Nvidia: Huang Says Chip Sales Will Double Next Year, Testing the $673B Forecast

3-Line Briefing

• Nvidia CEO Jensen Huang said Thursday he expects the company to sell twice as many chips next year as it does this year, a forecast delivered at a UK summit with King Charles III alongside representatives from Google DeepMind, OpenAI and Anthropic.
• The comment builds on Nvidia's existing guidance of 70% revenue growth to roughly $673 billion for the fiscal year ending January 2028.
• Nvidia does not disclose total chip unit volumes; the last hard data point is 6 million Blackwell GPUs shipped over four quarters, as of last fall.

What "Twice as Many Chips" Actually Signals

Huang's framing matters because it is a unit claim, not a dollar claim. Nvidia's standing 70%-growth, $673 billion guidance for the fiscal year ending January 2028 is a revenue figure that blends data-center GPUs with CPUs, switch chips, optical-networking silicon, laptop chips, Jetson chips for robots and cars, and the chip inside Nintendo's Switch 2 console. Stacking a unit-doubling claim on top of that revenue target implies either steady average selling prices on data-center GPUs even as volume scales, or a shift in mix toward higher volumes of lower-priced parts — and Huang's remarks do not say which.

Huang attributed the demand to broad-based AI investment "in almost every single country" Nvidia operates in, tying the forecast to enterprise and sovereign AI buildout rather than any single customer. That is a demand-side justification, not a supply-side one: nothing in his comments addresses whether Nvidia's manufacturing and packaging capacity, the constraint that has historically bound Blackwell output, can actually clear twice the chip volume next year.

The Baseline Nvidia Has Already Guided To

The 6-million-Blackwell-GPU figure Huang cited last fall is the only concrete unit data point Nvidia has given for its flagship data-center line, and it covers four quarters that predate this week's remarks. Set against the fiscal 2028 target of about $673 billion in revenue on 70% growth, a chip-volume doubling next year would need to surface in upcoming quarterly reports as either accelerating Blackwell and Rubin shipment counts or broader unit growth across Nvidia's non-GPU lines, which Huang did not break out. Nvidia has not disclosed a total current chip-sales figure, so the doubling claim has no public baseline to be checked against yet.

Risks and the Next Checkpoint

• No unit disclosure: Nvidia doesn't report total chips sold, so Huang's doubling claim can't be independently verified until it shows up in revenue and margin trends.
• Mix ambiguity: the forecast doesn't specify how much of the doubling comes from data-center GPUs versus CPUs, networking, laptop or Jetson silicon.
• Safety overhang: Huang was in the U.K. specifically to discuss AI safety and told the summit that "when a product is not safe, we should hold it back and keep engineering it" — a signal that safety review, not demand alone, could pace future product releases.
• Loose timing: the remark was reported only as made "Thursday," with no specific date attached to when the next confirming data point arrives.

The clearest test of Huang's claim lands with Nvidia's next quarterly report, when investors can check whether data-center revenue and any disclosed shipment commentary are tracking toward the pace implied by both the chip-doubling remark and the standing 70%-growth, $673 billion fiscal 2028 target.

This article was independently written by OneDayTrading from public reporting. Read the original (CNBC)