Nvidia Corp.

$68.13B in Q4 revenue and a $1T-plus chip-sales forecast through 2027 frame Nvidia’s coverage as the build-out of an AI infrastructure platform, not just a GPU cycle.

Who they are

Nvidia is the central supplier of AI chips and rack-scale computing systems in this coverage, with Jensen Huang’s product and demand forecasts serving as a focal point for stories about data-center expansion, AI models and the financing required to deploy them. Its role extends beyond selling accelerators: the company appears as a partner and investor in infrastructure providers including Nebius, CoreWeave and SB Energy, while its Nemotron architecture is used in Salesforce’s Koa model.

The recent arc

Coverage accelerated from product-led AI-chip demand into a broader infrastructure-and-capacity story. Nvidia’s February results reported $68.13B in Q4 revenue, with Data Center revenue of $62.3B, followed in March by Huang’s expectation that flagship chips could generate more than $1T in sales through 2027. The late-2025 GTC narrative also emphasized Blackwell volumes and Huang’s rejection of the AI-bubble thesis.

The recent arc

The 2026Q3 peak has been driven by the next deployment wave around Vera Rubin and the customers financing it: AM Intelligence ordered 9,000 Vera Rubin systems for a planned 1GW project; Lambda raised short-dated debt to buy GPUs for lease to Microsoft; and reports said some customers were told system prices could rise more than 15% from early 2027. Nvidia also moved to deepen its ecosystem, including a planned $3.5B MediaTek investment, a $2B Nebius investment tied to more than 5GW of systems by 2030, and up to $105B of support for SB Energy’s Ohio campus that OpenAI is set to lease.

The tension

The coverage circles a contest between extraordinary AI-infrastructure demand and the durability of the capital supporting it. Nvidia’s sales outlook, Blackwell and Vera Rubin roadmap, and partnerships with cloud builders reinforce its position, but Peter Thiel’s fund sold Nvidia shares amid AI-bubble concerns and Lambda’s debt financing illustrates how much expansion depends on leveraged or externally funded buyers. Competitive pressure also remains visible in Nvidia’s NVLink Fusion strategy, which permits non-Nvidia CPUs or accelerators in its rack-scale systems, including amid attention to AMD, while potential US restrictions on shipments to Malaysia and Thailand underscore China-linked regulatory exposure.

Why it matters

If this trajectory holds, Nvidia’s influence will increasingly rest on its ability to organize the full AI-computing stack—chips, interconnects, memory, model architectures and financed data-center capacity—rather than on accelerator sales alone. That can make partners such as Microsoft, OpenAI, Nebius and MediaTek more consequential to Nvidia’s growth, but it also concentrates exposure to data-center execution, customer financing, pricing acceptance and export-control policy. Whether the investment cycle proves durable is the key uncertainty running through the coverage.

Related: AI · Jensen Huang · Vera Rubin · AMD · Microsoft · Nvidia reports Q4 revenue up 73% YoY to $68.13B, above $66.21B est., D