Nvidia builds a $90 billion moat to lock in AI partners

Nvidia is transforming its cash from selling AI chips into capital, committing nearly $90 billion to ensure no one can build an AI future without it.

Nvidia Corp. is deploying its massive cash flow to reshape the artificial intelligence industry, committing approximately $90 billion in 16 months to investments and partnerships that lock customers and suppliers into its technology. The strategy, which spans more than 145 companies, aims to make Nvidia’s hardware and software the indispensable foundation of the AI economy, but has also drawn formal inquiries from regulators in the U.S., European Union, and the United Kingdom.

"Nvidia’s investment logic is 'ensuring both solutions will always work well together,'" said Patrick Little, chief executive of chip design startup SiFive, who confirmed his company received an investment after agreeing to make its technology compatible with Nvidia’s proprietary NVLink interconnect standard.

The scale of the spending, detailed in company filings, represents about 40 percent of Nvidia’s operating cash flow in its most recent fiscal year. The outlay dwarfs the venture activity of other tech giants, with Alphabet’s comparable investment ratio standing at just six percent of its cash flow. The deals include a $2 billion stake in chip designer Marvell Technology Inc. and a complex agreement with cloud firm Iren, where Nvidia will invest up to $2.1 billion while also paying Iren $3.4 billion over five years to rent GPU capacity.

This campaign to build a sprawling, loyal network of partners serves as a powerful moat against two rising threats: large customers like Amazon and Google developing their own AI chips, and the complete loss of the Chinese market due to U.S. export controls. While solidifying its market dominance, the web of investments, which makes Nvidia a customer, supplier, and shareholder to many firms simultaneously, has raised antitrust concerns globally.

From Cloud Challengers to Supply Chain Control

Nvidia’s investment targets reveal a clear strategy to control key chokepoints in the AI value chain. The company has funneled capital into a new generation of cloud computing providers, such as CoreWeave and Iren, creating both new customers for its GPUs and alternatives to the large-scale cloud providers who are also its biggest competitors. CEO Jensen Huang has stated that CoreWeave "wouldn't exist without Nvidia's support," highlighting the kingmaker role the company is now playing.

Beyond cloud infrastructure, Nvidia has committed $95 billion as of January to lock in its supply chain. This includes multi-billion dollar deals with photonics companies Coherent Inc. and Lumentum Holdings Inc., as well as a $3.2 billion warrant investment in Corning Inc., a key producer of the fiber optics essential for high-speed data centers. These moves come as the entire AI industry faces a critical shortage of high-bandwidth memory (HBM), a bottleneck that has sent shares of memory makers like Micron Technology Inc. soaring more than 80% in the past month. By funding supplier expansion, Nvidia secures its own access to critical components and increases its bargaining power in a capacity-constrained environment.

A Strategy Forged by Geopolitical Risk

The aggressive push to build a diversified, loyal ecosystem is heavily influenced by the total collapse of Nvidia’s business in China. Before U.S. export restrictions intensified, China accounted for at least 20 percent of the company's data center revenue. Today, that figure is zero. The financial impact was stark, including a $4.5 billion charge in the first quarter of 2026 related to now-unshippable inventory. This revenue void helps explain the urgency behind Nvidia's strategy to foster new markets and partners.

The company is actively encouraging its portfolio of 145-plus companies to adopt its open-source Nemotron AI model, aiming to replicate the success of its proprietary CUDA software platform, which represents one of its strongest competitive barriers. The logic, as expressed by Huang, is to ensure developers worldwide build on an American tech stack, preventing the rise of a parallel ecosystem centered on rivals like Huawei's Ascend platform. For investors, this strategy represents a long-term bulwark against competition, though the recent underperformance of Nvidia's stock—up 19% in a month where the SOXX semiconductor ETF rallied over 40%—suggests the market is actively weighing the significant, unquantified risk from the ongoing regulatory probes in Washington, Brussels, and London.

This article is for informational purposes only and does not constitute investment advice.