While gold, real estate, and stocks have long been the go-to investment choices, Nvidia CEO Jensen Huang is betting on a new frontier: artificial intelligence chips. The company announced on Monday that it had signed memorandums of understanding with Blackstone, BlackRock, Goldman Sachs, and other large asset managers to build financing platforms for Nvidia's customers. The goal is to mobilize more than $500 billion of third-party capital for the buildout of AI infrastructure over time.
"This is really the first time that technology chips have become an investable asset class," Huang said in an interview. "They are revenue-generating assets now. They are productive, they are long-lived, they are fungible, they are flexible." His comments mark a significant shift in how the industry views GPUs, which historically have been treated as rapidly depreciating hardware. Nvidia's latest efforts challenge that assumption, transforming AI compute capacity into long-term, bankable infrastructure.
"We began by building chips; today, we are helping create a new class of productive, investable infrastructure: AI factories," Huang said in a press release. He noted that in the case of artificial intelligence, compute is revenue, and Nvidia is uniquely positioned to enable this shift. "That is why we are bringing the world's leading long-term capital providers together to independently underwrite AI infrastructure. These financing platforms will help customers access scarce compute at scale and build the DSX AI factories that will power every industry and country in the age of AI," he added.
How AI chips became infrastructure
The concept of treating compute as infrastructure is a departure from the traditional IT procurement model, where companies buy servers and software as one-time expenses with short depreciation cycles. Huang argues that the rise of AI demands a new framework. "Fundamentally, what's different about this industry and this way of doing computing is that the computer is now part of the infrastructure, like electricity, like the internet, and so you have to think about it like infrastructure," he explained.
This framing aligns with the surge in data center construction and the massive capital expenditures by hyperscalers such as Microsoft, Amazon, Google, and Meta. These companies are pouring billions into GPU clusters to train and deploy large language models, but the cost is staggering. Nvidia's financing initiative could help a broader set of enterprises—not just the tech giants—access AI compute without bearing the full upfront cost of purchasing expensive hardware.
Under the proposed structure, asset managers would own the AI infrastructure and lease it to customers, similar to how real estate investment trusts own properties and collect rent. This model has precedent in other capital-intensive industries, such as aircraft leasing or energy infrastructure. By partnering with financial institutions, Nvidia is essentially creating a new asset class that can be financed, securitized, and traded, opening the door to institutional investors seeking exposure to the AI boom without directly buying tech stocks.
Who is involved and what they said
The roster of partners includes some of the most influential names in finance. BlackRock, led by Chairman and CEO Larry Fink, is one of the key collaborators. Fink commented that the AI buildout will require unprecedented investment and a skilled workforce to turn that investment into infrastructure that can power future growth. Blackstone, the world's largest alternative asset manager, is also on board, along with Goldman Sachs, a leading investment bank with deep expertise in structured finance.
KKR, another major private equity firm, is participating as well. Co-CEOs Joe Bae and Scott Nuttall emphasized the importance of delivery over ambition. "As we have scaled our approach to digital infrastructure, we have learned that delivery, not ambition, is the hard part," they said. Their remark reflects the practical challenges of building and operating AI data centers, which require specialized electrical systems, cooling technologies, and supply chain coordination.
Nvidia's move is also a strategic response to the cyclical nature of semiconductor demand. By bringing in third-party capital, the company can smooth out demand fluctuations and ensure that its GPUs are deployed and generating revenue even during periods when direct purchases from cloud providers slow down. It also strengthens Nvidia's ecosystem lock-in: customers who finance through these platforms will likely commit to using Nvidia's hardware and software stack for years.
Historical context: From gaming to AI
Nvidia was founded in 1993 by Jensen Huang, Chris Malachowsky, and Curtis Priem. The company initially focused on graphics processing units for gaming, but over the decades it expanded into high-performance computing, automotive, and data centers. The turning point came in 2012 when researchers at the University of Toronto used Nvidia GPUs to train a deep learning model that won the ImageNet competition, kickstarting the modern AI revolution. Since then, Nvidia has evolved from a niche graphics card maker into the world's most valuable semiconductor company.
The company's CUDA software platform, introduced in 2006, allowed developers to use GPUs for general-purpose computing, not just graphics. This parallel processing capability turned out to be ideal for the matrix calculations that underpin neural networks. As AI models grew larger, so did the demand for Nvidia's hardware. The A100 and H100 data center GPUs became the gold standard for AI training and inference, with companies willing to pay premium prices for these scarce components.
Today, Nvidia's data center business is the primary growth engine. In its most recent fiscal quarter, data center revenue exceeded $30 billion, far outpacing its gaming segment. The company's market capitalization has soared past $3 trillion, making it one of the most valuable companies in the world. However, this growth has also attracted scrutiny from regulators, customers, and investors who worry about the sustainability of AI spending.
The investment case for AI chips
The traditional view of GPUs as quickly depreciating assets stems from their rapid obsolescence in consumer gaming. But in the data center, AI chips are now designed with long lifecycles and are often used for five years or more. They are also fungible: a GPU can be repurposed for different workloads, from training a new model to serving inference requests. This flexibility makes them more like infrastructure assets than disposable components.
Huang's pitch to investors is that AI factories—facilities filled with thousands of GPUs interconnected by high-speed networking—generate revenue by selling compute time. Much like a power plant sells electricity, these factories can sell computing capacity by the hour. This recurring revenue model is attractive to long-term investors such as pension funds and insurance companies, which typically seek stable cash flows.
The financing platforms could take several forms, including debt arrangements, sale-leaseback deals, or structured equity vehicles. Asset managers would underwrite the expected revenue from the AI infrastructure and raise capital from institutional investors. Nvidia would provide the hardware and software stack, while the asset managers would manage the financial engineering. The $500 billion figure represents the potential scale over several years, not an immediate commitment.
Some analysts have drawn parallels to the emergence of data centers as a real estate asset class in the early 2000s. Once considered a technical cost center, data centers are now owned by REITs and investment funds. Nvidia wants to do the same for AI compute, creating a liquid market for chips that goes beyond direct purchases.
Challenges and skepticism
Nvidia's push comes amid a turbulent period for AI stocks. While hyperscalers are spending billions on AI infrastructure, some investors question whether these massive investments will actually bear fruit. Concerns have been raised about the energy requirements of AI data centers, the availability of electricity, and the potential for overcapacity. If AI demand stalls, the value of chips as collateral could plummet.
Additionally, the financing model relies on accurate projections of future compute demand. If AI adoption slows or if competing chips from AMD, Intel, or custom accelerators eat into Nvidia's market share, the revenue generated by these AI factories might fall short of expectations. Lenders and investors could face losses, as they did during the dot-com boom when fiber optic networks were overbuilt.
There are also regulatory considerations. Concentrating AI infrastructure in the hands of a few asset managers could raise antitrust concerns, especially if these partnerships give Nvidia outsized influence over the AI supply chain. Regulators in the U.S. and Europe are already scrutinizing data center deals and cloud market competition.
Despite these risks, Huang remains optimistic. He has frequently compared AI to electricity, arguing that the technology is as transformative as the industrial revolution. In his view, the world's demand for AI compute is insatiable, and the bottleneck is not demand but supply. By unlocking new sources of capital, Nvidia hopes to accelerate the deployment of AI infrastructure and solidify its position as the leading enabler of this new era.
The involvement of established financial institutions lends credibility to the concept. BlackRock, Blackstone, Goldman Sachs, and KKR have deep experience in infrastructure finance and are unlikely to commit to projects without rigorous due diligence. Their participation could also help standardize the asset class, making it easier for other investors to participate in the future.
As the AI trade continues to evolve, the line between technology and finance is blurring. While investors may not be able to buy a physical GPU on the stock exchange, Nvidia's initiative could make chips an indirect investment vehicle. Whether that will stabilize or amplify the cyclical nature of the semiconductor industry remains to be seen. But one thing is clear: Jensen Huang is determined to make AI compute not just a technological marvel, but a financial asset as fundamental as gold or stocks.
Source:MSN News