$100B will flow to OpenAI while NVIDIA instantly gained $170B in market cap - a masterclass in co-opetition where AI titans collaborate and compete simultaneously.

When circular investments create exponential value

The deal structure reveals a sophisticated partnership strategy

Nvidia invests $100B in OpenAI through non-voting shares, which OpenAI uses to purchase Nvidia’s data center chips.

What analysts flag as “circular” investment is brilliant ecosystem alignment.

Nvidia secures its largest customer while OpenAI gets compute infrastructure for 10 GW of processing power, ~ like serving 8M households.

This would involve 4-5M of Nvidia chips. Total cost of this 10 GW of AI infra would be ~$400B including chips, land and physical infrastructure (FT).

NVIDIA is getting hands-on across the entire stack

It recently took a $5B stake in Intel Corporation that will cause tighter GPU/CPU integrations, plus targeted bets on datacenter infrastructure startups. The blueprint: finance demand, secure capacity, lock in distribution.

The customer scale justifies everything

OpenAI now serves 700M weekly active users, demonstrating AI demand that warrants this partnership level.

Since the OpenAI deal is additive to everything previously committed, AI infrastructure buildout is accelerating faster than even most optimistic forecasts.

Market validation meets growing co-opetition

The market’s immediate response validates the partnership logic. On the announcement Nvidia jumped ▲4%, Nasdaq ▲0.5%.

Oracle, which is building datacenters, surged ▲6% alongside Nvidia’s gains, recognizing how this investment amplifies commitments for cloud infrastructure spending.

Nvidia’s move pressures other major players

Google, Amazon, Microsoft and Meta now face a choice: match these AI capex investments or risk falling behind. Spending decisions could mean tens or hundreds of billions annually.

What’s more, they all closely work with Nvidia and (many of them) with Open AI.

Meanwhile, hyperscalers were already in a CapEx arms race:

Amazon investing $100B+ annually, Microsoft guided “over $30B” per quarter, Google lifted FY25 CapEx to ~$85B, plus $72B commitment from Meta.

Combined, the four were already expected to spend $400B+ on capex in the coming year (WSJ).

For perspective: the entire Apollo project cost ~$280B (over 13 years), adjusted for inflation.

Partnership Playbook for the AI Era

Meet partnership-as-infrastructure model.

Winners won’t just build great products—they’ll architect capital, capacity, and channels together.

They compete and collaborate fiercely in equal measure, participating in each other’s ecosystems while building their own.

How are you adjusting your GTM to this new era of AI-cloud partnerships?