Nvidia CEO Jensen Huang told reporters in Taipei that the oft-cited "up to $100 billion" figure was never a firm commitment but rather a non-binding letter of intent. He emphasized that Nvidia will invest "one step at a time" and continues to believe strongly in OpenAI’s technology. Following Huang’s remarks, Nvidia shares slid by roughly 1.8% in premarket trading, as investors reassessed the timing and scale of the potential capital deployment and its impact on near-term cash flow.

Over the past quarter Nvidia has deployed more than $30 billion in strategic stakes across the AI ecosystem. The company increased its position in CoreWeave with a $2 billion investment, boosting its stake to 11.5% in the data-center specialist. Weeks earlier, Nvidia struck a $4.9 billion deal to acquire Intel’s Habana Labs assets and invested $20 billion to purchase Groq’s IP and talent. These transactions reflect a deliberate strategy to secure GPU capacity, custom ASIC technology and specialized teams to support its full-stack AI infrastructure roadmap.

Despite Nvidia’s rally over the past three years, the stock currently trades only modestly above the S&P 500 on a forward price-to-earnings basis, underscoring investor belief in sustained AI demand. Analysts at Deutsche Bank project Nvidia’s server GPU revenue to grow at a 45% compound annual rate through 2027, driven by enterprise data-center deployments. With Nvidia’s latest 5-nanometer GPU architecture ramping production in Q2 2026 and order backlog exceeding $75 billion, the company remains positioned to capture the lion’s share of incremental AI infrastructure spending.