ClaimsJensen Huang

“In just two years, we increased the token generation rate from 22 million to 700 million, a 350-fold increase.”

Jensen Huang, Nvidia CEO, keynote 16 March 2026 Jensen Huang's GTC Speech (Full Text) (PANews)

Factual AI compute, hardware and data

In two years Nvidia raised token generation rate in a 1GW data center from 22 million to 700 million, a 350-fold increase.

Mostly true True in substance, wrong in a detail.

Factual: Statements about the present or the past, checked against the record: did they hold up?

The second judge would not resolve this claim; the verdict shown is the first judge's.

Said for an organisation they lead or work for; it counts on both records.

What would count as true or false

Test against NVIDIA's own keynote figures and any supporting data for token generation rate per 1 GW data center over the two-year period. The parts are the starting rate (22 million), the ending rate (700 million), the multiplier (350x) and the period (two years). Note that 700/22 is about 32x, not 350x, so the claim as stated is internally inconsistent. TRUE: the source figures support a starting and ending rate and a multiplier that are all consistent with the claim, within about 10%, over about two years. MOSTLY TRUE: the multiplier of about 350x and the 700 million endpoint are supported, but the starting figure was misstated (for example 2 million rather than 22 million). MIXED: the endpoints of 22 million and 700 million are supported (an increase of about 32x), but the stated 350-fold multiplier is wrong. MOSTLY FALSE: only one figure is supported, or the period is materially longer than two years. FALSE: none of the figures is supported by NVIDIA's data or independent measurements.

Set before the research began.

Evidence (4)

  1. E1 Transcript: NVIDIA GTC 2026 Keynote with Jensen Huang NVIDIA · 16 March 2026 Own record

    “But with this architecture, we're going to take our token generation speed token generation rate from 2 million to 700,000,350 times increase.”

  2. E2 Nvidia GTC Highlights – Nvidia Goal to Make All Data Ground Truth for AI NextBigFuture (Brian Wang) · 17 March 2026

    “token generation rate improved 350× in just two years (from ~22M to ~700M tokens per system).”

  3. E3 AI inflection point: As Nvidia's Jensen Huang outlines vision for agents and the AI factory, he forecasts big jump in revenue - SiliconANGLE SiliconANGLE · 16 March 2026

    “We’re going to take our token generation rate from 2 million per second to 700 million,” Huang said.”

  4. E4 GTC 2026 roadmap: Rubin Ultra, LP35, Feynman, LP40, Rosa longvideonotes.com · 1 October 2026

    “In a 1 GW factory, two years of architecture change takes output from 2 million to 700 million tokens per second, a 350x jump.”

Reasoning

NVIDIA's keynote transcript gives the starting token generation rate as 2 million and a 350x increase, though the endpoint text is garbled ('700,000,350 times increase') (E1). SiliconANGLE (E3) reports Huang giving the rate as 2 million per second to 700 million, and longvideonotes (E4) gives 2 million to 700 million tokens per second, a 350x jump, in a 1 GW factory over two years. These sources support the 700 million endpoint, the 350x multiplier and the two-year period. The 22 million starting figure appears only in E2 and conflicts with the 2 million in E1, E3 and E4, which is the figure consistent with 350x. E2 repeats 22M, but 700M divided by 22M is about 32x, not 350x, so E2 does not support the stated multiplier. Under the criteria, a misstated starting figure with the multiplier and endpoint supported is mostly true.

Cite this claim

Jensen Huang, “In two years Nvidia raised token generation rate in a 1GW data center from 22 million to 700 million, a 350-fold increase.” (said 16 March 2026). Verdict: mostly true. Reliability API, claim c_b670f1391e63, method 0.1.0+03d23c0e, dataset release v1.1.2. https://epistacklabs.com/claims/c_b670f1391e63

This claim is in dataset release v1.1.2, with the same verdict.

Reply to this claim

For Jensen Huang, or someone speaking for them: a correction or counter-evidence. Replies are read before anything is shown; an approved reply shows your name, role, date, statement and link. Your email address is never shown and is deleted 30 days after the decision (privacy).

More factual claims by Jensen Huang

  1. Mostly true Factual 4 September 2026
    “we invented a technology called a GPU, and basically a GPU does parallel processing tasks”

    NVIDIA invented the GPU, which does parallel processing tasks such as fluid dynamics, particle physics, quantum chemistry, graphics and AI.

  2. Resolved false Factual 1 June 2026
    “Today, NVIDIA’s token cost is the lowest in the world, not by 10%, by X factors, orders of magnitude, all because we did extreme co-design”

    NVIDIA's token cost is the lowest in the world, lower than others by large factors and orders of magnitude, due to extreme co-design.

  3. Mostly true Factual 1 June 2026
    “It’s the first one to be PCI Express Gen 6.”

    Vera is the first CPU to support PCI Express Gen 6.

  4. Mixed Factual 17 March 2026
    “We have demonstrated that NVFP4 can achieve inference without loss of precision, while delivering significant performance and energy efficiency improvements, and it is equally suitable for training.”

    NVIDIA has demonstrated that NVFP4 can achieve inference without loss of precision, with significant performance and energy-efficiency gains, and is equally suitable for training.

  5. Mixed Factual 17 March 2026
    “Over the past two years, the computational demands for inference have increased approximately 10,000 times, while usage has increased approximately 100 times.”

    Over the past two years, computational demand for inference rose about 10,000 times while usage rose about 100 times.

All of them

Verdicts are made automatically; how verdicts are made.