The Week’s 10 Biggest Funding Rounds: AI Drives Another Spree Of Megadeals

Inference is eating the cap table

Crunchbase’s roundup puts AI at the center of another week of large U.S. financings. Baseten’s Series F was the headline number: $1.5 billion, its fourth fundraise in 18 months, for systems software used to run AI application workloads.

Groq followed in the same lane, closing $650 million in new funding led by Infinitum and Disruptive. The company said the capital will go toward scaling its AI inference cloud technology and infrastructure. That round came a little over six months after a transaction in which Nvidia hired away Groq’s founder and key team members and licensed its technology.

The pattern is not subtle. Investors are not just underwriting models; they are underwriting the plumbing required to run them. Inference has become the cost line everyone can see. If model usage grows, someone has to absorb the compute bill. These rounds suggest venture capital still believes specialist infrastructure companies can capture enough of that spend before hyperscalers and chip vendors compress the margins.

Foundation-model money is still available — at a price

The same Crunchbase list also included General Intuition, which raised $320 million in Series A funding at a $2.3 billion valuation. Khosla Ventures led the financing, with participation from backers including Jeff Bezos and General Catalyst. The company is developing a foundational AI model based on gameplay.

Mirendil, described as a frontier lab building systems that excel at AI R&D, said it raised a $200 million seed round led by Andreessen Horowitz and Kleiner Perkins. Nvidia is also listed as a backer.

A $200 million seed round is not seed-stage finance in any traditional sense. It is institutional risk capital being deployed as if access to talent, compute and distribution windows must be bought upfront. That can work. It can also bury ordinary venture discipline under a very expensive burn rate.

Outside AI, there were still large checks. AppsFlyer reportedly secured more than $1 billion in Series E funding at a $2.7 billion post-money valuation, with reported backers including Unity, Meta, Moloco and Google. Ollin Biosciences raised $330 million in Series B funding for ophthalmic therapies. Peregrine Technologies secured $250 million in Series D financing at a $6.8 billion valuation. Quantifind closed $200 million in growth financing led by Summit Partners.

But AI took the oxygen. That is the useful signal.

The wider deal market is becoming top-heavy

The venture numbers sit inside a broader capital market shift. Briefs Finance, citing PwC, reported that global dealmaking is on pace to reach $4 trillion in 2026, which would make it the busiest M&A year since 2021. Deals valued at $5 billion and above reportedly account for nearly half of M&A activity so far this year, up from 39% a year earlier and 26% in 2024.

The same report framed AI as a common thread in large transactions. It cited SpaceX’s agreement to acquire Cursor for $60 billion, Salesforce’s move to buy Fin for $3.6 billion, and Qualcomm’s talks to buy Modular for around $4 billion. PwC’s Brian Levy described the market as increasingly K-shaped, with the largest players pulling away while mid-market dealmakers face geopolitics, price gaps, slower growth, higher borrowing costs and a backlog of private equity assets waiting to be sold.

For AI investors, the practical takeaway is blunt. Liquidity is concentrating around companies that either control strategic AI infrastructure or can be bought by balance sheets large enough to ignore ordinary valuation gravity. Everyone else is still living with interest rates, exit pressure and a cap table that needs a credible path to cash — not just another larger round.