Serenity’s account just printed a 49.4% drawdown from its December 2026 peak. Not a flash crash. A slow bleed. Most retail portfolios at that level would have triggered a margin call or a panic liquidation. Serenity? He posted a one-liner: "Remains optimistic."
Most will dismiss this as ego. I see something else: a masterclass in volatility harvesting. Let's cut through the noise.
Context: The AI Bottleneck Portfolio
Serenity’s concentration is three small-cap AI infrastructure stocks: AXTI (AXT Inc.), SIVE (Sivers Semiconductors), AAOI (Applied Optoelectronics). All operate in the "bottleneck" layer of AI compute – optical interconnect and semiconductor substrates. His thesis: as AI clusters scale from thousands to tens of thousands of GPUs, the optical link between them becomes the limiting factor.
- AXTI supplies indium phosphide (InP) substrates for 800G/1.6T laser chips. Global #2 behind Sumitomo.
- SIVE develops photonic chips for co-packaged optics (CPO). Revenue from AI: less than 5% today.
- AAOI builds high-speed transceivers with vertical integration – they self-source lasers. Gross margin: 28% and compressing.
Serenity’s bet: a revenue inflection in H2 2027 when 1.6T deployments go from samples to volume. He turned $1 into $45 in 2026 on the back of the 800G super-cycle. Now the account sits at ~$22.7 for every original dollar. He can stomach a 49.4% drawdown because his cost basis is effectively zero.
Core Analysis: The Math of Survival
Let me run the numbers the way I audit a liquidity pool.
Assume initial capital = 1 unit (say, $10k). Peak after 4502% = 46.01 units. Drawdown of 49.4% from peak = 46.01 * 0.506 = 23.28 units. Still a 23.28x return. The real pain starts only if the drawdown exceeds 95% from peak – that would take him below the original capital.
Most traders forget that absolute return matters more than percentage drawdown. Serenity isn’t losing money. He’s giving back a portion of asymmetric gains. The market is testing his conviction, not his solvency.
Where did the drawdown come from? Three forces:
- Macro rotation: January 2027 saw a shift from growth to value, hitting high-beta stocks hard. AAOI dropped 22% in a week on zero news.
- Competition fear: Reports of Chinese module makers ramping 800G at 15% lower ASP. AAOI’s gross margin compressed from 32% to 28% in Q4 2026.
- Timing skepticism: Funds started discounting the H2 2027 inflection, pricing it earlier. When real orders didn’t materialize by early 2027, the premium evaporated.
All three are temporary. The structural tailwind remains: AI cluster scale is doubling every nine months. Optical throughput per GPU must double to keep up. The suppliers are capacity-constrained.
Contrarian: The Drawdown is a Gift, Not a Warning
The consensus read: 49.4% drawdown = thesis broken. That’s retail logic. Smart money reads the opposite.
If the H2 2027 inflection is real, then the current drawdown represents an entry point with a better risk/reward than the peak. Why? Because the upside is the same, but the entry price is 49% lower. Serenity knows this. He posted “optimistic” not as a prayer, but as a signal that he’s adding.
I’ve seen this pattern before. In 2020, during the DeFi Summer, I accumulated UNI-V2 LP tokens after a 40% dump in the CLP pool. Panic sells, liquidity buys. The market forgets that drawdowns redistribute wealth from the weak hands to the prepared.
But there is a contrarian risk inside the contrarian trade: the H2 2027 timing might slip. Technology cycles are rarely linear. 1.6T modules could get delayed to 2028. CPO could disrupt traditional transceivers faster than expected. SIVE’s wafer-level integration is unproven at scale. If the inflection moves to 2028, Serenity’s portfolio could bleed another 40% in the interim.
That’s why position sizing matters. Serenity’s account is likely a high-conviction satellite, not his core net worth. Without knowing his leverage (if any), I assume he’s using no debt. The 49.4% drawdown doesn’t wipe him – it tests his thesis.
Takeaway: Actionable Levels and Catalysts
If you want to follow Serenity’s playbook, here’s the structure:
- Don’t buy the whole basket. Only AAOI has the near-term revenue catalyst (800G still growing). AXTI is a proxy. SIVE is a lottery ticket. Allocate 70% to AAOI, 20% to AXTI, 10% to SIVE.
- Set a timing trigger. Buy when AAOI drops below $X (its book value), not when it’s down 20%. Use limit orders.
- Monitor one catalyst: AAOI’s Q1 2027 earnings call (April). If management guides 800G revenue grew sequentially and confirms 1.6T samples with a hyperscaler, the drawdown was a buying opportunity. If they lower Q3 guidance, exit.
- Code doesn’t care about your feelings. The market will push the price below your pain threshold. Have a stop-loss at 25% below your entry.
Serenity’s 49.4% drawdown isn’t a failure. It’s a re-pricing of long-duration risk. Yield is the bait, rug is the hook – but only if you don’t survive the volatility. The real question: will H2 2027 deliver?
Track the order flow. Ignore the noise. The bottleneck doesn’t disappear because the price drops.
