David Tepper's Bold Bet: Doubling Down on Nvidia While Dumping Intel in Latest 13F Filing

Billionaire investor David Tepper’s latest 13F filing just revealed a striking portfolio move that contradicts his own deep-value investing philosophy. Managing close to $6.5 billion at Appaloosa, Tepper announced on February 10 that he’s dramatically shifting his AI play: buying 55,001 shares of Nvidia while aggressively selling 60% of his Intel position (1.5 million shares). It’s a decision that has Wall Street scratching its head.

The Nvidia Paradox: Why Tepper is Buying an Expensive Stock

What’s puzzling about Tepper’s move is his renewed interest in Nvidia at valuations that seem anything but cheap. The chipmaker’s current price-to-sales ratio hovers around 30—well above historical norms and firmly in bubble territory by traditional metrics. For a value investor like Tepper, this appears counterintuitive.

Yet the logic is hard to ignore. Nvidia’s dominance in AI-accelerated data centers is unmatched. The Hopper H100 GPU and its successor, the Blackwell architecture, are generations ahead of competitors in raw computing speed. More importantly, Nvidia’s CUDA software ecosystem has created an almost impenetrable moat. Developers across Microsoft, Meta, Amazon, and Alphabet have invested heavily in building AI applications on CUDA, making switching costs prohibitively high.

By adding to his Nvidia position with an 8.8% increase from September 30, Tepper may be betting that this ecosystem advantage will persist despite bubble-like valuations. The question remains: is he buying Nvidia’s technical superiority or riding an AI hype cycle that could implode?

Intel’s Mysterious Sale: Abandoning a Deep-Value Opportunity

The Intel exit is where Tepper’s decision becomes truly perplexing. Dumping 60% of Appaloosa’s Intel stake in Q4 looks like exactly the kind of panic selling that contradicts his value-investing track record. On paper, Intel should be Tepper’s ideal turnaround play.

Yes, Intel’s execution has been disastrous. The company was late pivoting to enterprise GPUs, surrendered CPU market share to Advanced Micro Devices, and is burning cash building its foundry business from scratch. These are legitimate concerns. But the bull case is equally compelling.

Intel’s Gaudi AI chips offer a compelling price alternative just as Nvidia’s GPU scarcity begins to ease. With demand backlogs clearing naturally over the next few quarters, enterprises will have the luxury of choice—and Intel’s pricing will look increasingly attractive. Add President Trump’s push for domestic AI intellectual property protection and U.S. chip sovereignty, and Intel becomes a natural beneficiary of efforts to break Nvidia’s GPU monopoly.

There’s also the possibility of an Intel foundry spinoff, which could unlock value by separating profitable chip divisions from struggling new business segments. Meanwhile, Intel stock trades at only a 9% discount to book value—historically rock-bottom for a company of its size and profitability potential.

Tepper’s aggressive 60% exit may prove to be one of his most regrettable calls if Intel’s turnaround actually accelerates.

The Real Story: Betting on Monopoly Over Value

Tepper’s portfolio shift reveals something important about 2025’s investment climate. Even disciplined value investors are struggling to resist Nvidia’s gravitational pull. The company’s near-monopoly in mission-critical AI infrastructure is real, and the switching costs are genuine.

But that same monopoly power that attracted Tepper is also the reason to worry. History shows that companies leading “next big thing” innovations consistently get overestimated by investors in the early adoption phase. When those bubbles burst, market leaders suffer the most brutal crashes.

David Tepper’s 13F filing shows a man wrestling with this exact dilemma: his discipline telling him Intel at deep discounts is the smart play, but his market instincts screaming that Nvidia’s moat is too valuable to ignore. For now, the moat is winning. Whether that proves wise depends entirely on whether AI lives up to the hype.