We don't talk enough about the emotional architecture of conviction. Last week, a friend sent me a screenshot of an investor's public portfolio — down 49.4% in a year. The tickers were obscure: AXTI, SIVE, AAOI. Small-cap AI suppliers. The investor, Serenity, had been a legend in niche Twitter circles for turning a small stash into 4,500% returns over 26 years. But now his account was bleeding. The community was split: some called it recklessness, others called it diamond hands. I called it a mirror.
Because I've seen this pattern before. Not in AI stocks, but in crypto. In the same way Serenity is betting on the bottlenecks of artificial intelligence — optical modules, substrate materials — we in Web3 bet on the plumbing of decentralized finance: sequencers, bridges, data availability layers. The same emotional muscle is required. The same risk of being early. The same likelihood of being caught in a drawdown that feels like the end of the world, but is actually just the beginning.
The Context: Bottlenecks and Beliefs The article that dissected Serenity's portfolio was written by an AI investment analyst, but it could have been about any crypto native who went all-in on scaling solutions in 2021. Serenity's thesis was simple: AI's physical infrastructure has pinch points — components that cannot be easily substituted, supplied by small, underfollowed companies. He identified three: a photonics firm (AAOI), a substrate materials maker (AXTI), and a specialised sensor company (SIVE). All were down 40-60% from their peaks. Yet Serenity didn't panic. He said he had already taken his original capital out years ago; this was just profit erosion. He called it "a test of conviction" and predicted a revenue inflection point by the second half of 2027. The analyst gave that prediction a B-minus confidence.
Now, replace the tickers with ETH, ARB, OP, or any Layer-2 token. Replace 2027 with the next Bitcoin halving cycle. The narrative is identical. In 2022, many of us watched our portfolios drop 80%. We said the same thing: "We've been here before. Fundamentals haven't changed. The bear market didn't destroy the technology." And we were right. But that doesn't make the drawdown less painful, or less instructive.
The Core: Time Arbitrage and Technical S-curves Let me tell you a story from 2017. I was a 20-year-old CS student in Nairobi, manually auditing the Ethereum smart contract code of The DAO hack. I spent 150 hours tracing the reentrancy vulnerability, realising that code wasn't just logic — it was a social contract. That experience taught me that the biggest risks in decentralized systems aren't technical; they're misaligned incentives and premature celebration. Serenity's AI stocks face a similar risk: the technology might work, but if the market's timeline is shorter than the innovation's S-curve, the price gets crushed.
The analyst noted that Serenity's 2027 inflection point aligns with the typical 3-5 year lead time for data centre buildouts. In crypto, we see the same pattern: Ethereum's Layer-2 scaling was promised for 2022, but meaningful adoption really hit in 2024. Those who bought into the hype too early — say, during the 2021 bull market — endured a 60-80% drawdown before the thesis played out. The key question isn't whether the thesis is correct; it's whether you have the capital, the psychology, and the time horizon to survive the valley.
The Contrarian: The Blind Spots of Conviction But here's the counter-intuitive truth that Serenity's cheerleaders ignore: conviction can be a trap. The analyst rated the risk of technological substitution as "high" — an optical module like AAOI could be replaced by co-packaged optics (CPO) from larger players. In crypto, we've seen similar: ZK-rollups were supposed to dominate, but Optimistic rollups won the first battle on EVM compatibility. Many small-cap crypto protocols that promised to be "the bottleneck" are now dead or forgotten. Binance's BNB Chain convinced projects to deploy with low fees, and the OP Stack vs ZK Stack war is not about technical superiority — it's about who can convince more projects to deploy first. That's a business development challenge, not a code challenge.
Serenity's portfolio is also missing a critical hedge: what if the AI capex cycle peaks earlier? The analyst warned of a "J-shaped reversal" — overinvestment in 2025-2026 leading to oversupply and price collapse in 2027. In crypto, we call this the "infrastructure paradox": everyone builds for the next billion users, but if the users don't come fast enough, the capital dries up and the builders fail. The bear market didn't kill the protocols; it killed the ones without real traction.
About Me I wrote this not as a market commentator, but as a decentralized protocol PM who has lived through these cycles. In 2022, while others panicked, I researched ZK-rollup scalability solutions and built a visualization tool for proof generation times. That project didn't make money, but it taught me something valuable: the ability to focus on the technology during the drawdown is what separates survivors from casualties. Serenity's resilience is admirable, but his strategy is one of extreme concentration. Most of us can't afford that.
Takeaway So what do we take from Serenity's story? Not that conviction is wrong — we need believers to fund the bottlenecks of the future. But that conviction without diversification is a high-stakes bet. The next time you see a crypto project promising to be the "bottleneck" of the next billion users, ask yourself: is it the AI photonics of Web3, or just another optical illusion? The answer will emerge not from the whitepaper, but from the second half of 2027 — or whatever cycle your timeline demands. Until then, we build, we endure, and we keep our capital safe enough to survive the test.
