Open question: do synthetic hours count? NVIDIA generated the equivalent of 6,500 human hours in eleven, and no standard says what an hour is
NVIDIA's press release of 18 March 2025 says it generated 780,000 synthetic trajectories in eleven hours, which it calls the equivalent of 6,500 hours of human demonstration data, and that mixing them with real data improved its humanoid model by 40 percent; its technical blog the same day says over 750,000. The CAICT need estimate is in real hours. ISO's draft humanoid dataset standard covers training, validation and test sets without saying how synthetic and real hours compare.
The verdict
Open question
No settled answer exists. The ledger poses the question, links the live debate, and does not answer it.
Key facts
What the sources say
- Record ID
- EAD-2026-0025
- Kind
- Open question
- Jurisdiction
- Global
- Last verified
- Added
- NVIDIA newsroom, 18 March 2025: 780,000 synthetic trajectories, the equivalent of 6,500 hours or nine continuous months of human demonstration data, in eleven hours; a 40 percent improvement over real data alone.
- NVIDIA developer blog, same day: over 750,000 synthetic trajectories in eleven hours. The two company sources disagree by thirty thousand.
- The CAICT estimate of ten million hours is stated as effective real data. Whether a synthetic hour is a fraction of a real one, or none, or more, is not defined anywhere the ledger found.
- ISO/CD 26264-1's abstract speaks of training, validation and test datasets and of quality requirements; the catalog entry does not mention synthetic data.
Figures
Every number, with who measured it and when
- 780,000 trajectories
Synthetic trajectories generated in eleven hours, per the press release
NVIDIA newsroom, Isaac GR00T N1 release, primary source, as of .
- 6,500 hours
Human demonstration hours the company says that equals
NVIDIA newsroom, Isaac GR00T N1 release, primary source, as of .
What it changes
For a robot maker collecting or buying hours
If synthetic hours count at par, the ten million hour gap closes in weeks of compute and the training grounds are a detour. If they count for little, the grounds are the moat. Every buyer of hours is betting on the exchange rate, and no standards body has published one. A maker should keep the two kinds of hours separately counted in its own records until someone does.
Sources
What this record was verified against
- NVIDIA newsroom, Isaac GR00T N1 releasePrimary · 18 March 2025
- NVIDIA developer blog, Isaac GR00T N1Primary · 18 March 2025
- ISO catalog, ISO/CD 26264-1Primary · 13 September 2026
Related
Records that sit beside this one
NVIDIA's synthetic supply: 780,000 trajectories in eleven hours, and an open simulated set of about 273,000
United States · verified 13 September 2026
NVIDIA newsroom, 18 March 2025: 780,000 synthetic trajectories generated in eleven hours; the developer blog of the same day says over 750,000.
The hours gap: a CAICT report puts the need on the order of ten million hours and the world's usable supply at under a million
China · verified 13 September 2026
The report's framing, per the Xinhua summary: a 55 billion parameter model needs about 1.1 trillion tokens; if all of it came from real robot data, that is roughly 21,200 robots working a full year, or effective data at the ten million hour scale.
ISO is drafting a humanoid robot datasets standard, ISO/CD 26264-1, at committee draft stage
Global · verified 13 September 2026
ISO catalog entry 93011: status under development, stage 30.60 close of comment period, edition 1, technical committee ISO/TC 299.
The largest open supply of hours is human, not robot: Ego4D holds over 3,670 hours of first person video
Global · verified 13 September 2026
Ego4D site: over 3,670 hours of daily life activity video, 923 unique participants, 74 locations in 9 countries, released 17 February 2022.
Open question: does a person in a training ground's footage have a say? Three regimes, three different answers, none written for robots
Global · verified 13 September 2026
GDPR and the EDPB video guidelines: footage of an identifiable person is personal data; it is biometric only when processed to identify. Workers are owed a lawful basis and information, and consent is a weak basis at work.
Open question: who owns the hours a robot records at a customer's site? Only the EU has a rule, and it is about access, not ownership
Global · verified 13 September 2026
Data Act articles 4 and 5: the user may access readily available data and have it shared with a third party; article 6(2)(e) bars the third party from a competing product. The Regulation does not use the word ownership.
Cite this record
Free to reuse under CC BY 4.0, with attribution. The record ID EAD-2026-0025 is permanent and is never reused.
- In a sentence
- According to the GAGE Embodied AI Data Ledger (as of 13 September 2026), open question: do synthetic hours count? nvidia generated the equivalent of 6,500 human hours in eleven, and no standard says what an hour is.
- APA
- GAGE (Global Academy of Generative-AI Education). (2026). Open question: do synthetic hours count? NVIDIA generated the equivalent of 6,500 human hours in eleven, and no standard says what an hour is. Embodied AI Data Ledger. Retrieved 13 September 2026, from https://www.gage.academy/tools/embodied-ai-data-ledger/records/EAD-2026-0025-do-synthetic-hours-count
- MLA
- "Open question: do synthetic hours count? NVIDIA generated the equivalent of 6,500 human hours in eleven, and no standard says what an hour is." Embodied AI Data Ledger, GAGE (Global Academy of Generative-AI Education), 13 September 2026, https://www.gage.academy/tools/embodied-ai-data-ledger/records/EAD-2026-0025-do-synthetic-hours-count.
- Chicago
- GAGE (Global Academy of Generative-AI Education). "Open question: do synthetic hours count? NVIDIA generated the equivalent of 6,500 human hours in eleven, and no standard says what an hour is." Embodied AI Data Ledger. Last modified 13 September 2026. https://www.gage.academy/tools/embodied-ai-data-ledger/records/EAD-2026-0025-do-synthetic-hours-count.
- Permalink
- https://www.gage.academy/tools/embodied-ai-data-ledger/records/EAD-2026-0025-do-synthetic-hours-count
Last updated . Every record re verified . The ledger is checked every Monday, and the same day for any announcement by a data regulator.
Back to the full ledger, or every record for Global and every open question record.
GAGE briefings tell you which AI regulation deadlines are coming, what they actually require of you, and when a program opens.