Nvidia posts record quarter, beating forecasts as AI-driven capex surge continues.

Nvidia’s fourth‑quarter haul of $68.1 billion—up 73 % year‑over‑year—was driven almost entirely by its data‑center segment, which generated $62 billion, according to the company’s earnings release. The breakdown shows $51 billion in GPU compute sales and $11 billion in networking products such as NVLink, underscoring how the firm has turned its graphics pedigree into a full‑stack AI platform (Nvidia News). The surge pushed full‑year fiscal‑2026 revenue to $215.9 billion, a 65 % jump from the prior year, and lifted GAAP gross margins to 75 % for the quarter, a slight edge over the 71 % annual average (Nvidia News).

CEO Jensen Huang framed the results as evidence that “the agentic AI inflection point has arrived,” noting that enterprise customers are racing to build “AI factories” capable of handling the exponential growth in token demand (Nvidia News). He highlighted the role of the new Grace‑Blackwell architecture, which pairs NVLink with next‑generation inference engines to cut token‑processing costs by an order of magnitude, and hinted that the upcoming Vera‑Rubin platform will extend that advantage (Nvidia News). Huang’s comments echo the TechCrunch interview in which he said the world’s token consumption has become “completely exponential,” pushing even six‑year‑old cloud GPUs to full utilization and driving price increases (TechCrunch).

Analysts had expected a slowdown after the 2023 AI boom, but Nvidia’s performance outpaced consensus forecasts across the board. Reuters reported that while the results beat estimates, investors are still pressing the company for larger cash returns, a sentiment reflected in the firm’s $41.1 billion share‑repurchase program for fiscal 2026 (Reuters). Bloomberg noted that the stock’s rally may temper as the market digests the magnitude of Nvidia’s growth and the sustainability of its capex‑driven demand cycle (Bloomberg). Nevertheless, the company’s non‑GAAP earnings per diluted share rose to $1.62 for the quarter and $4.77 for the year, reinforcing its profitability despite the massive scale of its operations (Nvidia News).

The data‑center boom also highlights a geopolitical nuance: Nvidia reported zero revenue from chip exports to China, even after the U.S. lifted some restrictions. Small shipments of H200 products to Chinese customers have been approved but have not yet generated sales, leaving the China market effectively untapped for now (TechCrunch). This absence of Chinese revenue underscores the firm’s reliance on Western cloud providers and enterprise buyers, a dynamic that could shift if export policies evolve.

Looking ahead, Nvidia forecast first‑quarter sales above analysts’ estimates, signaling confidence that the AI‑driven capex wave will continue to accelerate (Reuters). The company’s strategy of bundling high‑performance GPUs with proprietary interconnects and software stacks appears to be paying off, positioning it as the de‑facto supplier for the next generation of large‑language models and generative AI services. If the “agentic AI” narrative holds, Nvidia’s revenue mix may tilt even further toward AI‑specific workloads, cementing its role as the backbone of the emerging AI industrial revolution.