When There's Nothing to Diligence

Last Tuesday, a company called AMI Labs raised $1 billion.

Let that land for a second. One billion dollars. And here's the part that made me almost spit out my coffee during our Monday partners meeting: they have no product. No revenue. The first year will be devoted entirely to research.

The investors? Nvidia. Samsung. Toyota Ventures. Jeff Bezos. Eric Schmidt. People who presumably know what they're doing. And yet - by every framework I learned in my first year at The Firm, this deal makes no sense.

When I left BigTechCo to join a VC firm, I thought my superpower was obvious. I could read code. I could spot a brittle architecture from a mile away. I could ask the CTO questions that made them sweat - not because I was trying to be adversarial, but because I actually understood what "eventual consistency" means in practice, not just on a whiteboard.

Technical due diligence was supposed to be my moat. And for a while, it was. I'd sit in diligence meetings and catch things that my finance-background colleagues couldn't: the startup claiming "AI-powered" when it was really a bunch of if-else statements. The company whose "scalable infrastructure" was held together by a single senior engineer's tribal knowledge. The SaaS product that would need a complete rewrite to handle 3x growth.

That playbook still works. It works great, actually - 70% of investors now require technical due diligence, and the market for it is projected to nearly double over the next decade. SWE skills in VC have never been more valued.

But AMI Labs broke my brain a little.

Because the new game - the one where the biggest checks are being written - isn't about evaluating what exists. It's about evaluating what might exist. And that requires a completely different set of muscles.

Here's what I've been wrestling with. In the old model of technical DD, I'd evaluate a company across a clear rubric: code quality, architecture scalability, security posture, cloud unit economics, technical debt. These are measurable things. You can put numbers on them. You can write a memo that says "this system will need a $2M rewrite in 18 months" and everyone in the room understands what that means.

But when a company has no code, no architecture, no system - what's left?

People. And ideas. And the conviction that certain people with certain ideas will figure it out.

I've been doing this job long enough now to realize that this isn't actually new. The best early-stage investors have always been making people bets. What's new is the scale. We're not talking about a $5M seed round where you're betting on a founding team's potential. We're talking about a billion dollars deployed against a research thesis.

The Flyover cuts through the noise mainstream media refuses to clear.

No spin. No agenda. Just the day's most important stories — politics, business, sports, tech, and more — delivered fast and free every morning.

Our editorial team combs hundreds of sources so you don't have to spend your morning doom-scrolling.

Join 2.3 million Americans who start their day with facts, not takes.

Start Reading for Free

After the AMI Labs news hit, I spent a few days talking to colleagues who've done diligence on similar pre-product, research-heavy deals. Here's what I've pieced together about how the evaluation actually works:

First, publication and patent analysis replaces code review. Instead of reading a codebase, you're reading papers. You're evaluating the quality and novelty of a team's research output, their citation networks, and their intellectual property position. This requires deep technical knowledge, but of a different kind - you need to understand the frontier, not just the stack.

Second, team graph analysis replaces architecture review. Who has worked with whom before? What's the density of the team's collaboration history? Have these researchers shipped things together, or is this the first time they're in the same room? The best pre-product bets I've seen are on teams with a track record of collaborative breakthroughs, not just individual brilliance.

Third, market timing conviction replaces traction metrics. Instead of "show me your MRR growth," it's "why will this technology be ready in 3 years and not 10?" This is where having a SWE background is still incredibly useful - you can sense whether a technical timeline is realistic or delusional in a way that someone from pure finance can't.

Here's the part I keep coming back to: this shift makes me uncomfortable. And I think it should make every technically-minded VC uncomfortable.

Because the thing that made us valuable - our ability to cut through hype with hard technical analysis - is being sidelined at the exact moment when the biggest bets are being placed. The deals that will define this era of venture (think about it: one-third of the $560B invested in AI has gone to just five companies) are increasingly pre-product, pre-revenue, and sometimes pre-prototype.

Does that mean technical DD is dying? Absolutely not. For the vast majority of deals - the Series A SaaS company, the B2B platform raising a Series B, the growth-stage fintech - the old playbook is more relevant than ever. The bar is actually rising: investors are now evaluating cloud unit economics, integration readiness, and organizational health in ways that would have seemed overkill five years ago.

But for the deals at the very top of the market - the ones that keep showing up in your Twitter feed - the game has changed.

If you're a SWE considering the jump to VC, or if you're already here like me, here's what I'd say: your technical skills are your foundation, but they're not your ceiling. The best technical VCs I know have evolved from "I can read your code" to "I can read your field." They understand not just how systems are built, but why certain technical bets are worth making before anything is built at all.

That's a harder skill to develop. It requires reading papers, attending conferences, building relationships with researchers, and developing taste for which problems are on the verge of being solvable. It's less satisfying than a clean code review. There's no pull request to comment on, no architecture diagram to red-line.

But it might be the most important skill a technically-minded VC can develop in 2026.

Because the next billion-dollar check won't be written after a code review.

It'll be written after a conversation.

- SWEdonym

Most outlets tell you what to think. The Flyover just tells you what happened. Free, fast, fact-focused news across politics, business, sports, and more. Join over 2.3 million readers — no paywall, no agenda.

Join Free Today