OpenAI’s record-breaking $110 billion fundraising round, announced on February 27, marks not just the largest private capital raise in tech history—it also signals the arrival of a new financing architecture for artificial intelligence: Compute-backed financing. Rather than a traditional venture round, the deal ties OpenAI’s financing directly to the cloud, compute, and semiconductor suppliers that will power its next generation of AI systems.
The round, led by Amazon ($50 billion), Nvidia ($30 billion) and SoftBank ($30 billion), values OpenAI at a $730 billion pre-money valuation, with more investors expected to join as the raise remains open.
Industry observers say the most consequential part of OpenAI’s $110 billion raise is not the headline number but its structure. Each major investor is also a core infrastructure supplier, and their financial commitments are intertwined with long-term compute deals, effectively guaranteeing that OpenAI will spend vast sums on their platforms over the coming decade.
A significant share of the round from Amazon and Nvidia is likely composed of services, compute credits, or extended cloud infrastructure commitments rather than pure cash—a structure with clear precedent in OpenAI’s earlier fundraising, where large portions of investment from infrastructure partners took the form of non-cash contributions, according to TechCrunch.
“Leadership will be defined by who can scale infrastructure fast enough to meet demand,” OpenAI said in its announcement, underscoring the centrality of compute to its strategy.
The compute-backed model also helps OpenAI remain relatively asset-light. While it raises equity linked to future infrastructure consumption, the capital expenditures required to build out physical data centres are increasingly borne by its partners.
Alongside its $50 billion equity commitment, Amazon has inked a large expansion of OpenAI’s cloud spend. The company will increase its existing $38 billion AWS agreement by another $100 billion over eight years, making AWS the exclusive third-party cloud distributor for OpenAI’s newly launched enterprise platform, Frontier.
As part of the pact, OpenAI will also commit to consuming 2 gigawatts of AWS Trainium compute, Amazon’s in-house AI accelerator technology.
The structure effectively creates a strategic loop: Amazon invests in OpenAI, OpenAI commits to enormous long-term spending on AWS infrastructure, and AWS, in turn, deepens its position in the enterprise AI ecosystem.
Nvidia’s $30 billion investment similarly aligns with a sweeping infrastructure partnership. OpenAI said it has secured next-generation inference compute from Nvidia as part of the deal, reinforcing the chipmaker’s role at the heart of OpenAI’s model training pipeline.
This expands upon the strategic roadmap between the two companies. While OpenAI and Nvidia previously signed a Letter of Intent (LOI) to eventually deploy 10 gigawatts of AI infrastructure—a scale equivalent to the output of roughly 10 nuclear reactors—this specific $30 billion round formalises the first 5 gigawatts of that capacity. This rollout, which will utilise an estimated 2 to 2.5 million GPUs, is slated to begin coming online in the second half of 2026.
The circularity of this model has been noted before. Analysts have described earlier Nvidia-OpenAI arrangements as a “virtuous loop,” where Nvidia invests heavily in OpenAI and OpenAI spends heavily on Nvidia hardware.
SoftBank, the third major participant in the round, adds another $30 billion to the funding pot, continuing its pattern of large-scale backing of frontier model companies. Though the company has not disclosed linked infrastructure commitments, it is a significant party to OpenAI’s broader ecosystem of data centre funding and compute procurement.
OpenAI’s $110 billion round sets a new precedent: mega rounds for frontier AI will increasingly be driven by strategic suppliers underwriting their own future demand. For cloud platforms and chipmakers, owning a slice of the leading AI lab is less about financial return and more about securing infrastructure dominance in an era where compute is the new oil.
The model could reshape how the entire AI sector is funded—tying capital to compute commitments in ways that bind suppliers and model developers far more tightly than in traditional tech ecosystems.
First Published: Mar 02, 2026, 16:23
Subscribe Now