North American startup funding shattered records in the first half of 2026, driven almost entirely by AI. In Europe, deal counts hit a six-year low while AI absorbed 60 percent of the money. Two headlines, same six months, opposite directions. And sitting in the middle of that mess is biotech, which did something almost nobody talks about because it doesn’t make a good chart: it held steady.
I review AI tools for a living. My inbox is a landfill of launch emails, and I’ve developed a decent nose for the difference between a product and a pitch deck with a UI attached. So when I look at funding data, I’m not reading it as an investor. I’m reading it as someone who has to test whatever comes out the other end. And from that seat, “held steady” is the most informative number in the whole set.
Records and record lows, at the same time
Set the headlines next to each other and the shape becomes clear. Crunchbase News reports North American funding at record highs on the strength of AI. Tech Times reports European deal counts at a six-year low, with AI taking 60 percent of the capital that did move. Emerging markets VC rebounded, per alternativeswatch.com, on a China tech surge. Meanwhile BioPharma Dive describes a widening biotech funding gap even as overall VC recovers.
Fewer deals, bigger checks, concentrated in one sector. That’s not a boom in the way people usually mean it. It’s a redistribution. The total goes up because a handful of rounds got enormous, not because more founders got funded. If you’re a European founder building anything that isn’t a model or a wrapper around one, the record-breaking headline describes a party you weren’t invited to.
Why “steady” is the tell
Biotech didn’t spike. It also didn’t collapse. In a market where capital is visibly stampeding toward one category, flat means the money staying put is there for reasons that have nothing to do with which sector is fashionable this quarter.
That’s the part worth sitting with. Biotech capital is stuck to a timeline it can’t argue with. Trials take years. Regulators take as long as they take. Nobody ships a therapeutic in a weekend sprint. The result is that biotech funding is roughly immune to vibes, which is why the number barely moved while everything around it did.
Compare that to what I see in AI tooling. A category gets hot, forty products appear in six weeks, thirty-five of them are the same product with different fonts, and by the time I’ve finished testing them a third have pivoted or gone quiet. That churn is a direct downstream effect of how fast the money moves. Cheap, abundant, fast capital produces a lot of software that exists mainly because funding existed.
What this means if you buy AI tools
Concentration at the funding layer shows up in your stack a few months later. A few practical things I’d keep in mind:
- Well-funded doesn’t mean durable. A large round in a crowded category can mean the company has runway to keep experimenting on you. Check whether the product solves something you’d still need if the hype cooled.
- Thin deal counts mean thin competition later. Europe’s six-year low in deal count is a preview of narrower tool options in a year or two. Fewer companies got started, so fewer alternatives will exist when you go shopping.
- Sector-specific tools deserve a different bar. Software built for biotech workflows sits next to customers whose timelines run in years. That tends to produce slower, less flashy products that actually stay put. In my testing, boring longevity beats a slick demo almost every time.
- Watch the widening gap, not the rebound. BioPharma Dive’s framing matters: overall VC recovered while the biotech gap widened. Aggregate recovery can hide sectors getting squeezed. The same logic applies to tool categories inside AI.
The unglamorous read
I don’t think biotech’s flat line is a triumph. A widening funding gap is a real problem for the people trying to raise into it, and steady is not the same as healthy. But it does say something useful about what happens when a sector can’t be rushed. The money that stays is patient because the science forces it to be.
AI doesn’t have that constraint, which is its advantage and its liability. You can ship in a week, so a lot of things get shipped in a week. Some of them are excellent. Most of them are what happens when capital arrives before a clear reason to build.
My honest take, from someone who tests this stuff: the sector everyone is ignoring is behaving more rationally than the one setting records. That’s not a prediction about returns. It’s an observation about what happens to product quality when funding stops being a filter. When you’re evaluating your next AI tool, ask whether it exists because someone needed it or because someone raised for it. That question is getting harder to answer, and these funding numbers are exactly why.
🕒 Published: