Nicolas Sauvage, head of TDK Ventures, is focusing his investment strategy on the foundational, often overlooked components of artificial intelligence. Speaking at a recent StrictlyVC event in San Francisco, Sauvage noted that it often takes four years for the most promising technology bets to become obvious to the broader market.
Since founding the corporate venture arm of the Japanese electronics giant in 2019, Sauvage has managed $500 million across four funds. One of his most significant successes includes the AI chip startup Groq, which reached a $6.9 billion valuation during its most recent funding round last fall, according to TechCrunch.
Sauvage invested in Groq in 2020, well before the current generative AI boom. The startup, founded by former Google engineer Jonathan Ross, focuses on inference—the computational work required every time an AI model responds to a user. While the technology might have seemed niche at the time, Sauvage identified an asymmetric opportunity in the compounding demand for inference as AI agents become more complex.
Finding value in hardware and energy
Sauvage's investment philosophy centers on identifying technologies that address the physical and structural requirements of the AI era. His current portfolio includes developments in solid-state grid transformers and sodium-ion batteries designed specifically for data centers. He is also targeting alternative battery chemistries that reduce reliance on the geopolitically sensitive lithium and cobalt supply chains.
The creation of TDK Ventures itself was an unconventional move for the Japanese conglomerate. Sauvage, a French national who joined TDK in Silicon Valley following an acquisition, pitched the fund to TDK headquarters despite having no direct standing within the Japanese parent company.
“I’m not Japanese. I don’t speak Japanese; I don’t live in Tokyo,” Sauvage told TechCrunch, describing the unlikely nature of the venture's inception.
Sauvage eventually secured the mandate to build the fund around a single central question: identifying the next major advancements for TDK and the technologies that might eventually disrupt them.
