Investor Identity
Melvin has one thesis, and everything he owns is a different expression of it: AI needs an enormous amount of physical infrastructure, and building it is going to take longer and cost more than anyone currently thinks.
Memory chips, data centers, networking gear, power, robots - those aren't five bets to him. They're five doors into the same room.
Where he plays
AI infrastructure, top to bottom. Memory semiconductors, cloud providers renting out GPUs, optical networking, chip designers, power, and humanoid robotics, plus a handful of smaller satellite bets around the edges.
Background
Melvin is an AI writer at Milk Road, focused on value-driven equity investing, particularly within AI and technology.
His approach centers on identifying companies whose long-term fundamentals are stronger than current market sentiment suggests, and sharing those opportunities with the PRO community.
Portfolio Structure
How he invests
His entire philosophy fits in one Warren Buffett line: “Anytime stocks go down, I like it. I am a net buyer of stock.”
When everyone else is panicking, Melvin's first instinct is to go look at what just went on sale. That's not a personality quirk, it's the strategy. Most of his best entries came from somebody else's bad week.
He runs a wider book than most conviction investors, usually over 15 positions plus cash, with two names doing the heavy lifting and everything else sized down from there. There's no maximum position size because conviction sets the size.
Core infrastructure names can get big. Speculative moonshots stay deliberately small.
These are long-term holds by default, but not on autopilot. Money moves when the story changes or when something better shows up.
What he's looking for (and what he avoids)
Two things get his attention:
- A structural wave in AI infrastructure the market hasn't fully priced yet
- A selloff or dislocation that puts a good company on sale for reasons that have nothing to do with the company
Everything also has to clear one gate: a real, established business. Not a story that could evaporate overnight.
On the way out, his rules are guidelines rather than triggers. He usually pulls his original money off the table once a position doubles, unless conviction is unusually high. A 10-15% trim is common around a 50% gain. Losses get cut early instead of nursed. And taxes never get a vote in the investing decision.
But the real reason he sells is because the story broke, or he found somewhere better to put the money.