Hi everyone,
Welcome to my first quarterly portfolio update. As I mentioned in January, I have decided to shift my public portfolio updates from monthly to quarterly. While I still conducted my own internal reviews at the end of January and February, stepping away from posting every 30 days has been fantastic for my investing psychology. It forces me to look at the long-term trajectory of these businesses rather than agonizing over month-to-month stock price fluctuations.
Here is my portfolio as of the end of March 2026, alongside the tracking from the previous months:
2026 YTD Return: ~ -40%
Drawdown from Nov 2025 Peak: -55%
Portfolio Tracking (Sorted by End of March Allocation):
| Ticker | End of Mar | End of Feb | End of Jan | End of Dec ('25) |
|---|---|---|---|---|
| IREN | 21.05% | 23.20% | 23.84% | 17.80% |
| NBIS | 17.93% | 15.31% | 11.20% | 10.96% |
| ALAB | 17.26% | 17.21% | 12.44% | 10.54% |
| APP | 16.67% | 17.95% | 15.30% | 22.99% |
| NVDA | 13.42% | 13.40% | 15.01% | 14.89% |
| NET | 6.52% | 5.22% | 4.21% | 4.79% |
| RBRK | 6.20% | 6.65% | 5.61% | 8.40% |
| HIVE | 0.96% | 1.06% | 1.09% | 0.51% |
| EOSE | 0.00% | 0.00% | 11.30% | 9.10% |
Q1 Portfolio Changes (Specifically in February):
-
Sold out of EOSE: Completely closed my position.
-
Trimmed NVDA / Added ALAB: I sold about 25% of my NVDA position and used the proceeds to build up my ALAB allocation.
Reflections and Lessons Learned in Q1
1. The EOSE Mistake: Trust and Management
I sold out of EOSE in February, and frankly, looking back, I never should have bought it in the first place. I never had high conviction in the business. The final straw was their recent earnings call, where the management team came across as highly unprofessional, and their actual reported numbers drastically missed their previous guidance.
Lesson Learned: When evaluating early-stage companies with low revenue, we have to be incredibly strict. An inexperienced or unreliable management team cannot be trusted with our capital. If the conviction isn’t there from the start, it doesn’t belong in the portfolio.
2. The AI Infrastructure Thesis: Exponential Token Burn & SaaS Disruption
Compared to a year ago, my portfolio has heavily rotated into AI Infrastructure. Despite the recent market volatility, my conviction here remains sky-high.
If we look at the actual utility of AI over the past couple of years, the evolution of compute consumption is staggering:
-
Phase 1 (Chatbots): Having a basic conversation with ChatGPT consumes a relatively tiny amount of tokens. The compute cost is minimal, maybe less than $1 a day per power user.
-
Phase 2 (Coding Agents): Recently, we’ve seen the rise of agentic coding. A single developer can spend 10 hours and $100 to burn through ~10 million tokens “vibe coding” a custom, personalized task management app (like a lightweight Monday.com). This is exactly why I am no longer bullish on many simple, high-fault-tolerance software SaaS companies. If a user can use AI to spin up a “good enough” custom alternative for cheap, those traditional SaaS moats are dead. In contrast, mission-critical sectors with zero fault tolerance—like cybersecurity (e.g., RBRK, NET)—are much more insulated from this AI disruption.
-
Phase 3 (Super Agents): Now, with tools like the OpenClaw super agent, AI is highly accessible to everyday, non-technical users. Normal people can easily burn through 10 to 50 million tokens in a single day automating daily workflows.
AI is becoming exponentially more useful, and as a result, the “token burn rate” per user is growing exponentially. I believe the market is currently underestimating how this shifts the demand curve. With compute demand scaling at this magnitude, the downstream requirement for AI infrastructure will continue to rise exponentially. I am positioning my portfolio to capture that reality.
3. Surviving the Macro Shocks
My portfolio is down roughly 40% YTD and 55% from my November 2025 high. This has been a brutal stretch, driven by wave after wave of macroeconomic and geopolitical shocks. We went straight from the AI Capex “Show Me the ROI” rotation (Dec-Mar), to sticky inflation and the death of 2026 rate cut hopes, and most recently, the massive market disruption and oil price shock caused by the US-Iran war.
It is painful, but it reinforces a core truth: No one can perfectly predict the market or macroeconomic environments.
Trying to time these massive macro and geopolitical events is a losing game. To achieve long-term profitability, we have to stick to what we are actually good at—evaluating individual businesses—while strictly managing our position sizing and risk.
Thank you all for the great discussions this quarter. I look forward to reading your updates.