NVIDIA's AI Dominance: Why Billionaire Halvorsen's Reduced Stake Still Points to Big Gains

Andreas Halvorsen, the Norwegian billionaire behind Viking Global Investors, is known for his contrarian bets and razor-sharp risk management. In late 2024, he made headlines by trimming his position in NVIDIA (NVDA) by 10%, reducing his stake to 2.03 million shares. Yet, despite this adjustment, Halvorsen’s residual holdings in the AI chip giant suggest a calculated strategy—one that still sees massive upside potential. Here’s why investors should pay attention.

The Halvorsen Play: Prudent Risk Management or a Bullish Signal?

Halvorsen’s partial exit from NVIDIA wasn’t a rejection of its long-term prospects but a tactical move to mitigate near-term risks. The reduction came amid heightened scrutiny of U.S. export controls on advanced chips to China and growing competition from AMD’s AI-focused offerings. However, the fund’s retention of 88% of its original NVIDIA stake underscores Halvorsen’s belief in the company’s unmatched dominance in AI infrastructure.

The key data point lies in NVIDIA’s revenue mix: 88% of its 2025 revenue comes from its data-center segment, which powers AI training and inference for tech giants like Alphabet, Microsoft, and Amazon. Even as Halvorsen boosted his Alphabet stake (a major NVIDIA customer), he left enough exposure to NVIDIA to capitalize on this $300 billion AI hardware market, projected to grow at 21% annually through 2030.

The Case for NVIDIA’s Continued Growth

1. Technological Supremacy: NVIDIA’s Hopper (H100) and upcoming Blackwell architectures offer 10x performance gains over older GPUs, with pricing premiums of 100-300% over competitors. This “moat” is critical in a sector where AI model complexity (measured in parameters) is doubling every 3.4 months.
2. AI Adoption Surge: 70% of enterprises are now investing in AI infrastructure, with 60% using NVIDIA’s DGX systems or cloud-based equivalents. Even as Alphabet invests in in-house chips, its $75 billion annual spend on AI infrastructure in 2025 remains heavily reliant on NVIDIA’s GPUs.
3. Valuation Check: Despite recent volatility, NVIDIA trades at 24x forward earnings, a discount to its 30x+ P/E during 2022’s AI hype cycle. With AI data-center revenue growing at 50% YoY, this valuation looks increasingly compelling.

The Contrarian Edge: Why Halvorsen’s Pivot to Alphabet Doesn’t Undermine NVIDIA

Halvorsen’s 186% stake increase in Alphabet isn’t a vote against NVIDIA but a multi-pronged bet on the AI ecosystem. Alphabet’s Google Cloud and Gemini AI models directly depend on NVIDIA’s GPUs, while its undervalued stock (17x forward earnings) offers a safety net. This “double-down” on AI’s hardware-software axis aligns with Halvorsen’s long-standing philosophy: own the entire value chain.

Risks and Reality Checks

• Regulatory Headwinds: U.S. export controls on AI chips to China could limit growth in key markets.
• Competition: AMD’s MI300X and Intel’s Ponte Vecchio aim to erode NVIDIA’s lead, though neither has matched its software ecosystem.
• AI Overvaluation: Halvorsen’s partial exit hints at concerns over near-term froth, but the long-term AI adoption curve remains intact.

The Bottom Line: NVIDIA’s Upside Remains Underappreciated

Halvorsen’s residual NVIDIA stake and his broader bets on AI-driven tech signal a decade-long opportunity, not a short-term trade. With $30 billion in annual data-center revenue and a 90%+ share of the premium AI GPU market, NVIDIA’s moat is as wide as ever.

Even if Halvorsen’s unnamed “10,000% return AI stock” materializes, NVIDIA’s role as the backbone of the AI economy ensures its place as a core holding for the foreseeable future. For investors, the path forward is clear: NVIDIA’s valuation, technological leadership, and secular growth story make it a buy at current levels, despite near-term noise.

In sum, Halvorsen’s nuanced moves—trimming exposure to risks but maintaining core stakes—highlight NVIDIA’s enduring value. As the AI revolution accelerates, this stock isn’t just a bet on silicon—it’s a bet on the future of computing itself.