H100s were introduced in 2022, so they could not have been coming off five-year contracts.
Resolved true The record shows it happened or was so.
Factual: Statements about the present or the past, checked against the record: did they hold up?
What would count as true or false
Judged as of 2026-07-09. True: Nvidia first announced or launched the H100 in 2022, so that fewer than five years had passed by July 2026 and no five-year contract on H100s could have run its full term. Mostly true: the H100 dates from 2022 on one reasonable reading (for example announced in March 2022 with volume shipments in late 2022 or 2023), and the conclusion about contracts still holds. Mostly false: the H100 was introduced in a different year but still within five years of July 2026, so the conclusion holds while the stated year is wrong. False: H100s were available early enough (before mid-2021) that a five-year contract could have ended by July 2026.
Set before the research began.
Evidence (3)
-
E1 NVIDIA Hopper in Full Production GlobeNewswire (NVIDIA press release) · 20 September 2022
“NVIDIA today announced that the NVIDIA H100 Tensor Core GPU is in full production, with global tech partners planning in October to roll out the first wave of products and services based on the groundbreaking NVIDIA Hopper™ architecture.”
-
E2 NVIDIA Hopper GPUs Expand Reach as Demand for AI Grows NVIDIA (investor relations) · 21 March 2023
“The H100 began shipping in the fall in individual and select board units from global manufacturers.”
-
E3 Nvidia Data Center GPUs — Timeline What Led To · 2 September 2026
“Nvidia announced the Hopper H100 in March 2022 and said it began shipping the GPU in November that year; Hopper was later extended with the H200.”
Reasoning
NVIDIA's September 2022 release says the H100 was in full production, with partners rolling out the first products in October (E1). NVIDIA's March 2023 release says the H100 began shipping the previous fall (E2). A third-party timeline dates the announcement to March 2022 and the first shipments to November 2022 (E3). The H100 was therefore introduced in 2022. Fewer than five years had passed by July 2026, so no five-year contract on H100s could have run its full term.
Cite this claim
Michael Burry, “H100s were introduced in 2022, so they could not have been coming off five-year contracts.” (said 9 July 2026). Verdict: resolved true. Reliability API, claim c_ac819f7fd89e, method 0.1.0+c876030c, dataset release v1.1.2. https://epistacklabs.com/claims/c_ac819f7fd89e
This claim is in dataset release v1.1.2, with the same verdict.
Reply to this claim
More factual claims by Michael Burry
-
Mixed
“This caused additional expense across 9 months of $677 million.”
-
Mixed
“His contracts run five years, his A100s stay fully booked, and a batch of H100s coming off contract were leased right back at 95% of the original price.”
-
Mostly true
“During the same quarter, Meta (META) extended its useful server lives to 5.5 years, reducing the depreciation hit to net income by $2.9 billion.”
-
Mostly true
“Intrator’s consistent message is that Nvidia’s chips get rented out on successive contracts so the idea depreciation schedules should be less than 5 or 6 years is ignorant.”
-
Resolved true
“He told Jim Cramer his customers are willing to rent GPUs for six to seven years”
Intrator told Jim Cramer that his customers are willing to rent GPUs for six to seven years.
Verdicts are made automatically; how verdicts are made.