Opinion
The best time to build, the hardest time to raise
Execution got cheap, so we stopped underwriting it. What we buy now is the derivative: the pace of learning.
This is an extraordinary moment to start a company. Entire technology stacks are being rebuilt at once, across industries that hadn't changed in decades. Teams reach real products and customers faster than at any point in this industry's history.
There is also more capital in this market than there has ever been. Global venture funding hit a record $510b in the first half of 2026 (Crunchbase) - more in those six months than in all of 2025, and past the previous half-year record set at the top of the last cycle. Roughly 86% of US venture dollars went to AI.
At the same time, it has never been harder to raise money.
Since the fourth quarter of 2024 there have been 353 completed seed rounds involving Israeli companies, according to PitchBook. The median round grew from $5M to $8M while quarterly deal count fell from 77 in the first quarter of 2025 to 27 in the most recent quarters.
Seed companies are also less likely to reach a Series A, and slower when they actually do. Carta's data on US startups shows roughly 17% of the 2022 seed cohort raising a Series A within two years, against 25–30% for the 2018 cohort.
Concentration used to be a late-stage challenge about few mega-rounds and it has now moved to the early stage.
Consensus is expensive, and usually not worth it
The market has moved decisively to consensus. Entry valuations are high and there is always a bigger, more expensive round around the corner, sold on the promise of a generational company.
Most of the time that price is not buying a better company but it fits the pattern other investors recognize and want, which makes the next round much more likely to happen.
Once in a while it really is the jackpot. The trillion-dollar outcomes of the past few years are real enough that everyone now underwrites for the possibility of one, and the size of the prize makes that rational. The cost is that capital drains away from everything that doesn't look like it.
Our business used to be defined by the opposite instinct. In 2018, when a16z managed roughly $8b, Marc Andreessen described the firm as “trying to preserve the contrarianism kind of at the core of what we do, the strong, non consensus views”. This year, at about $90b AUM, the advice is the reverse: "Don't ever do diamonds in the rough, only do diamonds”, because if something has merit and is investible for venture, "there are plenty of smart, hungry VCs" already hunting for it. At $90b, the search stops being for what is mispriced and becomes a search for what is certainly going to be enormous so paying up is the rational move (especially when the alternative is missing, which you really can’t afford to).
For those of us not managing $90b, this leaves us with 2 questions: Can companies still survive outside the consensus? And is finding them still an edge?
The KPIs stopped meaning what they meant
Raising a Series A is much harder than it was eighteen months ago, and the metrics that would have earned one then don't earn one now, both on the traction side and the ambition side. The reason is that execution got cheap.
When agents do the work of a team and five people ship what used to take twenty engineers, hitting a milestone no longer tells you much about the people who hit it.
For twenty years, distance travelled was a usable proxy for the quality of a team, because getting anywhere at all was hard. Today, arriving there says much less about you, and standing out takes a lot more.
The ambition bar also rose, because every investment is now looked at through the lens of whether it could become enormous. Every pitch now promises a vast outcome with a TAM to match, which is why investors discount it accordingly.
Investors have shifted away from underwriting the pure metrics and started underwriting the “derivative”: instead of what ARR you reached, how fast you got there and how much you learned on the way. In a market where a competitor can rebuild your product in a quarter (especially if they just raised one of the new, larger seed rounds), a team's rate of learning is often the only thing that compounds, and the closest thing to a moat that's left. We bet that even if the next competitor raises 3x as much and the tech stack keeps on improving, our team figures out their next move before they do.
What this means for founders
Two things follow.
First, put your rate of learning into the story you tell, not only the metrics growth. What did you believe six months ago, what do you know now, and what did it cost you to find out? Founders who are actually moving fast find that a much easier conversation than a traction slide, which is so hard for investors to interpret or get excited about in 2026.
Second, know which path you are on and what your capital plan is. The upside is that capital has fragmented along with the companies, and funds are increasingly explicit about which kind they back. Some companies are on a fundraising path, where the objective is to reach the metrics and the legibility that make the next round happen (which, also, looks very different if they’re building a hard science venture vs an AI roll-up). Others are on a breakeven path, where the objective is to reach sustainability and stop chasing capital in the short term. Both can produce excellent outcomes but they are not the same story and they are not told to the same investors (who also don’t like to figure this out for you).
It has never been easier to build something real. It has never been harder to get it funded. Starting a company has always required a degree of irrationality. It just maybe requires more of it today.
Emmanuelle Lipski is a Partner at 10D.
