OpenAI currently generates $25B/year, and it is not a profitable company..
Yesterday I also saw that there is only 22% the will IPO this year, so most likely 2027, this is where they could get part of the money. They are expected to IPO at $1T valuation and companies usually sell 15% to 25% of the shares which would give them 150-250B..
This is type of the investment I would never consider regardless of how fast it is growing..
AVGO ($350B)
OpenAI has signed a deal with Broadcom to design OpenAI accelerators and new chip infrastructure.. The deal stretches over 4 years, but the later hardware deployment estimates are over 10 years..
ORCL ($300B)
This 5-year cloud computing deal targets to heavily expand data center capacity and its tied to "Stargate" data center initiative..
MSFT ($250B)
Ongoing compute commitments for Azure cloud infrastructure and inference costs.. The deal is signed over 6 years..
AMZN ($100B)
Built on an initial $38B compute agreement OpenAI extended the deal with AWS for additional $100B that is supposed to be paid out over 8 years..
$NVDA ($100B could go up to $600B)
Mass purchasing and leasing of the top GPUs, massively intertwined with a big strategic investing round.. There are talks between two companies that would dramatically increase the size of the deal up to $600B total.. Nvidia would finance OpenAI's 350B purchase of AI Nvidia chips..
$AMD ($90B)
Multi year deals to purchase AMD instinct accelerators over 4-5 years.. Gradual scaling toward the 6-gigawatt ceiling across future generations of AMD hardware, projected through approximately 2030. AMD also has option to get 10% of OpenAI stake..
$CRWV ($22B)
Specialized data center space and GPU cloud rental to assist model training. The deal is signed for the next 5 years..
Do you think OpenAI can find a way to pay out all of this commitments or the link is going to break at some point?