Ken Fisher doesn’t make quiet bets. The billionaire founder of Fisher Investments — one of the largest independent money managers in the world — has put $16.05 billion into Nvidia, making it the top holding in his AI-focused portfolio. That’s not a speculative flier. That’s a conviction call at scale. And yet Nvidia shares are down about 4.5% over the past six months while everything else in tech seemed to rip. So what does Fisher see that the market is missing?
The bull case, as laid out by Nvidia’s own management and backed by Fisher’s thesis, centers on a fundamental shift in how AI is used. The industry is moving from training large models to running them — deploying AI agents at enterprise scale. That transition, called inference, demands constant, ongoing chip compute rather than one-time training runs. It turns Nvidia’s data center business from a cyclical capex boom into something closer to a recurring utility. CEO Jensen Huang has pointed to $1 trillion in data center revenue as a realistic ceiling. Goldman Sachs estimates Nvidia will drive 21% of total S&P 500 earnings growth in 2026 alone.
Valuation is another angle. At roughly 20x to 22x forward earnings, Nvidia is now trading nearly in line with the broader S&P 500 — remarkable for a company growing earnings this fast. That compression happened as the market spent months worrying about geopolitics, export controls, and AI bubble fears. Meanwhile, Big Tech collectively announced over $600 billion in capital expenditures for 2026, most of it running through Nvidia’s hardware.
None of this is risk-free. Export restrictions on advanced chips remain a real constraint. Chinese competitors are closing the gap on specific workloads. And a two-decade-long history of tech narratives teaches humility about $1 trillion targets. But at current valuations, with a verified demand cycle and one of the most credible bull-case sponsors in the business sitting on a $16 billion position, the skeptic’s burden is heavier than it looks.
