Nvidia chief executive Jensen Huang told reporters on Thursday that the company will sell twice as many chips next year as it does in the current one, a volume target that implies a growth trajectory few semiconductor companies have ever sustained. The comment came at a U.K. summit attended by King Charles III and representatives from Google DeepMind, OpenAI and Anthropic, where Huang framed the projection as a consequence of near-universal demand for AI infrastructure rather than a supply-driven milestone.
The volume claim has no denominator
Nvidia does not disclose total chip shipments, so the doubling pledge cannot be verified against a baseline. The only concrete figure Huang has offered recently is the 6 million Blackwell GPUs shipped over four quarters, disclosed last autumn. That number covers only the data-center graphics processors that have become the company’s flagship product; it excludes central processors, switch and optical-networking chips, laptop silicon, Jetson modules for robotics and automotive, and the custom processor inside Nintendo’s Switch 2 console. Without a total unit count, “twice as many” is a direction, not a metric.
The revenue forecast that preceded it
The volume target follows a financial forecast issued earlier this month: 70 percent revenue growth for the fiscal year ending January 2028, reaching roughly $673 billion. Analyst models adjusted upward after that guidance. If the unit-volume doubling materializes at anything like current average selling prices, the revenue implication would exceed even that elevated consensus. But the source is silent on pricing assumptions, product mix shifts, or whether the volume increase includes lower-margin consumer and automotive parts that would dilute the revenue impact.
Safety rhetoric at a royal summit
Huang used the platform to reiterate a position on AI safety that has become a recurring talking point in recent weeks. “When a product is not safe, we should hold it back and keep engineering it,” he said. The statement aligns with the summit’s stated purpose but sits awkwardly alongside a sales forecast that assumes unconstrained deployment across “almost every single country.” The tension between gating unsafe products and doubling shipments in twelve months is not addressed in the company’s public rationale.
What to watch next
The next test is whether supply-chain partners, foundry capacity, advanced packaging, high-bandwidth memory, can support a unit ramp of that magnitude without margin erosion. Nvidia’s guidance has historically been directional rather than contractual, and the absence of a disclosed unit baseline makes accountability difficult. Investors should listen for any quantification of the current run-rate on the next earnings call; until then, the doubling claim is a narrative device, not a covenant.