Nicolas Sauvage, founder of TDK Ventures, is investing in the unglamorous infrastructure that makes AI systems work at scale — inference chips, solid-state grid transformers, sodium-ion batteries, and robots that do one thing reliably.
Sauvage founded the corporate venture arm of Japanese electronics giant TDK in 2019. The unit now manages $500 million across four funds. His investment thesis is simple: identify the bottleneck four years out, then find the founders already working on it.
The portfolio
The highest-profile bet so far is Groq, the AI chip startup valued at $6.9 billion during its most recent funding round in fall 2025. Sauvage wrote a check into Groq in 2020, well before the generative AI boom made infrastructure bets obvious. Groq focuses on inference — the computational work that happens every time a model responds to a query. Founder Jonathan Ross, one of the engineers who built Google's Tensor Processing Units, designed the chip by building the compiler first, stripping the architecture down until "you can't remove one part and have it still work," Sauvage said.
Other portfolio companies include:
- Agility Robotics, which builds robots for the single, mundane task of moving things from one place to another in warehouses facing workforce shortages.
- ANYbotics, a Swiss company that builds ruggedized robots for environments too hazardous for human workers.
What's next
Sauvage is watching three areas closely:
Physical AI. Not all of robotics, but robots with a highly specific job to be done. The through-line is clarity of purpose — the robots he's betting on don't try to do everything; they do one hard thing reliably.
The compute stack shift. GPUs dominated training. Inference chips like Groq's are reshaping what happens when a model responds. Now Sauvage argues CPUs are due for a renaissance. They're not the most powerful or fastest chips, but they're the most flexible and best suited to the branching, decision-making logic of orchestration — managing the choreography when an AI agent delegates a task, checks progress, and loops back across dozens of steps.
China's hardware iteration speed. A recent report from venture firm Eclipse documented what Sauvage describes as "vibe manufacturing" — the rapid, AI-assisted iteration of physical hardware prototyping, mirroring what vibe coding did for software. Chinese manufacturers are compressing the design-build-test cycle for physical products in ways Western supply chains aren't yet equipped to match.
The unsolved problem
Sauvage says one remaining unsolved problem is dexterity. Models are improving fast enough that physical AI