Not content with just selling shovels! Jensen Huang spent $90 billion in 16 months, investing in almost the entire AI industry.

```

Nvidia is transforming itself from a chip manufacturer into the core capital force of the entire artificial intelligence industry.

According to a report by the Financial Times on Tuesday, over the past 16 months, Nvidia has committed about $90 billion to investments and cooperative deals, covering more than 145 companies, including AI model developers, cloud computing service providers, and infrastructure suppliers.

This scale is already on par with the largest venture capital arms of major tech companies, making Nvidia the most aggressive dealmaker in the technology sector. Nvidia will report its quarterly results this Wednesday, which the market sees as a key indicator of global AI spending.

Meanwhile, Nvidia announced that it will set up its first research center in Singapore, its second research institution in the Asia-Pacific region, focusing on embodied AI and improving the efficiency of AI infrastructure. This move is highly in line with Singapore’s strategy to position itself as a regional AI hub, further demonstrating Nvidia's ambition in the global AI landscape.

While this large-scale transaction offensive accelerates the expansion of the AI ecosystem, it also places Nvidia at the intersection of customers, suppliers, and potential competitors, and has already attracted the attention of global regulatory agencies.

Where does the $90 billion come from: Two parallel funding streams

According to the Financial Times, citing company disclosures and PitchBook data, Nvidia committed about $47 billion to investments and partnerships in the fiscal year ending January 25, and a further $43 billion in the following four months.

This spending accounts for about 40% of Nvidia's latest fiscal year operating cash flow, far higher than Alphabet’s — traditionally seen as the largest tech investor in startups — approximate 6% cash flow investment ratio.

There are two parallel deal advancing mechanisms within Nvidia. The company has a venture capital division called NVentures, but according to insiders, the business development team led most of the recent deals, with investments often accompanied by broader commercial cooperation agreements. A San Francisco lawyer described that after Nvidia’s technical team contacts a target firm, business development representatives arrive “with checks in hand,” and “two parallel conversations advance simultaneously.”

Binding the ecosystem: From chip interconnection to open-source models

Nvidia’s investment logic is not purely financial returns, but to use capital as a link to deeply bind partners into its own technology ecosystem.

Patrick Little, CEO of semiconductor design startup SiFive, revealed that after the two sides agreed to make SiFive chip designs compatible with NVLink — Nvidia’s proprietary interconnection technology — Nvidia immediately invested in SiFive. He stated that Nvidia’s investment logic is "to ensure that both solutions always cooperate well."

Nvidia reached a similar arrangement with chip designer Marvell, taking a $2 billion stake this March while signing a cooperation agreement to ensure Marvell’s future custom chips are compatible with NVLink. Marvell is the chip designer behind Amazon’s Trainium AI accelerator.

On the software side, according to two people familiar with the matter, Nvidia is also actively encouraging its investees to adopt its open-source AI model Nemotron. Jensen Huang hopes Nemotron can replicate the success of CUDA — one of Nvidia’s strongest competitive moats with its proprietary software platform. A venture capitalist commented, “Founders are starting to realize that if you build on the Nvidia ecosystem, you can get funding from Jensen Huang.”

New cloud computing power: Client, supplier, and shareholder all in one

In Nvidia’s deal map, its bets on the new generation of cloud computing companies are particularly notable and also raise the most concern among analysts.

Jensen Huang once publicly stated that CoreWeave “would not exist without Nvidia’s support.” Nvidia supports these emerging AI infrastructure providers partly because tech giants like Google and Amazon are not only Nvidia’s biggest customers but also increasingly pose competitive threats by developing their own chips.

Earlier this month, Nvidia reached a deal with emerging cloud company Iren, committing to spend $3.4 billion over five years to lease its GPU computing power, and investing up to $2.1 billion for an equity stake. This means Nvidia is simultaneously acting as client, supplier, and potential shareholder.

Nvidia’s largest single transaction last year was a $20 billion agreement with chip designer Groq, covering technology licensing and talent acquisition. As AI workloads shift toward "inference" computing, Groq’s processors have a clear advantage in this area, and Nvidia has already launched products based on Groq technology.

Securing the supply chain: Locking in critical capacity

Beyond equity investments, Nvidia had also committed $95 billion as of the end of January to secure components supply and manufacturing capacity.

Recently announced major supplier deals include: investing $2 billion each in photonics companies Coherent and Lumentum, and investing $3.2 billion in warrants for fiber manufacturer Corning — the latter produces the fiber used in high-speed data centers. All three companies are major suppliers to Apple, and Nvidia’s investments are making it a core customer to these suppliers.

Moon Surana, portfolio manager at asset management firm Harding Loevner, said Nvidia’s funding supports suppliers’ expansion plans while “enhancing Nvidia’s bargaining power in the supply chain amid persistent capacity constraints.”

Regulatory risks and Singapore’s layout

The large-scale deal offensive has attracted global regulatory attention. Nvidia disclosed in its annual report that regulators in the US, EU, UK, and elsewhere have issued “wide-ranging information requests” about its “investments, collaborations, and other agreements with foundation model developers,” as well as agreements with customers, suppliers, and partners.

On the global layout front, Nvidia announced it will establish its first research center in Singapore, its second research institution in the Asia-Pacific region. According to CNBC, the center will focus on advancing embodied AI and improving AI infrastructure efficiency, working in coordination with university researchers, industry partners, and government agencies.

Singapore is positioning itself as a regional AI hub and on the same day announced the launch of a test platform to help private enterprises jointly design, deploy, test, and validate commercial AI robotics technologies. Industry leaders such as Certis, DHL, Grab, and QuikBot are expected to be the first to use the platform.

SiFive’s Patrick Little described Jensen Huang as having a “tunnel vision” for the next “five to ten steps” of the AI market and acting accordingly. “They are not interested in playing just one chess move; they want to see the pawn become a queen.”

Risk warning and disclaimerThe market involves risk, and investment must be made cautiously. This article does not constitute personal investment advice, nor does it take into account the individual investment objectives, financial situation, or needs of any specific user. Users should consider whether any opinions, views, or conclusions in this article are suitable for their own circumstances. Investment based on this is at your own risk. ```