After Microsoft invested $1B in OpenAI in 2019, the company scrambled to string together tens of thousands of Nvidia A100 GPUs, costing several hundred million
When Microsoft Corp. invested $1 billion in OpenAI in 2019, it agreed to build a massive, cutting-edge supercomputer for the artificial intelligence research startup.
Context & Ripple Effects
When Microsoft signed its $1B partnership with OpenAI in 2019, the headline number was equity — but the real bill arrived as hardware: assembling tens of thousands of Nvidia A100 GPUs for an OpenAI-dedicated supercomputer ran to several hundred million dollars, per this Bloomberg account. That made the 2019 deal as much a compute commitment as a financial one.
The pattern only escalated from there: the reported $10B multiyear investment in January 2023 deepened the Azure tie, and by late 2025 Microsoft held a ~$135B, ~27% stake alongside OpenAI's commitment to buy an additional $250B of Azure services. This story is the origin point of that structure — the moment cloud capital became AI compute capital.
First-order effects
- Microsoft absorbed a several-hundred-million-dollar GPU procurement bill on top of its $1B check, effectively prepaying OpenAI's research capacity out of its own datacenter budget before any revenue existed.
- Nvidia became the immediate beneficiary and bottleneck: Microsoft's ability to honor the 2019 agreement depended on securing tens of thousands of A100s during a period when those parts were the industry's scarcest input.
Second-order effects
- Rival clouds were forced to match the structure, not just the spending — offering frontier labs dedicated supercomputing capacity became table stakes for winning their workloads and the halo of hosting the field's flagship model.
- OpenAI's leverage grew with each round: having demonstrated its ability to pull billion-dollar infrastructure commitments from Microsoft, it could later shop its capacity needs elsewhere, including the Nvidia–OpenAI investment deal that put the chip supplier itself in the investor seat.
Third-order effects
- The 2019 episode established the template now standard across the industry: AI funding rounds are inseparable from compute purchase commitments, with equity, cloud credits, and GPU supply negotiated as one package rather than separately.
- If the pattern holds, the boundary between investor, supplier, and customer keeps dissolving — Microsoft is simultaneously OpenAI's backer, landlord, and vendor — which concentrates systemic risk in the circular flow of capital among a handful of players.
The trend: AI investment has evolved from cash-for-equity into bundled compute-and-capital packages, where a lab's funding round is also its datacenter procurement contract.