NVIDIA is continuing to increase its capital deployment in the global artificial intelligence ecosystem. In the past few months of 2026, it has committed more than US$40 billion in equity investments to a number of AI-related companies, further consolidating its dominant position in the field of AI infrastructure.
According to US financial media CNBC, a large part of this huge investment comes from a single bet on OpenAI: NVIDIA previously announced that it will invest approximately US$30 billion in OpenAI to support the latter’s continued expansion and technology iteration in large model training and inference. At the same time, the chip giant has also made intensive investments in multiple listed companies, announcing seven billion-dollar equity investments, including the recently disclosed investments of up to $3.2 billion in glass manufacturer Corning and up to $2.1 billion in data center operator IREN.
In the field of start-ups, NVIDIA is also making frequent moves. TechCrunch's previous review showed that Nvidia participated in 67 venture capital projects for AI startups in 2025, covering multiple tracks from basic models, vertical industry applications to development tools. According to statistics from financial data service provider FactSet, so far in 2026 alone, Nvidia has participated in about twenty financing rounds of unlisted AI startups, and continues to consolidate the peripheral ecology of its "AI empire" through capital support.
A large part of this series of investments are also customers of Nvidia, which has also triggered continued controversy in the market and industry about "circular transactions". Critics believe that when Nvidia sells large-scale computing power and chips to customer companies and then reinvests funds in the form of equity investment, the repeated flow of funds among the same group of companies may amplify valuation bubbles and distort the market's judgment of real demand and profitability to a certain extent.
However, Wall Street institutions also give different perspectives. Wedbush Securities analyst Matthew Bryson said that Nvidia's current investment behavior "falls completely within the theme of circular investment," but if these invested companies eventually gain a foothold in their respective segments, it will help build a wider and deeper "moat" for Nvidia. In his view, by deeply binding key customers and ecological partners, Nvidia can not only further lock in the mid- to long-term demand for its new generation of AI chips such as Blackwell and Rubin, but is also expected to gain stronger voice and bargaining power in AI infrastructure, data center construction, and upper-layer applications.
Against this background, Nvidia has shifted from a pure chip supplier to an "AI infrastructure operator" integrating hardware, software and capital. With the previous expectations given by the company's management - the cumulative revenue of the Blackwell and Rubin series chips is expected to reach at least US$1 trillion by the end of 2027 - the market capitalization giant obviously hopes to firmly bind itself to the core players of the next generation AI infrastructure through intensive equity investment, thereby occupying a more durable structural advantage in the new round of AI competition.