How We're Thinking About the Market

We've been slow. Probably too slow. The misses tend to follow a pattern: we knew the founders well but passed because the valuation or thesis felt off. Should have gone in anyway. The AI landscape has shifted fast, and figuring out where to place conviction versus where to stay disciplined has been a real-time learning process.

That said, we'd rather miss a few than repeat the 2021 vintage pattern: funds that deployed fast into hot rounds at peak valuations are largely underwater now. They followed social consensus instead of doing the work. The fact that Alex, Amit, and Matt are all actively building has been critical here. We're not sure how you make sound decisions in this space without doing the work yourself.

The goal is to make money, and that means staying disciplined in first and second rounds where the ownership economics work for our size. Building software costs have dropped so dramatically that you can bet on the founder almost independent of traction, but you still need to evaluate motivation, market, and whether the company accrues real equity value. A lot of capital is being raised right now; not all of it produces venture-scale outcomes. That's the filter.

Portfolio construction is changing too. The old playbook of anchoring on an ownership target doesn't hold up given how heterogeneous deals are now. We're shifting to a more holistic view: a spectrum of bet types rather than running the same play. That means a much higher bar for hot or overpriced deals when we do them (how big can it get, what's the funding trajectory, is it kingmaker-shaped, how strong is the market pull), continued discipline on the core early-stage book, and selective incubations like Zeer where outsized ownership creates real optionality. It's naive to assume cheap-but-great early-stage deals are sitting around waiting; a lot of that has been priced in. We have to think across the portfolio, not deal by deal.

We're also looking harder at areas that weren't on the chessboard when we started the fund: energy, deep tech, companies with real science behind them (Archetype and Andromeda are early examples). AI is the dominant theme. CrewAI, Archetype, Rubie, DataChain, and Zeer all have AI at their core. But we're backing infrastructure and tooling where durable value accumulates, not chasing the hype cycle. That's always been our thesis.

Odds and Ends

Beyond Fund II, the broader Earl Grey portfolio continues to develop. We've seen several companies reach unicorn status: SWORD Health, Hightouch, Writer, and NexHealth. We've also seen 10+ exits including acquisitions by Datadog, Dropbox, Fivetran, and more. Fund II will continue to benefit from a decade of relationships, a large portfolio/founder network, and that we are actively shipping.

One thing worth noting on the market: later-stage opportunities have become surprisingly attractive. You can get early-stage-type returns even after big growth rounds in certain AI companies. It's not what Earl Grey is doing today, but it's something we've had to check our assumptions on, and we could explore it or collaborate on opportunities if there's appetite.