The disclosure came in a report filed with the U.S. Securities and Exchange Commission this past Friday (Aug. 14), but what looked on the surface like a routine portfolio diversification exercise concealed something far more ambitious: the quiet transformation of Nvidia, for years the unassailable symbol of AI chip design, into an equity participant in its own most important customers. Jensen Huang, the company's CEO, did not need to ask questions or negotiate special conditions — simply holding shares of Intel and SpaceX already placed the Santa Clara giant inside the very cycle of bets that sustained the computing ecosystem of its most ambitious founders, Elon Musk and Intel's Lip-Bu Tan.
The SpaceX position, valued at approximately $21 billion at the end of the second quarter, arrived by indirect route. In January 2026, Nvidia contributed $10 billion to xAI's Series E funding round, the artificial intelligence venture owned by Elon Musk. Two months later, in February, SpaceX completed its merger with xAI in a deal that valued the combined entity at $1.25 trillion. All of xAI's equity converted into SpaceX shares, and Nvidia's xAI stake became approximately 122.8 million Class A shares of the aerospace company. When SpaceX debuted on the Nasdaq on June 12, 2026 — raising $85.7 billion in the largest IPO ever recorded — Nvidia's shares were worth about $21 billion, or roughly 33% of Nvidia's total disclosed portfolio. The SPCX share price has been volatile, however: it closed at $140 last Friday, reducing Nvidia's position to approximately $17.2 billion.
The other half of the bet — $30 billion in Intel — has a different origin. In September 2025, Huang and Lip-Bu Tan, the newly appointed Intel CEO, announced an agreement that sent shockwaves through semiconductor markets: Nvidia would buy $5 billion in newly issued Intel shares at $23.28 per share, totaling 214.776 million new shares. The deal required antitrust clearance, and the FTC approved it on December 18, 2025. On December 26, Intel closed a private placement four days after regulatory approval. Overnight, Nvidia became one of Intel's largest individual shareholders.
The financial return was spectacular. At the end of March 2026, the position was worth roughly $9.5 billion according to earlier filings. By June 30, with Intel trading at approximately $140 per share, the same 214.8 million shares were worth about $30 billion — an unrealized gain of roughly $25 billion in under seven months. Even after Intel's additional $20 billion offering, closed on August 12, Nvidia's position value retreated to about $22 billion, but it remained the company's single largest disclosed holding. Intel also announced in August that Nvidia and Intel would jointly develop chips for data centers and PCs, a partnership that would give Nvidia access to Intel's advanced packaging and fabrication technology while providing Intel with the most important customer endorsement in the industry.
What makes this combination of bets truly singular is that both companies — SpaceX and Intel — have committed to using Nvidia hardware exclusively as their chip supplier. On SpaceX's second-quarter earnings call, held August 4 and marking the company's first mandatory earnings call as a public company, Musk announced that SpaceX would build all of its AI infrastructure on Nvidia hardware. "We've decided to build exclusively on Nvidia, because we think the Vera Rubin architecture is the best architecture," Musk said. "We're exclusive to Nvidia." The announcement caused AMD shares to fall 6% on the same day, while Nvidia shares rose roughly 3-4%, signaling that investors understood the weight of the commitment.
The commercial scale of that commitment is enormous. SpaceX's AI segment generated $2.56 billion in the second quarter — a 312% sequential jump. Musk said the company expects to operate more than 2 gigawatts of compute capacity by the end of 2026, scaling toward approximately 10 gigawatts by the end of 2027. All that computing capacity is now committed to Nvidia hardware. The Colossus 1 data center in Memphis, Tennessee, houses more than 220,000 Nvidia GPUs and already generates third-party lease revenue: the Anthropic contract covers all capacity for approximately $1.25 billion per month, and a Google deal covers 110,000 Nvidia GPUs at $920 million monthly from October 2026 through June 2029. Musk also announced plans for Starmind, a proposed satellite network that would deploy Nvidia GPUs in orbit, extending the partnership beyond Earth's atmosphere.
For Nvidia shareholders, including index funds that track the S&P 500, the risk dynamic is clear: if Intel or SpaceX underperform, Nvidia takes the hit twice — as a chip supplier whose revenue shrinks and as an equity holder whose portfolio deflates. Huang has transformed Nvidia from a unilateral supplier into a financial stakeholder in its largest customers, doubling exposure to the AI infrastructure buildout without publicly announcing this structural change. Altogether, Nvidia has about $50 billion concentrated in two companies that depend on it for processing power and are, simultaneously, indirect stakeholders in the Nvidia ecosystem itself. The question that remains is whether Jensen Huang calculated this web of risk with precision or is betting that exponential growth in the AI ecosystem will render any oscillation transient.
Sources: CNBC, Fortune, TechTimes
✓ Independent sources cross-checked and verified before publishing
