Nvidia is investing $30 billion in OpenAI, and OpenAI is to spend 20+ billion on Nvidia chips.
Mixed Partly true and partly false in roughly equal measure.
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.
What would count as true or false
Judged as of 2026-02-27. Two parts: (a) Nvidia is investing about $30 billion in OpenAI; (b) OpenAI is to spend more than $20 billion on Nvidia chips, as a counterpart to or alongside that investment. True: dated reporting or company announcements around that date show an Nvidia investment in OpenAI of roughly $30 billion (about $27–33 billion counts) that was agreed, announced or under way, and an OpenAI commitment or plan to spend more than $20 billion on Nvidia chips or systems. Mostly true: both parts hold, but there is a small slip, for example the investment amount is a little outside that band but still close to $30 billion, or the chip-spend figure is stated loosely but is clearly above $20 billion. Mixed: one part is confirmed and the other is unsupported or materially different, for example the investment is confirmed but no chip-spend link or figure exists. Mostly false: there is a real Nvidia–OpenAI deal, but the claim badly misstates it, for example the investment amount differs substantially (such as the earlier $100 billion LOI framework misdescribed), or the deal had only been reported as talks when the claim states it as a done investment. False: as of that date there was no reported Nvidia investment in OpenAI of anything like $30 billion and no such OpenAI chip-spend arrangement.
Set before the research began.
Evidence (2)
-
E1 Scaling AI for everyone OpenAI · 27 February 2026
“This includes $30B from SoftBank, $30B from NVIDIA, and $50B from Amazon.”
-
E2 Scaling AI for everyone OpenAI · 27 February 2026
“We are also expanding our long standing collaboration with NVIDIA, including the use of 3 GW of dedicated inference capacity and 2 GW of training on Vera Rubin systems.”
Reasoning
OpenAI's 2026-02-27 announcement (E1) lists "$30B from NVIDIA" in its funding round, which confirms part (a). E2 says the collaboration with NVIDIA is expanding to include 3 GW of dedicated inference capacity and 2 GW of training on Vera Rubin systems. However, it gives no dollar figure and does not describe a commitment by OpenAI to spend more than $20 billion on Nvidia chips. Part (a) is confirmed, but the evidence does not support the specific chip-spend figure in part (b), so the claim resolves as mixed.
Cite this claim
Michael Burry, “Nvidia is investing $30 billion in OpenAI, and OpenAI is to spend 20+ billion on Nvidia chips.” (said 27 February 2026). Verdict: mixed. Reliability API, claim c_7e5b4c13cf70, method 0.1.0+c876030c, dataset release v1.1.2. https://epistacklabs.com/claims/c_7e5b4c13cf70
This claim is in dataset release v1.1.2, with the same verdict.
Reply to this claim
More factual claims by Michael Burry
-
Mixed
“The end of q2 and during q3 a lot changed in the developer landscape, and that is not in this prospectus”
-
Mostly false
“The Superinvestors of Graham & Doddsville all lost money 3 years in a row during the mid-1970s.”
The Superinvestors of Graham & Doddsville all lost money three years in a row during the mid-1970s.
-
Mostly true
“It’s a perfect storm for Build-A-Bear (BBW), a one-of-one retailer now heavily shorted, 25% of the float, almost 9 days to cover.”
-
Resolved true
“Here is the link where I made the case last week that beneath the surface the October 6, 2026 record close in the S&P 500 looks more like the March 2000 top than any other record high since 1980.”
-
Mostly true
“I have never said that LLMs would stop improving”
Michael Burry has never said that LLMs would stop improving.
Verdicts are made automatically; how verdicts are made.