Nvidia leans on its balance sheet to keep the AI boom rolling

Nvidia's early dominance in artificial intelligence made it the world's most valuable company, but rivals like AMD and Google have been closing the technology gap. Now the chipmaker is shifting its focus to another advantage: its enormous financial firepower.

After announcing a deal with Wall Street firms to pursue $500 billion in financing for its graphics processing units, Nvidia said it would provide up to $105 billion to support a massive OpenAI data center in Ohio. The move acts as a safety net for OpenAI if the ChatGPT maker's fortunes decline.

Nvidia's free cash flow has jumped 18-fold in three years, reaching $48.5 billion in the latest quarter. The company has also increased its dividend from a penny to 25 cents per share and launched an $80 billion stock buyback plan. It holds $30.2 billion in marketable equity securities, up from $12.9 billion a year earlier.

Recent investments include $30 billion in OpenAI and $1.5 billion in SB Energy, a SoftBank affiliate building the Ohio facility. Nvidia is also backing about 4 gigawatts of development at the site, with data centers expected to open between 2028 and 2030.

CEO Jensen Huang said frontier AI labs have huge demand for computing power but often lack the long-term credit profiles needed to secure infrastructure independently. A week earlier, Huang announced that GPUs are becoming a new asset class, with firms like Goldman Sachs, Apollo Global Management, Blackstone and BlackRock able to invest in them like real estate.

Analysts at Cantor dismissed concerns that Nvidia is effectively buying revenue, calling the latest deal a "clear signal that the current AI investment cycle will be elongated and durable." They added it creates "additional competitive moats" for Nvidia.

Competition is intensifying - Google began recognizing revenue from TPU system sales in the second quarter, and AMD reported over 100% growth in its data center business. Paul Meeks of Freedom Capital Markets said stepped-up competition eats into Nvidia's "outrageous margins," pushing the company to broaden its strategy.

AI bulls argue Nvidia is simply responding to demand, noting capacity remains the bottleneck. Anthropic's annualized revenue run rate hit $65 billion in July, up sevenfold from a year earlier, while OpenAI's reached $40 billion. Matthew Vegari of Clearwater Analytics said the "house of cards" narrative is "somewhat misguided," adding, "We might one day be at overcapacity. But that day isn't today."

Source: CNBC