$43 billion. That's how much Nvidia now holds in startup investments - a portfolio size that rivals the largest dedicated venture capital firms in the world, disclosed alongside another record quarterly revenue result on Wednesday.
The record revenue itself is almost routine at this point. Nvidia's GPUs handle the bulk of AI model training and inference (running models after they're trained) globally, and demand has not softened. But the investment disclosure reframes what Nvidia actually is. A $43 billion stake across AI startups means the company isn't only selling chips - it has direct financial interest in which AI companies win.
The complication: Nvidia simultaneously supplies the hardware these startups run on, invests in their equity, and competes with some of them through its own software and cloud platform products. Traditional semiconductor companies didn't face that conflict structure. There's no indication of wrongdoing, but the scale of the portfolio makes the overlap harder to dismiss as incidental.
On forward guidance, Nvidia signaled that revenue growth would slow in the coming quarter. That's deceleration, not contraction - the company's revenue base is now so large that matching prior growth rates requires adding billions more each quarter just to stay flat on a percentage basis. Still, analysts who have priced in sustained hyper-growth will read any guidance cut carefully.
For the companies building AI tools that people actually use day-to-day, Nvidia's position as the dominant hardware supplier gives it pricing power that flows through the entire stack. OpenAI, Anthropic, and the hundreds of smaller AI product companies all depend on Nvidia's chip availability and cost structure. The $43 billion investment portfolio means Nvidia now has a financial stake in those outcomes too - not just a vendor relationship.