SoftBank just made one of the year's boldest AI bets, selling its entire $5.8 billion Nvidia stake to fund its massive OpenAI investment. The move comes as 'Big Short' investor Michael Burry warns AI companies are artificially inflating earnings through questionable chip depreciation practices.
SoftBank just executed one of the most counterintuitive moves in AI investing, dumping its entire Nvidia position at what many consider peak valuations. But CEO Masayoshi Son isn't retreating from AI - he's doubling down in a different direction.
The Japanese conglomerate revealed in Tuesday's earnings that it sold 32.1 million Nvidia shares in October for $5.83 billion, completely exiting its position in the chip giant that's become synonymous with the AI boom. At first glance, the timing looks questionable - Nvidia has been one of the market's biggest winners, with its stock climbing over 180% this year alone.
But sources familiar with the transaction tell a different story. According to a person close to the deal who spoke to CNBC, the sale had nothing to do with concerns about AI valuations or Nvidia's prospects. Instead, SoftBank is redirecting that $5.8 billion windfall directly into its massive $22.5 billion commitment to OpenAI.
The move represents a strategic pivot from betting on AI infrastructure to betting on AI applications. While Nvidia provides the chips that power AI training, OpenAI builds the models that consumers and businesses actually use. It's classic Masayoshi Son - always chasing the next transformative platform.
Timing couldn't be more interesting, given the warnings now echoing from Michael Burry, the investor who famously predicted the 2008 financial crisis. Burry posted on X Monday that major AI companies are "understating depreciation" of AI chips, which "artificially boosts earnings - one of the more common frauds of the modern era." He's promising to reveal "more details" on November 25.
Burry's track record gives his warnings weight. His bet against subprime mortgages before the 2008 crisis made him a household name, and his ability to spot financial engineering ahead of market corrections is legendary. If AI companies are indeed manipulating depreciation schedules to inflate earnings, it could signal trouble for the entire sector.
But SoftBank's move suggests Son isn't buying Burry's thesis - at least not entirely. The company that infamously pumped $18.5 billion into WeWork only to later value it at $2.9 billion has developed a reputation for both spectacular wins and crushing losses. This time, Son seems convinced he's found the winning horse in OpenAI.
The $22.5 billion OpenAI investment represents one of SoftBank's largest single bets ever, putting it in the same league as the company's early Alibaba investment that generated massive returns. The difference is timing - OpenAI is already a proven entity with ChatGPT's explosive growth, not a startup with untested business models.
Market reaction has been mixed. Nvidia shares barely budged on the news, suggesting institutional investors view this as portfolio rebalancing rather than a vote of no confidence. Meanwhile, OpenAI's valuation continues climbing toward $150 billion in secondary markets.
The broader AI investment landscape is watching closely. If Son's bet pays off and OpenAI becomes the Google of AI, selling Nvidia at these levels will look brilliant in hindsight. But if Burry's warnings prove prescient and the AI bubble deflates, SoftBank could find itself holding another overvalued unicorn.
SoftBank's Nvidia exit signals a fascinating inflection point in AI investing. While Burry's warnings about earnings manipulation deserve attention, Son's massive OpenAI bet suggests the real money is moving from AI infrastructure plays to platform plays. Whether this proves genius or folly depends on who's right about AI's next phase - the hardware enablers or the application builders. With $22.5 billion on the line, SoftBank is betting everything on the latter.
