Compounding conviction: notes from our 2026 annual meeting

Simultaneous with the five-year anniversary of our launch, earlier this week we hosted Asymmetric's annual meeting in Boston. Our LPs flew in, five exciting founders presented, and we spent the day doing what we like doing most: comparing notes with the people closest to the work. The theme we chose for the year was "compounding conviction". The more I sat with it over the course of the day, the more it felt like the right frame not just for our firm but for the current moment in which we're all operating.

I want to share a version of what I covered in my session, because I think the through-line is one that matters for anyone trying to make sense of where private tech is heading. The playing field has fundamentally changed. Most of the capital in the market is chasing the obvious version of the opportunity. And the firms that will produce outlier returns from this vintage are going to look structurally different from the consensus.

The playing field has changed, and there's froth in the obvious places

The first thing to say plainly is that AI is rewriting industry economics in a way I haven't seen before in my career. Not just "transforming workflows" in the soft sense. Actually rewriting the unit economics of how work gets done. Insurance claims that took a team of adjusters a week now take an agent a few minutes. Manufacturing operations are learning on the floor. A small group of engineers can now build software that would have required a hundred people and a decade five years ago. And a single doctor can manage a patient panel literally multiple orders of magnitude greater than they could before their job was assisted with AI. Capital needs, for an enormous swath of categories that historically required real money to attack, have collapsed.

This is the genuine, structural, industrial fact. Underneath it, though, you also have a second trend, which is the predictable, very loud hype cycle layered on top. Both are true at once. The technology is real and there is froth in some places. The job of an early-stage investor right now is to hold those two facts in your head simultaneously without letting either one make your head spin.

Most of the venture market is currently choosing one of two binary sides. The dominant strategy is to lean all the way into the hype: write large checks into mega-platforms and their copycats, underwrite huge TAM horizontal markets, accept winner-take-all dynamics, and bet that you've picked the one company that gets to a trillion dollars of enterprise value. The growth of very late stage single asset SPVs has perfectly mirrored this approach. The small minority contrarian view, louder on Twitter than in practice, is simply to sit things out and wait for the bubble to clear.

Our view is that neither of those is correct. The right course today is to be deeply engaged in the structural shift while being deliberately disciplined about where and how you express it.

We choose to play in the non-obvious half of the market

If you draw a line down the middle of today's tech investing market, the two halves look almost nothing alike.

On one side, where most of the money is going, you have mega-platforms and the copycats trying to land in the middle of them, founders who are deep San Francisco insiders spinning out of hyperscalers, businesses built around horizontal core infrastructure, capital-intensive by design, with exits that depend on winning the entire market. That's a fine game to play if you're a $5B platform fund. It is a terrible game to play with a $137M fund focused on the early stage. Forgetting anything about expected return, the companies are simply too capital consumptive, and raising at such high initial valuations, that achieving (and defending) significant ownership is basically impossible.

On the other side, where we choose to play, are deliberately sub-scale categories run by disciplined operators, founders who are deep industry insiders bringing novel domain insights to verticals Silicon Valley has historically ignored, capital-efficient business models, as well as multiple creative paths to exit: PE, public strategics, and private strategics. Not "we have to be acquired by Google or we go to zero." Real businesses that real buyers want to own for decades.

The pitch on this side of the market is straightforward. AI takes industries that used to be tech-resistant, like recurring home services, manufacturing, retail dental, multi-family leasing, and vertical marketplaces nobody wanted to touch, and makes them addressable for the first time. The historical playbook in those industries was cost cutting, financial leverage, blunt scale aggregation and service-quality compromises. The new playbook is AI-native operating models that genuinely deliver better outcomes for the end customer while being structurally more profitable. Those are the kinds of businesses we want to own through and after the cycle.

Why our fund size and strategy fit this moment

The hardest thing to do in venture right now is be honest about what size your fund should be. The temptation, when LP appetite is strong and you've had a good run, is to raise more. The math on that decision is brutal. The dominant industry behavior of raising as much as the market will give you does not match what actually returns capital to LPs in cycles like this one.

We have intentionally moved in the opposite direction. Our funds are deliberately size-constrained because the math of "outlier MoIC" gets harder, not easier, as you scale. A right-sized capital base lets us be surgical about entry, write checks that meaningfully move the needle in companies we believe in, and not be forced to deploy into deals we don't love just to put the money to work. It also means that when one of our companies has an outlier outcome, that outcome actually shows up in the fund. Our Fund I has five investments that we think may each individually return the entire fund - and many more we expect to deliver a real fraction thereof.

Paired with that is a deliberate concentration. Fewer bets, larger ownership stakes, and a willingness to re-up early in our highest-conviction companies before the market reprices them. AI has compressed the timelines on which great companies become obvious. The window between "this is a thesis bet" and "this is priced in" is now months or quarters rather than years. If you wait for the proof points the consensus is looking for, you're paying a much higher price for the same business. We'd rather underwrite earlier, at the team and thesis stage, and follow on aggressively in the ones that work.

That approach only makes sense if you have the operating depth to be confident at the team and thesis stage. It's why our work is concentrated in a small number of verticals where we have real expertise: healthcare delivery, recurring home services, industrial and manufacturing, and tech-enabled serial acquisition. We're not spread across whatever happens to be trending. Domain depth is what lets you say yes early and quickly and mean it.

The shape of the return profile we're trying to build

There's a phrase we kept coming back to on Monday, which is "venture upside with PE consistency." That's the construction we're continuing to build. We want the right tail, the genuinely fund-returning outcomes that justify the asset class, but we want it on a base of companies that are real businesses, with multiple paths to liquidity, that we'd be happy to own through any market.

What that looks like in practice is something close to the opposite of what you might expect from a venture firm in 2026. Smaller checks aimed at larger outcomes. AI as the enabler of value creation, not the product being sold. Capital-efficient companies, not capital-intensive ones. Deep verticals over broad horizontals. Founders chosen for industry depth over pedigree. A portfolio you can describe company by company without resorting to "and one of these will be the next OpenAI."

It's a less glamorous version of venture than the one you'll read about on the front page of TechCrunch. We think that's exactly why it works. The opportunities at the obvious end of the market are well-known and aggressively priced. The opportunities at the non-obvious end, in industries that are genuinely being remade by AI but where the capital hasn't yet arrived, are where this era's outlier fund returns will come from.

Compounding conviction is a view on how we do the work. It means repeatedly doing something consistent, with discipline, for long enough that the advantages stack on each other. Right-sized fund. Right verticals. Right founders. Right cadence of follow-on. Done over and over, with patience, in a market that mostly wants to do something else.

That's what we talked about on Monday. That's what we're going to keep doing.

Rob

‍