Introduction
In 2025 the AI investment wave intensified worldwide. As a central engine, OpenAI reshaped the cloud and chip ecosystems through a series of large partnership agreements, directly driving partner stock rallies. From Oracle's 36.7% one‑day surge to AMD's 34% jump, these events have produced an "OpenAI effect": a single agreement can ignite market enthusiasm. However, with giants such as Nvidia and Intel increasingly involved, does this chain reaction signal an expansion of an AI investment bubble? Drawing on authoritative sources including Reuters, The Wall Street Journal, the Stanford AI Index report and Gartner forecasts, and comparing similar cases, this article analyzes sustainability. Overall, while AI investment has structural growth potential, valuation detachment from fundamentals, infrastructure constraints and other factors are brewing a watershed moment between an "internet‑style boom" and a dot‑com‑style bubble.
Event Chain Verification: Stock Frenzy Triggered by OpenAI Deals
OpenAI's strategic moves have become a catalyst in the AI ecosystem. First, on September 10, 2025, OpenAI signed a 5‑year, US$300 billion cloud agreement with Oracle covering cloud infrastructure and data center services. The deal is part of the Stargate hyperscale data center project, with Oracle supplying massive GPU clusters to support OpenAI's model training. After the announcement, Oracle's shares jumped 36.7% in a single day—the largest increase since 1992—and market value temporarily rose by over US$100 billion. The Wall Street Journal reported that the agreement accounts for a large share of Oracle's new revenue and accelerates its shift from traditional databases to AI cloud services.
Shortly after, on September 22, Nvidia announced up to US$100 billion in investment into OpenAI to build 10 gigawatts (GW) of data center capacity. The move tied Nvidia's H100/H200 GPU supply to OpenAI, reinforcing its dominance in AI hardware. Nvidia's stock rose 4.4% that day and had gained nearly 40% year‑to‑date. Reuters noted that although the investment will be injected in phases, it has already boosted Nvidia's market cap to over US$4 trillion.
On September 18, Nvidia further invested US$5 billion to take a stake in Intel, with the two companies collaborating on GPU‑CPU hybrid chips for data center and PC AI acceleration. Nvidia received roughly a 4% stake in Intel, and Intel's shares surged 22%–30%, bringing its market value back to US$150 billion. CNBC said the partnership helps alleviate Intel's lag in AI chips and strengthens its competitiveness against AMD.
The latest link materialized on October 6: AMD signed a multi‑billion to multi‑ten‑billion‑dollar AI chip supply agreement with OpenAI, granting OpenAI the right to purchase 16 million shares of AMD (about a 10% stake, at US$0.01 per share). The agreement covers deployment of 6GW of AMD GPUs, and AMD's stock rose 34% that day, increasing market value by about US$70 billion. The Wall Street Journal described this move as OpenAI diversifying its supply chain and reducing reliance on Nvidia.
These events are not isolated: OpenAI's "computational hunger" has driven a chain of partner stock gains, similar to the Nvidia linkage seen after CoreWeave's financing in 2024.
Comparable Investment Cases: The AI Ecosystem's "Domino Effect"
Comparable cases are numerous and highlight amplification effects in AI investment. In September 2025, Anthropic closed a US$13 billion Series F (led by Amazon and Google), valuing it at US$183 billion; although not publicly traded, that round boosted AWS‑related cloud equities by 5%. Meta committed US$72 billion in annual capital expenditure across the year to AI data centers and acquired a 49% stake in Scale AI for US$14.3 billion, directly lifting its stock by 15%. CoreWeave's US$14.2 billion GPU leasing deal with Meta drove CoreWeave's valuation to US$19 billion and indirectly benefited Nvidia, whose shares gained another 3%.
The Stanford 2025 AI Index shows private AI investment worldwide reached US$252.3 billion in 2024, up 44.5%, with generative AI accounting for US$33.9 billion. Gartner forecasts AI spending of US$1.5 trillion in 2025, nearly 2% of global GDP. Another example is xAI's US$6 billion financing (led by Sequoia) and integration of Musk's Grok model into the X platform, which briefly lifted Tesla's stock by 2%. These investments often create ripples: after Vercel raised US$300 million, its AI tool v0 saw a 10% linked rise in equity value.
However, bubble signs are emerging. Bain warned that AI must generate US$2 trillion in incremental revenue by 2030 to justify current investments; otherwise, history could repeat the dot‑com bust. Reuters commented that although the Oracle‑OpenAI US$30 billion deal lifted stock prices, it relied on debt financing (Oracle borrowing US$100 billion), amplifying risk.
Bubble Analysis: Boom or Mirage?
Is AI investment inflating a bubble? Optimists call it an "internet‑style boom": the Stanford report emphasizes that cumulative AI investments from 2013–2024 reached US$470 billion (U.S.‑led), with potential to drive US$15 trillion in GDP growth. Gartner's 2025 Hype Cycle suggests AI is moving from "inflated expectations" toward a "plateau of productivity," with enterprise adoption at 81%. OpenAI's deal chain (e.g., AMD's 6GW deployment) reflects real demand: model training requires exponential compute, and Nvidia/AMD share gains stem from supply‑demand imbalances.
But negatives cannot be ignored. The Atlantic likened the AI bubble to exceeding the peak of dot‑com telecom investments; by 2025 AI‑related market value topped US$5 trillion while representing only about 2% of global GDP. Reuters highlighted three expanding problems that could inflate a bubble: (1) infrastructure gaps—data centers require US$800 billion/year, but energy shortages delay ROI; (2) valuation disconnect—OpenAI's US$500 billion valuation implies 100× revenue, similar to Cisco in 1999; (3) intensified competition—while the AMD‑OpenAI deal lifted AMD's shares, Nvidia still holds about 90% market share, squeezing smaller firms. Fortune reported AI investment grew 26% in 2025 but only converted 10% into actual revenue, resembling the 2021 SPAC surge.
Compared with other cycles, the AI bubble appears more "rational": the dot‑com crash followed from absent profit models, whereas AI already has revenue—ChatGPT reportedly generates US$10 billion annually. Yet Morningstar warns that Oracle's debt‑funded deal could trigger cascading defaults; if interest rates rose to 5%, AI capital expenditure would shrink by 20%. Overall, bubble risk is moderate: a short‑term boom (2025–2026) followed by medium‑term correction (2027).
Conclusion
OpenAI's partnership chain—Oracle's US$300 billion agreement, Nvidia's US$100 billion investment, Nvidia's US$5 billion stake in Intel, and AMD's 10% equity arrangement—undoubtedly ignited the AI torch and drove stock market value increases exceeding hundreds of billions of dollars. But in light of Anthropic, Meta and other cases, and data from Stanford and Gartner, this is a double‑edged sword: while commitments to up to US$1.5 trillion in spending signal growth, gaps in revenue and stretched valuations pose bubble risks. Investors should beware FOMO‑driven irrational exuberance; as Reuters put it, this could be AI's dot‑com moment. In the long run, AI investment is more transformation than pure bubble—the key is converting compute into sustainable value. Regulation (such as EU AI legislation) and efficiency gains (for example, Mixture of Experts architectures) will likely determine the outcome.