AI Unicorns Surge: 42 New Billion-Dollar Startups in Q3

A cooler quarter that still minted unicorns

The third quarter of 2026 produced an odd-looking set of venture numbers. Headline funding totals fell from the first half of the year, yet the pace at which AI startups crossed the billion-dollar valuation mark accelerated. CB Insights counted 42 companies reaching unicorn status in the quarter, the highest quarterly figure in 18 months. Its read was that investors are still backing new AI unicorns even while placing fewer bets overall 3.

That combination of fewer dollars in aggregate and more companies crowned at billion-dollar valuations is the key to understanding where AI venture capital stands. It does not look like a market in retreat. The most plausible reading is that the first-half totals were inflated by a handful of outlier deals.

The OpenAI and Anthropic effect

Crunchbase's North American data shows the distortion most clearly. Total startup funding in the region came to about $92 billion in Q3. That was down 35% from the second quarter but still 50% higher than a year earlier 1. AI companies took roughly $61 billion of that, about two-thirds of the total 1.

The sequential drop has a simple explanation. Crunchbase attributes it mostly to the absence of new megarounds for OpenAI and Anthropic. Their financings of $110 billion and $65 billion helped push the first and second quarters to record levels 2. Remove those two companies from the comparison and the quarter-over-quarter decline loses much of its drama. Crunchbase says the dip does not appear to reflect any broad weakening in the venture climate 2.

Deal counts support that view. Early- and late-stage round counts in North America stayed close to Q2 levels 1. If investors were genuinely pulling back, the number of rounds would be expected to shrink along with the dollars.

Big checks keep flowing

The global picture follows the same pattern. Worldwide venture funding reached $159 billion in Q3 2026, according to one tally, which also described a quarterly pullback. AI nonetheless drove a record 27 rounds of $1 billion or more 4. Fewer total dollars alongside a record number of billion-dollar checks only makes sense if capital is concentrating in a narrower set of perceived winners.

In North America, late-stage and growth-stage companies raised $66.45 billion. That is up about a third from a year earlier, though sharply down from the first two quarters 2. The largest rounds were:

• Databricks at $5 billion
• AI infrastructure company Crusoe at $3.9 billion
• The Boring Co. at $3 billion
• AI coding startup Cognition at $2 billion 2

More than a dozen startups raised late-stage or growth rounds of $1 billion or larger 2. Crunchbase also noted that a few exceptionally large early-stage rounds, including one for an open-source AI company, lifted that category's total 2.

Where the data diverges

The sources do not agree on everything. CB Insights describes investors making fewer bets overall 3. Crunchbase says North American deal counts held roughly steady 1. The gap likely comes from differences in geography, methodology and how each firm classifies AI deals, rather than a true contradiction. Still, it is a reminder that "funding is down" or "funding is up" depends heavily on which yardstick is used.

What both datasets share is a story of concentration. Money is clustering in more mature companies 1 and in very large rounds 4. Meanwhile, the number of newly minted unicorns is rising 3.

What it means

Two readings are possible. The optimistic one is that the AI market is broadening. Valuations once reserved for a few frontier labs are now reaching a wider group of infrastructure, data and application companies, from compute providers like Crusoe to coding tools like Cognition.

The more cautious reading is that investors are betting heavily on a smaller number of names and paying unicorn prices to secure allocations. That raises the stakes if growth disappoints. Crunchbase's framing that AI giants are eyeing the public markets 2 matters here. Exits through IPOs will eventually test whether these private valuations hold up.

The evidence leans toward the first interpretation, with caveats. A 35% sequential drop that disappears once two companies are excluded, steady deal counts and a record number of billion-dollar rounds do not describe a cooling market. They describe one that has become dependent on a few giant deals, which will make swings like this quarter's routine. The 42 new unicorns suggest investor appetite remains strong, even if not every new entrant justifies its price.