AI Unicorns Multiply as Town, Daytona Hit $1B and M&A Heats Up

A week of instant unicorns

Two young AI companies reached $1 billion valuations within a day of each other this month. Their timing shows how fast private capital is moving in 2026, and it also shows where that money is having trouble getting back out.

On October 19, Town, a San Francisco AI assistant startup, announced a $108 million Series B co-led by Forerunner Ventures and Menlo Ventures. The round valued Town at $1 billion only four months after it launched publicly.11 Andreessen Horowitz, Serena Ventures and First Round Capital also invested. The company has now raised at least $160 million, including a $55 million Series A in June.1113 Town says it reached millions of dollars in annual recurring revenue within weeks of launch. Its product is a set of assistants called "Townies" that learn about a user's colleagues as well as the user.1113 SiliconANGLE was cautious about the price. It noted that the assistants still have to show they last before the valuation looks justified.13

One day later, Daytona said it had raised a $60 million Series B at a $1 billion valuation.12 Daytona builds what it calls an "agentic cloud," meaning sandboxed computing environments where AI agents can install software, run code and fail safely. Fortune and TechCrunch both reported that FirstMark led the round.15 The Recursive counted Daytona as Croatia's third unicorn, after Infobip and Rimac Automobili.15 Coverage of Daytona's total funding does not agree. One outlet reported $91 million and another about €85 million. CEO Ivan Burazin put the team at 35 people, which is very small for a company valued at $1 billion.15

Speed is now the pitch

Town and Daytona are not outliers. CB Insights counts 19 companies founded in 2026 that have already become unicorns, together worth $60.5 billion. For comparison, AI companies that reached $1 billion this year took a median of 3.4 years to get there.19 Several in this group were valued on their founders' records rather than their products. Yann LeCun's AMI Labs raised $1 billion at a $4.5 billion valuation, and its chief executive has said commercialization is years away. Igor Babuschkin's River AI raised $1.1 billion at $5 billion.19 Core Automation, founded by former OpenAI staff, reportedly went from a $1 billion valuation to a $4 billion target within weeks, with no product or revenue. Mirendil, founded by former Anthropic staff, is reportedly in talks at $5 billion three months after first reaching $1 billion.19

The quarterly data points the same way. CB Insights reports that 42 AI companies reached billion-dollar valuations in Q3, the most in 18 months, even though AI deal count fell to an 18-month low.22 Across all sectors, 68 new unicorns appeared in the quarter, bringing the global total to 1,471.23 In the US, 24 new AI unicorns were created, compared with 11 in Asia and 7 in Europe.23 One analysis found that the 195 unicorns minted in the first half of 2026 matched the whole of 2025. It also noted that many reached $1 billion through rounds that doubled their valuations in months rather than through revenue growth.27

This is a market paying for founders' track records and their position in a category. Town, which has early revenue, is the more defensible case. The research labs valued at billions before shipping anything are the cases where pricing has moved furthest from fundamentals.

Fewer, bigger checks

The headline totals for the quarter differ somewhat by tracker, but they agree on the shape of the market. Crunchbase put global Q3 venture funding at $159 billion, down 25% from Q2. AI took $102 billion of it, or 64%.25 KPMG's Venture Pulse counted $170.9 billion across 8,341 deals and $728.3 billion for the first nine months of the year.28 CB Insights put AI funding at $103.3 billion, down 33% from Q2's $153.1 billion.22

The trackers also agree on why the quarter fell. Neither OpenAI nor Anthropic raised a mega-round in Q3, after raising $110 billion and $65 billion earlier in the year.2425 Leaving those two out, money was still flowing heavily. A record 27 companies raised rounds of $1 billion or more, and together they took about a third of the quarter's capital.2529 Databricks raised $5 billion at a $190 billion valuation. Crusoe raised $3.9 billion at $30.9 billion. Cognition raised more than $2 billion at $48 billion, and Mistral raised about $3.5 billion.25 KPMG called the market "quite bifurcated": most of the capital goes to a small number of top-end AI deals.28 One venture analyst estimated that the average AI round is now about 4.6 times the size of a non-AI round.27

The concentration reaches past software. Biotech's largest Q3 round went to Chai Discovery, an AI drug-discovery company that raised $400 million at a $3.8 billion valuation.30

Acquisitions are where the money comes back

Record fundraising sits alongside a stalled exit market, and acquisitions are filling the gap. The PitchBook-NVCA Venture Monitor counts 992 active US unicorns worth a combined $5.7 trillion, with no clear route to liquidity. Only 18 companies went public in Q3, most of them in healthcare.21 Some older unicorns have sold at steep losses. Airtable sold for $1.3 billion and Miro for $1.4 billion, both nearly 90% below their previous valuations.21

Young AI companies are getting very different terms. On September 28, AMD agreed to buy World Labs, Fei-Fei Li's two-year-old world-models startup, in an all-stock deal worth $8.2 billion.13 Li will become AMD's executive vice president and chief scientist, reporting to CEO Lisa Su.3 Ars Technica described the deal as AMD moving to compete with Nvidia's Cosmos world-model work in robotics simulation.4 World Labs launched in 2024 with $230 million in funding, which puts the size of the return in context.7

AMD's purchase fits a wider pattern of chipmakers and cloud providers buying AI software. Nvidia agreed to buy Hugging Face for $12.93 billion in September, the quarter's largest M&A deal, and Stripe bought OpenRouter for a reported $7.5 billion.24 Amazon is reportedly in talks to buy Decart, which makes software that lets AI models run on different makers' chips, for about $7 billion. Decart was valued at nearly $4 billion in May.9 Anthropic explored a deal for Decart at around $6 billion and dropped it after due diligence.5 Calcalist noted that Amazon gave Decart early access to its Trainium3 chip, so a purchase would help Amazon's effort to compete with Nvidia. It also pointed out that Decart's revenue is far smaller than the price being discussed.2 Tech Startups estimates nearly $200 billion in AI startup acquisitions, investments and licensing deals over 18 months, led by SpaceX's $60 billion agreement for Cursor maker Anysphere.6

The pattern suggests that chip and cloud companies now treat model and tooling startups as strategic assets. They are paying for insight into future workloads and for help reducing customers' dependence on Nvidia, not for current revenue. That makes them the most reliable exit route for AI startups this year. For AI companies that are not useful in the chip competition, and for most unicorns from 2021, the outcome looks closer to Airtable's.

OpenAI and Anthropic set the ceiling

The largest private AI companies are now valued at levels comparable to big public tech firms. OpenAI is reportedly close to raising up to $30 billion. CNBC put the pre-money valuation at $1.25 trillion, below the more than $1.4 trillion that Reuters reported a few days earlier.16 The reason for the difference is unclear. OpenAI's last confirmed valuation is still the $852 billion it set on March 31, and the new round had not closed as of October 22.20 The round is being described as a bridge to a possible 2027 IPO. Reuters reported that OpenAI sped up the process ahead of Anthropic's expected public listing in mid-November.16 CB Insights' rankings place Anthropic first among private companies at $1.2 trillion, ahead of OpenAI.23

These two companies affect the rest of the market. Their rounds largely determine whether a quarter looks strong or weak, and a successful Anthropic IPO would test whether public investors will accept private-market prices for AI.

The takeaway

The data does not show an AI market slowing down. It shows money concentrating in fewer companies. New unicorns are being created faster than at any time since 2022, AI acquisitions are closing at high prices, and the leading labs are preparing to go public.2224 However, the number of companies valued at $1 billion or more is growing much faster than the exits available to them, which one analyst called a growing backlog.27 Founders like those at Town and Daytona can raise at high valuations now. They will still need a buyer such as a chipmaker or cloud provider, or an IPO market that values them just as highly.