Analysis
Semiconductor companies have quietly become some of the most active investors in AI and robotics startups this year, participating in more than 60 financings of $100 million or more, in rounds collectively worth over $250 billion, according to Crunchbase News. Nvidia leads the category by a wide margin, having led or co-led at least 11 private company financings in 2026 alone, including a $5 billion round for Ilya Sutskever's Safe Superintelligence in July.
Nvidia's investing activity is inseparable from its position as chip supplier to nearly every company it backs -- it was one of eight lead investors in OpenAI's $122 billion round in March, and this week separately disclosed a $21 billion stake in SpaceX and up to $105 billion in financing backing for OpenAI's Ohio data center. Intel, historically a smaller player in corporate venture, has also stepped up participation, though at a fraction of Nvidia's scale and check size.
“Intel, historically a smaller player in corporate venture, has also stepped up participation, though at a fraction of Nvidia's scale and check size.”
Robotics as the next frontier
Sixteen semiconductor-backed rounds this year have cleared the $1 billion threshold, and robotics has become a growing share of that total alongside foundational model labs. Nvidia has been explicit that it sees robotics -- humanoid platforms, warehouse automation, and industrial manufacturing systems like Cincinnati's 1872 -- as the next large compute market after LLM training and inference plateau in growth rate, and its Jetson Thor computing platform for humanoid robots is central to that bet.
The structural pattern across all of this is the same one showing up in Nvidia's OpenAI and SpaceX relationships: chip suppliers taking equity or extending financing to their own largest customers, which locks in demand and captures upside but also means semiconductor company earnings increasingly reflect the health of a concentrated set of private companies rather than distributed enterprise chip demand. When a chip company's investment portfolio and its customer list start to overlap this heavily, a slowdown at any single portfolio company shows up twice -- once in lost chip orders, once in a markdown on the balance sheet.