Kleiner Perkins Bets on Physical AI as the Source of the Next Trillion-Dollar Companies

Kleiner Perkins partner Ilya Fushman told Ed Ludlow on Bloomberg Tech that AI is still in the early stages of reshaping the economy. In his estimation, the next trillion-dollar companies could emerge in physical AI: self-driving cars and robotics. He has also become noticeably more optimistic about the IPO and M&A markets. For a venture fund, that is an important signal: bets are being placed across the entire stack, and exits appear to be getting more realistic.

An important caveat: only a teaser for the interview has been published. It contains no figures on company valuations, fund size or IPO timing. Below is what can be stated based on what was said, along with our reading of the context.

Four Layers of One Bet

Fushman lists several layers of the market at once. The first is frontier models, with Anthropic named as an example. The second is consumer agents, meaning AI that acts on a user's behalf instead of just answering questions. The third is autonomous cars, and the headline of the piece puts Waymo next to Anthropic. The fourth is robotics.

The logic is fairly clear. Models supply the intelligence, agents bring it to mass-market users, and autonomous vehicles and robots carry it into the physical world. The last layer is the most capital-intensive and the slowest, but Fushman names it as the likely source of the next trillion-dollar companies. The bar is high: only a handful of the largest tech corporations belong to that club so far, and all of them grew out of software, the cloud or chips.

Private AI Companies and the Exit Window

A separate point concerns the rapid growth of the leading private AI businesses. For a fund, the arithmetic is direct: the higher the valuations and revenue of portfolio companies, the more pressing the question of how and when investors will get their money back. Fushman says he sees reasons for optimism on both IPOs and M&A.

These are two different channels. IPOs suit large companies with predictable revenue that are ready to live under public reporting. M&A also works for mid-sized players: a startup with a strong team or technology can be acquired by a platform that needs talent, data or compute. If both channels are open, the venture cycle closes, and funds are more willing to take part in large late-stage rounds.

Key Takeaways from the Interview

• Market stage: AI is at an early phase of transforming the economy.
• Scope: frontier models, consumer agents, autonomous cars, robotics.
• Main emphasis: physical AI as a possible source of trillion-dollar companies.
• Momentum: rapid growth of the leading private AI companies.
• Exits: optimism on IPOs and M&A, with both channels in focus.
• Format: Bloomberg Tech interview, host Ed Ludlow, October 9, 2026.

What to Watch Next

Statements from a fund partner remain the view of an interested party: Kleiner Perkins makes money when valuations rise. So it is more useful to track verifiable signals than rhetoric.

First: the unit economics of a single ride or a single robot. For now, physical AI runs into the cost of hardware, regulation and the pace of scaling, so a trillion-dollar valuation remains a hypothesis. Second: actual IPO filings and large M&A deals among AI companies. These will show whether the fund's optimism matches the market's willingness to pay. Third: where capital is flowing within the stack. If rounds in robotics and autonomous driving start to catch up with model rounds, Fushman's forecast will gain support.

For founders, the practical takeaway is simple: funds on the level of Kleiner Perkins are ready to look beyond models and applications, but they want a clear path to scale. For readers following the industry, the main point is that the center of gravity in AI investment is gradually shifting from chat windows to the physical world.