China is doing to the USA what the USA did to the British. In hardware development, China already has the higher intercept and the steeper slope. It’s ahead and pulling further ahead. And it’s running a 24/7 shift. America won’t close that gap engineering at human speed. But I’m optimistic it doesn’t have to. I talked about why on Lumafield's Go/No-Go podcast. For years, the West dismissed China’s manufacturing advantage as cheap goods: “Sure, they build faster and cheaper, but American cars, robotics and technology are better.” That argument is getting harder to make. In robotics, drones, EVs and autonomy, Chinese products are now often the best in the world. What concerns me most isn’t where China is today. It’s the slope. We’ve seen this before. Britain pioneered modern manufacturing. America took those innovations, scaled them and moved faster. It happened again with AI. Much of the foundational work came from DeepMind in the U.K. but America moved faster to commercialize it. That’s the lesson, and it cuts both ways. The advantage goes to whoever scales innovation fastest. Right now, America leads in AI, and AI is how we scale hardware innovation at a speed no 24/7 shift can match. You can’t out-hour a 24/7 shift. But AI gives the West a real chance to change the equation entirely. Look at the gap between software teams using AI-native development and those still writing nearly every line by hand. It’s already enormous. The same thing is about to happen in hardware, and it’s an incredibly exciting time to be building. Agents working alongside engineers in a continuous loop to design, verify and deliver complex systems faster. That’s how America changes the slope. Full podcast linked below.

Flow on the Nasdaq - yes my Asian mum (and sister) were both very proud.

On Bloomberg TV, Roelof Botha and I talked about Flow's $50M Series B and why hardware design cycles are starting to collapse. In software, design cycles went from weeks → hours in less than a year. The key was verification. Every commit gets tested automatically, so AI agents can write code, run the tests and loop until it passes. In hardware, verification is still brutally manual. Every change has to be traced against millions of requirements across mechanical, electrical, thermal & software - design reviews, test campaigns, spreadsheets. It's so expensive and painful that you do it once, at a gate, and hope nothing slipped through. Flow turns hardware verification into a loop that runs on every change: make a change, > check it against every requirement → see the impact → iterate. Our agents listen to changes in CAD, Git, simulation and docs, then run impact analysis, flag conflicts and catch requirement failures as they happen. Agentic Systems Engineering is where we’re starting. Long term, the agents will get good enough to do real engineering work themselves: designing, analyzing and iterating alongside engineers. Thanks to Carmen Arroyo at Bloomberg for breaking the story. Her exclusive is linked in thread.

Flow has raised a $50M Series B at a $750M valuation, co-led by Antonio Gracias (Valor) and Gavin Baker (Atreides) with Sequoia Capital, Roelof Botha (SpaceX, Block), and more. When I became a mechanical engineer, I wanted to invent. In reality, 90% of the day to day execution work (CAD, excel, sim) was digital manual labour. That’s about to flip. The engineer’s role will change more profoundly than at any point since the invention of CAD, with humans focusing on invention, architecture, tradeoffs and judgement calls, and agents taking the grunt work. Anduril, Joby, Stoke Space, Rivian & many more - Flow is now the default platform for requirements and verification for frontier hardware teams. But it’s not just next-gen anymore. Industry giants including General Motors PPU and Volkswagen & Rivian’s joint venture (RV Tech) are reshaping their core development practices with Flow’s AI. The revolution is just starting. Models are getting better quickly at CAD, analysis, simulation, and tool use. In the next year, AI will move from integrating the work to doing the work. One day, this shift won't just enable us to iterate faster, it will enable us to design a class of products that we could not have dreamt of before. We’re hiring.

This is insanely important and “solves CFD”. Hardware design will forever be different with fast/solved NS.