth quoting directly. If web3 cannot tell the difference between a person and a swarm everything else becomes theatre.</p><p>That is not a rhetorical point. It is the description of an actual economic problem.</p><p>For a campaign platform trying to provide genuine engagement infrastructure, farming behaviour meant a continuous investment in verification that kept getting harder to maintain. Not because the technology was failing. Because the thing you were defending against kept getting more sophisticated at exactly the same rate you were getting better at defending against it.</p><p>Odaily’s review of airdrop history from 2023 to 2026 described how mechanisms evolved from genuine community rewards into what it called a systematic process of harvesting retail users, ultimately leading to a complete collapse of community trust. That collapse extended to the infrastructure layer. If farming was endemic, how do you prove the engagement you verified was real. And if you cannot prove it simply, how do you price the verification.</p><h3>What this meant for building here</h3><p>We built AirLyft’s verification system specifically because of this problem. On chain action verification rather than social post screenshots precisely because we had watched the gap between what was claimed and what was real become too wide to ignore.</p><p>On chain verification addressed the activity layer. But the content layer needed a different approach. Campaign submissions, posts, written entries, content responses, still needed to be checked for whether they genuinely met the campaign requirements rather than being copied, generic, or entirely off brief. We built an AI based content verification layer for this. Campaign owners could define what a valid submission looked like and each entry would be processed automatically, checked against those requirements, and approved or rejected before any manual review was needed. Redundant posts, unauthorised content, and submissions that had nothing to do with the actual campaign brief got filtered out. The ones that met the standard got through. It removed hours of manual review from projects running campaigns at any real volume and closed another gap that farming behaviour had been exploiting.</p><p>But building the right answer to a problem that keeps escalating is not the same as having a sustainable business model for doing it.</p><p>Projects that deeply understood the farming problem valued what we built. They had seen their own campaign data corrupted by inorganic participation. They knew what genuine verification was worth. The projects that had not yet confronted the reality of their farming exposure tended to treat verification infrastructure as a commodity. Something that should cost close to nothing because the metrics it produced looked similar to the cheaper version until you looked closely enough to see the difference.</p><p>That asymmetry in understanding was one of the persistent challenges of operating in this space. You were solving a problem at scale for people who were only beginning to understand they had the problem.</p><h3>The cycle that nobody completed</h3><p>Part of what we kept working on was the education problem alongside the technology problem.</p><p>The engagement cycle in web3 breaks at a specific point that the category still has not fully addressed. A user joins a campaign. They complete the required actions. They collect the reward. And then there is nothing left for them to do. No next step. No reason to stay. No follow-on campaign or task that gives them a reason to come back to the product they just interacted with.</p><p>A user in that position is not disloyal. They are simply finished. And a finished user in a token economy has one natural next action. Sell what they received.</p><p>We wanted the ecosystem to understand campaigns as repeating cycles rather than single events. The effort to reward ratio across a campaign, what a user has to do relative to what they earn, shapes whether they come back or move on. The funnel does not end at participation. It continues into whether there is something worth doing next. Whether the user who showed up today has any reason to be there in two weeks.</p><p>What onboarding actually means. What recurring campaigns are for. How a community is built from the inside out rather than filled from the outside in. These were the conversations we kept trying to start with the teams that were running campaigns through us.</p><p>That part of the problem has not been solved by us or by anyone else building in this space. It requires a different way of thinking about what a campaign is for. Not an event that drives a number. A moment in a longer relationship with a user who has reasons to stay.</p><h3>A note on what comes next</h3><p>In this note we shared what farming behaviour actually did to the economics of building campaign platforms. The scale of the farming problem across the category. The arms race between verification and evasion that made defending genuine engagement a continuous and escalating cost. And what that meant for platforms trying to price the value of real verification in a market where many projects had not yet fully understood what they were losing to the alternative.</p><p>Not a recovery narrative. Not a roadmap. Just what building in this specific environment required and what it cost.</p><p>These notes are part of an ongoing series from the team. We will keep sharing as the work continues.</p><p>– The Team</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ae803708dd90" width="1" height="1" alt=""><hr><p><a href="https://blog.kyte.one/what-farming-behaviour-did-to-the-economics-of-building-campaign-platforms-ae803708dd90">What farming behaviour did to the economics of building campaign platforms</a> was originally published in <a href="https://blog.kyte.one/">Kyte.One</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why web3 never found a sustainable way to pay for its own infrastructure]]></title>
            <link>https://blog.kyte.one/why-web3-never-found-a-sustainable-way-to-pay-for-its-own-infrastructure-2416701571ac?source=rss-1d32682518c8------2</link>
            <guid isPermaLink="false">https://medium.com/p/2416701571ac</guid>
            <dc:creator><![CDATA[AirLyft]]></dc:creator>
            <pubDate>Tue, 25 Aug 2026 14:03:40 GMT</pubDate>
            <atom:updated>2026-08-25T14:03:40.665Z</atom:updated>
            <content:encoded><![CDATA[<p>Web3 has a payment problem it has been avoiding talking about honestly.</p><p>Not the kind that gets solved by a new layer or a better wallet. The kind that comes from a category that trained its builders, its users, and its investors to expect infrastructure to be free or close to it.</p><p>We built AirLyft trying to solve the engagement problem in web3. And one of the hardest things we kept running into was not a product problem. It was this payment problem.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Ib0D9CbfyGQfZAaO9i1x7w.jpeg" /></figure><h3>How the grant culture worked</h3><p>For most of web3’s history the infrastructure that made the category function was not paid for by the people using it. It was paid for by foundations. By protocol treasuries. By grant programs funded by token sales that were themselves funded by the expectation that the token would appreciate.</p><p>Open call grants paid upfront to whoever agreed to take on the work. Research tracking developer migration in web3 noted that most developers who left the category in 2026 had likely entered under exactly that grant logic. Ecosystems competed to attract builders by subsidising the tools those builders would use. Analytics platforms. Indexing services. Growth tools. Campaign platforms. The implicit deal was that the protocol treasury would fund these things while the ecosystem was building toward a future where the tools would eventually pay for themselves.</p><p>AirLyft was not grant funded. We built the platform trying to generate real subscription revenue from the projects using it. That put us at odds with the market from the start. The tools we were competing alongside were often subsidised by protocol treasuries or foundation grants. Being the platform charging a stable subscription in that environment meant swimming against the current before the current even started running against us.</p><h3>What token funded infrastructure actually meant</h3><p>In practice funding infrastructure through token programs meant that the revenue underneath those tools was always one market cycle away from becoming unsustainable.</p><p>A comprehensive web3 fundraising analysis published in early 2026 found that only twelve percent of tokens launched in 2025 traded above their token generation event price by year end. The grant programs and subsidies keeping infrastructure tools running were funded by treasuries that had, in the majority of cases, declined sharply from the values the original funding assumptions were based on.</p><p>We experienced a version of this directly. When we initially built subscription access using KTE, the cost of access moved with the token price rather than staying fixed. When KTE was higher, access cost more in stable terms. When it fell, the economics shifted again. Neither direction produced the stable predictable pricing a subscription model needs to function. We have written about this before. It is the same structural problem the category built for everyone running infrastructure on token economics.</p><p>Even NEAR Protocol acknowledged the scale of the problem. A proposal for a NEAR Sovereign Fund published in 2026 described what it was trying to address as the chronic underfunding of web3 public goods. The proposal noted that infrastructure tools the ecosystem depended on had no sustainable revenue model and remained dependent on foundation grants that fluctuated with market conditions.</p><p>That is not a small structural crack. That is a structural failure the category had been deferring for years.</p><h3>What sustainable actually looked like</h3><p>The tools and platforms that kept running through the contraction were the ones that had found a way to charge for their services in real terms. Not through token subsidies. Not through grant dependencies. Through subscriptions or fees or contracts that paid regardless of what the market was doing.</p><p>We moved to stable subscription pricing specifically because of this. Not because it was the path of least resistance. Because it was the only model with a chance of being real. We have written about how that transition worked and how hard the resistance was in a market conditioned to expect otherwise.</p><p>Cryptonews published a piece in 2025 whose headline said it plainly. Web3 cannot rely on grants forever. The category was going to have to find a way to generate real revenue for the infrastructure it needed or watch that infrastructure disappear when the next cycle turned.</p><p>The category found out which version of that was true in 2025 and 2026.</p><h3>What the resistance actually was</h3><p>The platforms that tried to build subscription revenue in this category faced something specific. Not because the product was not valuable. Because the market had been conditioned for years to expect infrastructure to cost a token amount that moved with speculation rather than a stable price that reflected the actual cost of building and maintaining the service.</p><p>The resistance we encountered was not random and it was not about us. It was the direct result of years of a category subsidising its own infrastructure with money that was contingent on token values holding. When that contingency was tested the subsidies disappeared and so did the projects that had relied on them.</p><p>That is the structural problem underneath the specific difficulties we have described across recent newsletters. Not just ours. Every platform that tried to charge real money for real infrastructure in a market trained to expect it for close to nothing.</p><h3>A note on what comes next</h3><p>In this note we shared the structural reason web3 never found a sustainable way to pay for its own infrastructure. The grant culture that trained the market to expect tools to be free or nearly free. How we experienced that resistance directly when trying to build stable subscription revenue. What happened when the token subsidies funding those grants declined across the category. And what sustainable looked like for the platforms that figured it out.</p><p>Not a recovery narrative. Not a roadmap. Just the structural explanation for something the category has been living with and has not fully acknowledged.</p><p>These notes are part of an ongoing series from the team. We will keep sharing as the work continues.</p><p>– The team</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2416701571ac" width="1" height="1" alt=""><hr><p><a href="https://blog.kyte.one/why-web3-never-found-a-sustainable-way-to-pay-for-its-own-infrastructure-2416701571ac">Why web3 never found a sustainable way to pay for its own infrastructure</a> was originally published in <a href="https://blog.kyte.one/">Kyte.One</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[When the projects you built for start closing: what happens to the platforms that served them]]></title>
            <link>https://blog.kyte.one/when-the-projects-you-built-for-start-closing-what-happens-to-the-platforms-that-served-them-f2422714fc7b?source=rss-1d32682518c8------2</link>
            <guid isPermaLink="false">https://medium.com/p/f2422714fc7b</guid>
            <dc:creator><![CDATA[AirLyft]]></dc:creator>
            <pubDate>Thu, 20 Aug 2026 16:00:38 GMT</pubDate>
            <atom:updated>2026-08-20T16:00:38.051Z</atom:updated>
            <content:encoded><![CDATA[<p>Something happens when your clients start disappearing that is different from anything else that can go wrong in a business.</p><p>It is not like losing a deal. It is not like a product that is not working. It is watching the ecosystem you built for shrink around you. One project at a time. Until the thing you were serving is smaller than it was when you started.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hTwTh9pLLyyR-MOrdOLaHA.jpeg" /></figure><p>That is what 2026 has looked like for a lot of platforms in this space. Including us.</p><h3>One every two days</h3><p>By mid 2026 Foresight News was tracking the pace of web3 project closures across the previous 600 days. Their number was one project dying every two days. Not scams. Not fraud. Projects that had raised real money, built real products, and still found that the market had moved past them before they could get to sustainable.</p><p>The closures we have been writing about in recent notes are not abstract to us. They are the category of teams we built for. Projects trying to reach real users. Projects that needed a way to understand who their community actually was and what would make them come back. That is the problem we were working on and those are the teams we were working with.</p><p>When they close, a platform that built for those teams does not just lose a customer. It loses one more proof point that the problem was real and that the work mattered.</p><h3>The tooling squeeze</h3><p>What is harder to see from the outside is that the closures were not limited to the projects themselves. The platforms that served those projects started going too.</p><p>Zapper had been running for seven years. It was one of the most widely used portfolio and analytics tools in web3. It shut down on August 3, 2026. The phrase that appeared in coverage of the closure was specific. A casualty of the analytics tooling squeeze.</p><p>DappRadar went. Parsec went. Step Finance went. These were not failed experiments. They were infrastructure. The tools that the ecosystem was using to understand itself. And when the ecosystem contracted enough that the projects running on it stopped needing those tools at the rate they once did the tools became unsustainable.</p><p>There is a line from the analysis of the 2026 closure wave that has stuck with us. Infrastructure tools struggled to turn developer adoption into paying customers. That is not a marketing problem or a product problem. It is a structural problem. And it is one that platforms like ours have felt directly.</p><h3>Where that leaves us</h3><p>We built AirLyft to serve web3 projects. The campaigns they ran. The users they were trying to reach. The question of whether the people showing up for their campaigns were real users or just wallets optimised to collect rewards.</p><p>That problem is still real. The projects that were trying to solve it are fewer than they were. Some closed. Some pulled back to the point where sustained campaign work was not something they could commit to.</p><p>We are not outside this story looking in. We are inside it. That is the honest version of where things are.</p><h3>A note on what comes next</h3><p>In this note we shared what happens downstream when the projects you built for start closing at scale. The pace of closures across the category. The platforms that served those projects and then could not sustain themselves without them. And where that leaves a platform that built its work around this market.</p><p>Not a recovery narrative. Not a roadmap. Just what this period has actually looked like from the inside of it.</p><p>These notes are part of an ongoing series from the team. We will keep sharing as the work continues.</p><p>– The team</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f2422714fc7b" width="1" height="1" alt=""><hr><p><a href="https://blog.kyte.one/when-the-projects-you-built-for-start-closing-what-happens-to-the-platforms-that-served-them-f2422714fc7b">When the projects you built for start closing: what happens to the platforms that served them</a> was originally published in <a href="https://blog.kyte.one/">Kyte.One</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What it takes to keep something like this alive]]></title>
            <link>https://blog.kyte.one/what-it-takes-to-keep-something-like-this-alive-4d4b3ccf6423?source=rss-1d32682518c8------2</link>
            <guid isPermaLink="false">https://medium.com/p/4d4b3ccf6423</guid>
            <dc:creator><![CDATA[AirLyft]]></dc:creator>
            <pubDate>Tue, 11 Aug 2026 14:49:20 GMT</pubDate>
            <atom:updated>2026-08-11T14:49:20.271Z</atom:updated>
            <content:encoded><![CDATA[<p>We have been asked this question in various forms. Sometimes directly. Sometimes in the space between messages from the community.</p><p>What is actually keeping you going.</p><p>The honest answer is the belief that the problem is still worth working on.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*EmJUDnuhOEgFIL_y7ZgSzw.jpeg" /></figure><h3>Because the problem has not gone away</h3><p>Here is the honest state of web3 engagement in 2026.</p><p>Only five to ten percent of users become repeat dApp users within thirty days of their first interaction. Fewer than twenty percent of web3 wallet users return to a product after seven days. Seventy percent of wallet users make one transaction and then leave entirely. Industry research consistently shows that most dApps lose the significant majority of potential users within the first minutes of interaction and poor onboarding is almost always a primary reason.</p><p>Those numbers are not from a bad quarter. They are from right now. They represent the current measured state of user retention across the category. The problem we set out to work on is not only still there. It is documented, widely acknowledged, and still largely unsolved.</p><p>We said in an earlier note that we have been trying to answer one question for five years. Who are the users actually sticking. Why do they leave. What would it take to keep them. Those questions are still not answered by anything the category has built. Not by us. Not by anyone.</p><p>That is frustrating. It is also the clearest reason for why the work has not stopped.</p><h3>It takes believing the problem is worth it</h3><p>The straightforward version of how this was supposed to go was that we would build something useful, the market would gradually recognise it, and growth would follow from that. Markets in this space rarely follow the straightforward version. The capital moved toward AI and away from the category we were building in when the cycle turned. The projects that needed what we built were fighting their own treasury constraints at exactly the moment we were trying to show them what sustained engagement could look like.</p><p>Maintaining belief through all of that requires stripping everything else away and still finding the problem worth working on.</p><p>We still find it worth working on.</p><h3>It takes a team that chose the mission over comfort</h3><p>The people involved in this project had other options. The same skills that built AirLyft could have been applied elsewhere.</p><p>They stayed. Not because it was the path of least resistance. Because the problem felt worth it.</p><p>That kind of commitment does not get announced. It gets demonstrated in the quiet decisions nobody sees. The calls at hours that would have been easier to ignore. The maintained integrations for ecosystems that had gone quiet. The campaign ideas sent to partners because waiting felt wrong even when the return was uncertain.</p><h3>It takes absorbing things you did not choose</h3><p>The Luna crash was not something we planned for. The staking partners that changed or wound down were not decisions we made. The market contraction that turned project treasuries into survival budgets was not something any individual team could have prevented.</p><p>What you can choose is what you do after. This team kept finding a way.</p><p>That is not a heroic statement. It is just what happened.</p><h3>It takes staying honest with the people who stayed with you</h3><p>The retention numbers across web3 are not just a product problem. They are a trust problem. Teams keep running campaigns that do not tell them who their real users are. Users keep arriving and leaving without anyone fully understanding why.</p><p>This newsletter series is part of what keeping something alive requires. Not because transparency is a strategy. Because the community that has stayed with this project through a genuinely difficult period deserves honesty about what that period has looked like from the inside.</p><h3>What we still believe</h3><p>We still believe that five to ten percent repeat user rates within thirty days should not be accepted as the standard for a category that claims to be building the next version of the internet. That the gap between active wallets and genuinely engaged users is the most important unsolved problem in web3 right now. That the work done over the last several years to narrow that gap matters even when the conditions around us have not made it easy.</p><p>And we still believe the community of people who have stayed with this deserves to hear from us honestly and regularly.</p><h3>A note on what comes next</h3><p>In this note we shared what it actually takes to keep something like this alive. The data behind why the problem still matters. The belief that outlasted the easy version of the story. The team that chose to stay. The external shocks that had to be absorbed. And the commitment to staying honest with the people who stayed with us.</p><p>Not a roadmap. Not a recovery narrative. Just what this has required and what we still believe.</p><p>These notes are part of an ongoing series from the team. We will keep sharing as the work continues.</p><p>– The team</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4d4b3ccf6423" width="1" height="1" alt=""><hr><p><a href="https://blog.kyte.one/what-it-takes-to-keep-something-like-this-alive-4d4b3ccf6423">What it takes to keep something like this alive</a> was originally published in <a href="https://blog.kyte.one/">Kyte.One</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>