Sebastian Thrun revealed Dulo, a small stealth robotics startup, during an August 18 keynote at San Francisco's Actuate conference. Dulo's first-party page says it is developing foundation models for hardware design with a team drawn from Waymo, Google Brain and Stanford AI, but it gives no product, funding, customer or launch details.
Sebastian Thrun disclosed a small stealth robotics startup called Dulo during an August 18 keynote at the Actuate robotics conference in San Francisco. Dulo's first-party Stanford-hosted page says the team is building foundation models for hardware design with an ambition to enable faster manufacturing.
What Dulo has disclosed
Business Insider reported that Thrun revealed the company at the end of his conference talk but declined to discuss it in detail, describing it as very small, under stealth and focused on robotics. The report does not identify funding, customers, a product launch date or a deployment.
Dulo's sparse official page supplies the only public product description: foundation models for hardware design and a team with leadership experience from Waymo, Google Brain and Stanford's Artificial Intelligence Laboratory. It does not explain the model architecture, training data, supported hardware categories or how generated designs would be validated for manufacturability and safety.
Thrun previously led Stanford's team in the 2005 DARPA Grand Challenge and went on to help establish Google's self-driving-car effort, which became Waymo. He also co-founded Google Brain, Google X and Udacity, and later led the flying-car startup Kitty Hawk. That background helps explain the attention around Dulo, but it is not evidence that the new venture has a finished system.
Why the direction matters
Applying foundation models to hardware design would push generative AI beyond software and content into workflows constrained by materials, geometry, cost, supply chains and physical testing. Useful evidence will therefore require more than model demos: practitioners should look for defined design inputs and outputs, benchmark tasks, fabrication results, engineering review gates and comparisons with conventional CAD or optimization tools.
For now, Dulo is an early company reveal with a stated technical direction. Its capabilities, commercial model and manufacturing impact remain unverified until the team publishes technical material or working results.
Key Points #
- 1Sebastian Thrun disclosed Dulo at the Actuate robotics conference on August 18, describing it as a very small stealth robotics company.
- 2Dulo's first-party page says it is building foundation models for hardware design and names experience from Waymo, Google Brain and Stanford AI.
- 3No public evidence yet establishes Dulo's model architecture, funding, customers, launch schedule or demonstrated manufacturing results.
Scoring Rationale #
A prominent autonomy researcher has disclosed a new robotics venture pursuing foundation models for hardware design. The direction is relevant, but the company remains stealthy and has not published technical results, funding or a product timeline.
Sources #
Primary source and supporting public references used for this report.
Practice with real Logistics & Shipping data
90 SQL & Python problems · 15 industry datasets
[High-Value Overnight OrdersEasy](/problems/sql/high-value-overnight-orders)
[Delivered International ShipmentsMedium](/problems/sql/delivered-international-shipments)
[On-Time Delivery Rate by CarrierHard](/problems/sql/on-time-delivery-rate-by-carrier)
250 free problems · No credit card
See all Logistics & Shipping problems