One of Silicon Valley's most celebrated technologists has raised one of the largest seed rounds of the year for a startup that aims to do for engineering design what large language models have done for writing and software code.
Dulo, a San Francisco-based startup co-founded by Sebastian Thrun and Adrian Aoun, has raised $130 million in a seed round, according to a report by The Information on Thursday, October 29, citing a person familiar with the deal, and reported by Tech Startups. Investors include former Google chief executive Eric Schmidt, venture capitalist John Doerr, DST Global, Construct Capital and Lowercarbon Capital.
Founded in 2026, Dulo is building foundation models for engineering and hardware design.
The founders
Sebastian Thrun is one of the best-known figures in artificial intelligence and robotics. He led the Stanford team that won the 2005 DARPA Grand Challenge for autonomous vehicles, founded Google X and the company's self-driving car project, which became Waymo, and co-founded the online education company Udacity. He also founded Kitty Hawk, which worked on electric flying vehicles.
Adrian Aoun is a serial entrepreneur who founded Forward, a technology-driven healthcare startup, and previously worked at Google. The pairing brings together deep expertise in AI, robotics and building ambitious technology companies.
Foundation models for hardware
Foundation models are large AI models trained on vast amounts of data that can be adapted to many tasks. The best-known examples, such as OpenAI's GPT models and Anthropic's Claude, are trained mainly on text and code. Dulo aims to build models designed for engineering and hardware design, which involve geometry, physics, materials and manufacturing constraints.
Designing physical products, from mechanical parts and electronics to vehicles and industrial equipment, is a complex, time-consuming process involving computer-aided design (CAD) software, simulations and repeated testing. AI models that can generate, evaluate and optimise designs could significantly accelerate engineering work.
Few details disclosed
Dulo has not disclosed product specifications, customers, revenue, valuation or independent benchmarks, according to Tech Startups. The size of the seed round is extraordinary for a company at such an early stage, reflecting investor confidence in the founders and in the potential of AI for engineering.
A wave of mega seed rounds
Very large seed rounds have become more common in artificial intelligence, as investors compete to back prominent founders building foundation models, which require substantial computing resources to train. Such rounds would have been almost unthinkable a few years ago, when seed rounds typically ranged from a few hundred thousand to a few million dollars.
Critics argue that the size of these rounds reflects exuberance in AI investing, while supporters argue that building foundation models simply requires more capital from the outset.
The climate angle
The participation of Lowercarbon Capital, a climate-focused venture firm co-founded by Chris Sacca, suggests that investors see potential for AI-driven engineering to accelerate the development of climate technologies. Designing more efficient batteries, turbines, motors and industrial equipment could help reduce emissions, and faster design cycles could bring clean technologies to market sooner.
Competition
Dulo enters a field where established design software companies, including Autodesk, Dassault Systèmes, Siemens and PTC, are adding AI features to their products, and where several startups are building AI tools for engineering. Large AI developers are also expanding their models' capabilities into areas such as 3D design and simulation.
To succeed, Dulo will need to demonstrate that its models produce designs that engineers trust and that work in the physical world, where errors can be costly and dangerous.
Implications for India
India has one of the world's largest pools of engineering talent, and many global companies operate engineering design centres in India, working on everything from automobiles and aircraft to semiconductors and consumer electronics. Engineering research and development services is a large and growing industry for Indian companies such as Tata Elxsi, L&T Technology Services and others.
AI tools that automate parts of the design process could change the nature of this work, increasing productivity while requiring engineers to develop new skills. Indian engineering services firms are already investing in AI to stay competitive.
Why hardware design is hard for AI
Text and code are abundant online, providing vast training data for language models. Engineering design data is different: much of it is proprietary, held within companies' CAD systems and product databases. Designs must also obey the laws of physics, meet safety standards and be manufacturable at reasonable cost. Training AI models that understand these constraints requires specialised data, simulation tools and validation methods.
Dulo will need to find ways to access or generate high-quality training data, potentially through partnerships with manufacturers, synthetic data generated by physics simulations, or both.
Potential applications
AI for engineering design could have wide applications. It could help engineers generate initial designs from requirements, optimise parts for weight or strength, identify potential failures before physical testing and suggest manufacturing methods. In industries such as aerospace, automotive, consumer electronics and energy, faster design cycles could reduce development costs and accelerate innovation.
Generative design tools already exist in some CAD software, using algorithms to propose shapes that meet specified constraints. Foundation models could extend these capabilities by understanding natural-language instructions and broader engineering context.
Thrun's track record
Thrun's career has repeatedly involved betting on technologies years before they became mainstream, from self-driving cars to online education and flying vehicles. Not all of his ventures have succeeded commercially, but his work on autonomous vehicles laid foundations for an industry now operating robotaxis in several cities. Investors are betting that his vision for AI in engineering will prove similarly influential.
What comes next
Dulo will use its funding to build its foundation models, which will require significant computing resources and data. Its progress will be watched closely given the prominence of its founders and the size of its seed round. If successful, it could help define a new category of AI focused on the physical world, with implications for manufacturing, engineering and innovation across industries.
For engineers around the world, including India's large engineering services workforce, Dulo's progress could offer an early glimpse of how their work may change in the years ahead.
Because the company has not yet disclosed its product, early customers or performance benchmarks, investors and industry observers will be watching for the first concrete evidence of what Dulo's models can do and how they compare with existing design tools used by engineers today.
