Dulo Raises $130 Million Seed to Build AI for Hardware Design

Dulo, an artificial intelligence startup co-founded by Sebastian Thrun and Adrian Aoun, has raised $130 million in seed funding to develop what the company calls an Engineering Superintelligence, according to The Information. The round, confirmed by Aoun in a LinkedIn post, ranks among the largest early-stage financings for a startup pursuing AI-driven engineering.

What is Dulo actually building?

The company is developing foundation models designed specifically for engineering and hardware development. In effect, Dulo wants to extend the foundation-model approach that reshaped language and software into the physical engineering world. Its stated goal is AI that can assist with increasingly complex hardware design work.

Aoun put it directly in his announcement: "Earlier this year Sebastian Thrun and I co-founded Dulo on a mission to create an Engineering Superintelligence, building foundation models for hardware design." The company was founded earlier this year, according to his post.

Who put up the money?

The $130 million seed round drew a prominent roster of individual technology investors and venture capital firms. Backers include Marc Benioff, John Doerr, Eric Schmidt, Ram Shriram and Yuri Milner, alongside Construct Capital, Lowercarbon Capital, Bling Capital, BoxGroup, SV Angel, Haun Ventures and Scrible Ventures, among others. Aoun acknowledged the group in his post: "We appreciate the support from Marc Benioff, John Doerr, Eric Schmidt, Ram Shriram, Yuri Milner, Construct Capital, Lowercarbon Capital, Bling Capital, BoxGroup, SV Angel, Haun Ventures, Scribble Ventures, and many others."

That concentration of billionaire tech operators and established seed funds at the earliest stage signals unusually high conviction — or unusually high hype — before the company has shipped a product. The source does not disclose a valuation, revenue figures or a product timeline.

Why hardware engineering is a harder AI problem

The technical bet is not a straightforward extension of large language models. Hardware engineering imposes constraints that generative AI systems built for digital content do not face. According to the source, these include:

• Physical tolerances
• Materials properties
• Manufacturability
• Simulations
• Failure modes

Those factors create a different set of technical challenges than those confronting general-purpose language models. Dulo's success will depend on whether foundation-model techniques, which learn patterns from vast datasets, can master domains where data is sparse, expensive and unforgiving of error. The company has not yet detailed how it plans to train such models.

A shift in where AI capital is going

Dulo's approach also reflects growing investor interest in applying advanced AI models to specialized scientific and engineering disciplines. Rather than positioning AI as a productivity layer for knowledge workers — writing emails, summarizing documents — the company targets the underlying process of designing physical products and systems.

The founders bring complementary track records. Thrun has extensive experience developing and commercializing emerging technologies; Aoun is an entrepreneur and investor. Neither founder, according to the source, has disclosed specific engineering domains, customers or partnerships the company will pursue first.

What the money buys — and what it does not

A $130 million seed round, roughly ten times a typical top-decile seed deal, gives Dulo substantial runway to build foundation-model technology without immediate revenue pressure. It also raises the bar: investors at this scale will expect progress toward AI capable of transforming how engineers design hardware, not just research demos.

For now, the company has capital, pedigree and a thesis. It has not yet shown a model, a customer or a metric — and until it does, the Engineering Superintelligence claim remains exactly that. The next test is whether Dulo can translate one of AI's best-funded seed rounds into hardware designs that engineers actually use.