Nvidia has crossed $40 billion in equity commitments to outside companies in 2026, an extraordinary pace of capital deployment for a hardware vendor and a clear signal that CEO Jensen Huang is no longer content to be merely the picks-and-shovels supplier to the artificial intelligence boom. As reported by CNBC, the company has signed at least seven multibillion-dollar agreements with publicly traded firms this year and participated in roughly two dozen private investment rounds, ranging from a $30 billion check to OpenAI to fresh deals this week with data center operator IREN and 175-year-old glass maker Corning. The strategy is at once a competitive moat, a vertical integration play, and, depending on whom you ask, a textbook example of the vendor financing that helped inflate the dot-com bubble.
For shareholders, the question is no longer whether Nvidia is the dominant AI infrastructure company in the world. That argument was settled when the company crossed a $5.2 trillion market capitalization, making it the most valuable business on the planet. The new question is whether Nvidia’s accelerating bets on the rest of the AI stack are buying durable strategic position or pre-funding demand that would have arrived anyway. The answer matters enormously for anyone holding semiconductor exposure, AI infrastructure plays, or the broader S&P 500, given Nvidia’s outsized weight in the index.
The Scale of the Spending
Nvidia generated roughly $97 billion in free cash flow during its last fiscal year, a number that gives the company an unusual level of strategic latitude. Most companies that throw off that much cash spend it on dividends, buybacks, and incremental capex. Nvidia has done the buybacks, but it is pouring an outsized share of its cash into the equity of other companies that the chipmaker either depends on, sells to, or wants to influence.
The pace of 2026 deals is striking. The company has already signed at least seven multibillion-dollar investments with public companies, including the IREN agreement announced this week giving Nvidia the right to invest up to $2.1 billion, and the Corning deal that allows it to invest up to $3.2 billion in the storied glass and optics manufacturer. Both stocks popped on the news, which is its own form of value transfer to existing shareholders before any product roadmap actually ships.
Beyond the public companies, Nvidia has participated in roughly two dozen private rounds, including stakes in some early-stage businesses, according to FactSet. The single largest commitment, by an enormous margin, is the $30 billion check Nvidia wrote to OpenAI in late February 2026, deepening a partnership that began more than a decade ago and that has grown increasingly intertwined since the launch of ChatGPT in November 2022. Nvidia also participated in massive funding rounds for Anthropic and Elon Musk’s xAI, with the xAI investment closing shortly before xAI merged with SpaceX in February.
The aggregate scale of these moves makes Nvidia, in effect, one of the largest venture and growth-stage investors in the AI ecosystem, sitting alongside or ahead of dedicated capital pools managed by traditional venture firms and sovereign wealth funds.
The OpenAI Backstory
The OpenAI investment is worth understanding in detail because it captures the dynamic that defines the rest of Nvidia’s strategy. The $30 billion final commitment in late February 2026 was actually a step down from what had originally been announced. In September 2025, Nvidia and OpenAI publicly described a deal that would see Nvidia put in up to $100 billion over time as OpenAI deployed 10 gigawatts of Nvidia’s systems.
That larger deal “never got off the ground,” in CNBC’s framing, as OpenAI pivoted away from operating its own data center buildout and instead leaned more heavily on infrastructure partners like Oracle, Microsoft, and a handful of others. Huang said publicly in March that putting $100 billion into OpenAI is “probably not in the cards,” and that the $30 billion deal “might be the last time” Nvidia writes a check before an OpenAI initial public offering that could happen as soon as later in 2026.
The interesting feature of the OpenAI relationship is that it cuts both ways economically. Nvidia is investing capital into OpenAI. OpenAI, in turn, is one of the largest single buyers of Nvidia GPUs in the world, either directly or through cloud partners that purchase capacity to run OpenAI workloads. The economic flow runs in two directions through the relationship, which is precisely the structure critics have flagged as concerning.
The Vendor Financing Critique
Matthew Bryson, an analyst at Wedbush Securities, has captured the bear case in a memorable phrase. Nvidia’s investments and infrastructure buildouts, Bryson wrote in a note, fit “squarely into the circular investment theme” that has been driving fears around the durability of AI market dynamics. The implication is direct. If Nvidia is funding the customers that buy its products, and those customers’ ability to keep buying depends on continued capital injections from Nvidia, then revenue growth on the income statement is partially a recycling of Nvidia’s own balance sheet rather than independent end-market demand.
This is the structural critique that haunted the late-1990s telecom buildout, when equipment vendors like Lucent extended generous vendor financing to startup carriers, who then used the financing to buy more equipment, until the music stopped and a wave of carrier bankruptcies torched the equipment vendors’ balance sheets. The analogy is imperfect because Nvidia’s investments are equity rather than vendor credit, but the directional concern is similar.
Bryson, for his part, ultimately argues the bull case wins on net. The investments, he wrote, demonstrate Nvidia’s strategic vision and create a “competitive moat” if the company can execute. That is also the case Huang himself has been making in public appearances. “There are so many great, amazing foundation model companies, and we try to invest in all of them,” Huang said in an April podcast. “We don’t pick winners. We need to support everyone.”
The rebuttal to the vendor-financing critique has three elements. First, Nvidia is investing across the entire stack, not just in customers, which limits the circularity argument. Second, the OpenAI bet is profitable purely on equity terms even if no GPU revenue ever materialized, given the trajectory of OpenAI’s valuation. Third, the demand picture, however supplemented, is being matched by real revenue that has lifted Nvidia’s stock more than elevenfold over four years, an outcome that requires actual end-market consumption, not just intracompany shuffling.
The Components Strategy
The most strategically defensible element of Nvidia’s 2026 spending is the cluster of investments in component manufacturers. Nvidia put $2 billion into Marvell Technology in March 2026 as part of a strategic partnership focused on silicon photonics, and the same amount into Lumentum and Coherent, two firms developing photonics technologies essential for the next generation of AI infrastructure.
The Corning deal extends this pattern. As part of the agreement, Corning is building three new US facilities dedicated to optical technologies for Nvidia, which will likely be transitioning from copper interconnects to fiber-optic cables as it builds out its rack-scale systems. The shift from copper to fiber is not a minor update. It is a fundamental redesign of how data moves inside Nvidia’s largest reference systems, and the bottleneck for actually executing it is the supply of advanced optical components.
Chip analyst Jordan Klein at Mizuho captured the logic in an email reported by CNBC. Nvidia’s deals with component makers are “super smart by the CFO and team and a great use of cash,” Klein wrote, because they accelerate development of critical technology and products that are otherwise in short supply. This is classic supply chain management dressed up as financial investment. Nvidia is using its balance sheet to buy out a manufacturing constraint that would otherwise slow its product roadmap.
Klein is more skeptical of the so-called neocloud investments, the deals with companies like CoreWeave and Nebius Group that are building data centers specifically to run Nvidia hardware and lease compute. Those deals, he wrote, “feel more questionable to me and likely investors.” The concern is sharper here. “It smells like you are pre-funding the purchase of your own GPUs and products,” Klein wrote, while acknowledging that the cloud providers offer power and data center capacity that Nvidia genuinely needs.
The Neocloud Bets
The neocloud investments are the cleanest test of Nvidia’s strategic thesis. In January 2026, Nvidia put $2 billion into CoreWeave as part of a deal that involves building out data centers using Nvidia’s technology. Around the same time, Nvidia invested $2 billion in Nebius Group, an AI cloud company, as part of an agreement on AI infrastructure deployment, fleet management, inference, and AI factory design.
Both companies have a similar profile. They lease compute capacity, primarily to AI training and inference customers, on infrastructure that is overwhelmingly Nvidia-based. The investment thesis from Nvidia’s perspective is that these companies provide capacity and operational expertise, particularly around power and physical site management, that is in scarce supply. The investment thesis from the neocloud perspective is that Nvidia capital plus Nvidia hardware allocation plus the implicit endorsement of being a Nvidia-aligned operator is a powerful competitive package.
Ben Bajarin at Creative Strategies put the systemic risk plainly. “The risk is that if the cycle turns, the market starts questioning how much of the demand was organic versus supported by Nvidia’s own balance sheet,” Bajarin told CNBC. That is the precise question. In a benign demand environment, the neocloud bets compound on themselves. In a sharp downturn, they could amplify the cyclical pain rather than buffer it.
For investors comparing this dynamic against past AI bubble worries and the debate over AI startup valuations, the takeaway is that Nvidia’s strategy has not eliminated cyclicality. It has rerouted it. Instead of a single-product chip cycle, Nvidia now has cyclical exposure across financial assets, customer relationships, and component dependencies in a tightly coupled web.
Earnings, Disclosure, and What to Watch
Nvidia is set to report its fiscal first-quarter 2027 results in less than two weeks. That report will give shareholders a clearer picture of the size and accounting impact of the expanding investment portfolio. During the prior fiscal year, Nvidia invested $17.5 billion in private companies and infrastructure funds, “primarily to support early-stage startups,” according to its annual SEC filing. The company disclosed that those investments include AI model companies that purchase its products either directly or through cloud service providers, the language that maps to the circularity concern critics flag.
The disclosure on the balance sheet is even more striking. Non-marketable equity securities, which are the private company investments, swelled to $22.25 billion at the end of January 2026 from $3.39 billion a year earlier. That is roughly a sevenfold increase in a single year. The company also reported gains on those holdings, plus on publicly traded equities, of $8.92 billion, up from $1.03 billion in the prior fiscal year, in significant part due to the Intel investment that has appreciated from $5 billion to over $25 billion in a matter of months.
Three things are worth watching in the upcoming earnings report. The first is the disclosure of any new related-party revenue streams flowing from invested companies back into Nvidia’s top line. The second is any update on impairments or markdowns of existing investments, which would signal that the portfolio is starting to take losses. The third is Huang’s commentary on whether the pace of investment will sustain or begin to taper, given that the company has now committed more in 2026 than it did in all of fiscal 2026.
For longer-term investors thinking about how big-cap technology earnings have been rewarded by the market, Nvidia’s strategy is the boldest version of a pattern that runs through Microsoft, Alphabet, Meta, and Amazon. The largest AI players are using their unprecedented free cash flow to lock in supply, fund their own demand, and shape the architecture of the next computing era. The model has worked so far. The risk profile, particularly if the AI capex cycle ever stalls, is concentrated in ways equity research has only begun to fully unpack.
The Bigger Picture
Nvidia’s $40 billion of 2026 commitments is, in dollar terms, about 40 percent of its prior fiscal year free cash flow. That is an enormous capital allocation choice in service of a single strategic theme. Huang has said in multiple recent appearances that the company’s investments are “focused very squarely, strategically on expanding and deepening our ecosystem reach,” and the empirical evidence supports the framing. Nvidia is no longer just a hardware vendor. It is the central financial node in the AI infrastructure economy.
Whether that position is sustainable depends on three answers. Can the AI demand cycle continue at anything close to its current pace? Can Nvidia’s investments avoid concentrating risk in ways that compound rather than diversify? And can regulators, particularly in the United States and Europe, refrain from intervening as Nvidia accumulates an effectively monopolistic position across multiple layers of the AI stack?
For shareholders, the simplest framing is that Nvidia has converted an unbeatable cash flow advantage into an unprecedented investment program. If it works, the company emerges from the AI cycle with structural ownership of the supply chain. If the cycle turns, the same balance sheet that enabled the buildout will absorb the markdowns. Either way, the next year of Nvidia disclosures will be among the most consequential in the company’s history.
