As the artificial intelligence (AI) compute race enters a 'money‑burning' era, tech giants NVIDIA and Oracle have each unveiled major collaborations that have ignited tech stocks. On September 22, NVIDIA announced it would invest up to $100 billion to support OpenAI in building at least 10 gigawatts (GW) of AI data‑center systems; the news sent its shares up 4%, hitting a record high. Just a week earlier, Oracle had agreed a $300 billion cloud‑services memorandum of intent with OpenAI, to be executed over five years starting in 2027; after that announcement Oracle shares jumped as much as 43% in a single day, the largest one‑day gain since 1992, briefly propelling founder Larry Ellison to the top of the world's richest list. These blockbuster deals not only underscore OpenAI's 'hunger' for AI infrastructure but also amplified market optimism for chips and cloud services, helping push the Nasdaq to record highs for three consecutive days.
Under the AI boom, compute has become the 'new oil.' OpenAI, the force behind ChatGPT, is accelerating toward artificial general intelligence (AGI), but training models requires immense numbers of GPUs and large‑scale data centers. NVIDIA's H100 and Blackwell chips dominate the market, while Oracle's cloud platform offers elastic scaling. Wall Street analysts say these partnerships mark a shift from 'proof of concept' to 'industrial‑scale deployment,' and estimate global AI infrastructure investment could exceed $1 trillion by 2030. However, the hefty spending raises concerns: long payback cycles, massive energy consumption and potential bubble risks.
Oracle's $300B 'Deal of the Century': From ERP Veteran to AI Cloud Powerhouse
The story began on September 10. The Wall Street Journal first reported that OpenAI and Oracle had signed a cloud‑services agreement valued at $300 billion, one of the largest single contracts in tech history. The deal covers five years of compute procurement starting in 2027—spanning cloud servers, storage and AI‑optimized services—and is set to cover roughly half of OpenAI's planned U.S. data‑center buildout. Oracle CEO Safra Catz confirmed on an earnings call that the contract is a core part of the company's $317 billion in new orders and said it would lift cloud revenue share from about 25% to over 50%.
Oracle's turnaround is striking. Long known for ERP and databases, the software giant trailed AWS and Azure in cloud services, but the AI wave changed the game. As early as 2024, Oracle partnered with OpenAI to build the Stargate super data center and invested tens of billions. The $300 billion agreement further cements OpenAI as an anchor customer and positions Oracle Cloud Infrastructure (OCI) as a preferred AI training platform. Catz said, 'This is not a one‑off deal but a strategic partnership to help OpenAI achieve AGI while delivering steady cash flow for us.'
The market reaction was explosive. On September 11 Oracle opened up 36% and closed up 43%, temporarily erasing competitors' market‑cap gaps and sending Ellison briefly past Elon Musk on the richest‑person list. Trading volume surged to about ten times normal levels, and banks such as JPMorgan raised price targets to $180, calling Oracle's shift from database vendor to AI infrastructure leader. Yet optimism carries risks. Moody's warned the deal is largely 'non‑binding' in parts and depends on OpenAI's financing and technical readiness; if an AI bubble bursts, Oracle could face significant write‑offs. Fortune bluntly noted, 'The $300 billion deal rings an AI bubble alarm—grand in headline but with returns far from guaranteed.'
NVIDIA's $100B 'Compute Bet': The Chip Empire's AI Crown
A week later NVIDIA took center stage. On September 22 NVIDIA and OpenAI announced a strategic partnership: NVIDIA will invest up to $100 billion to deploy more than 10 GW of AI systems in stages to train OpenAI's next‑generation models such as GPT‑6. The plan includes supplying millions of H100/B100 GPUs and Grace Hopper superchips, with the first GW system slated to come online in 2026. NVIDIA CEO Jensen Huang called it 'the largest AI infrastructure project in history,' saying it will accelerate OpenAI's transition from 'chatbot' to 'general intelligence.'
NVIDIA's positioning is natural. As the AI chip leader, its GPUs account for more than 90% of the market, and fiscal 2025 revenue exceeded $200 billion. OpenAI's appetite for compute is acute: CEO Sam Altman has repeatedly cited shortages, and while OpenAI has already secured hundreds of millions with Microsoft, NVIDIA's deep tie‑up is more like a crown jewel. The agreement also includes joint R&D on custom chips to guard against competition from AMD and Intel.
Stocks celebrated. On September 22 NVIDIA hit an intraday high of $184.55 and closed at $183.61, up 3.9%, briefly pushing market value above $3.5 trillion. The Nasdaq extended record highs, and the Dow and S&P 500 benefited from tech gains. Analysts such as Blayne Curtis at Barclays raised ratings to 'overweight,' forecasting NVIDIA's revenue could double by 2026. CNBC commented, 'NVIDIA isn't merely investing in OpenAI—it's betting on the entire future of AI.'
The AI Compute 'Arms Race': Opportunities and Worries
These two deals link the AI ecosystem: Oracle handles 'cloud,' NVIDIA supplies 'chips,' and OpenAI is the 'downstream engine.' Business Insider noted that a 10 GW data center is roughly equivalent to the power demand of 100 nuclear plants, highlighting an energy bottleneck. OpenAI's ambition is clear: Altman has said 'trillions of dollars' will be needed to build AGI factories, and these contracts are only the tip of the iceberg.
The opportunities are huge: tech stocks are reshaping the landscape. Oracle has transformed from a 'sunset' business to an AI upstart with a market value above $1 trillion, while NVIDIA continues its chip dominance. Barron's projects cloud and chip markets could grow 30% in 2025. But risks remain: many deals are non‑binding and dependent on OpenAI financing (such as a recent 6.6 billion funding round); energy constraints, geopolitics (including the US‑China chip tensions) or a bursting bubble could reverse gains.
Looking ahead, the AI compute race will intensify. Microsoft and Amazon may enter more deeply, and regulation (such as the EU AI Act) could tighten. Jensen Huang posted on X: 'AI is not a bubble; it's an infrastructure revolution.' For investors it's a frenzy; for the global economy it's a crossroads of opportunity and challenge.