Masayoshi son is going big on OpenAI

Shuli Ren, Bloomberg

When Masayoshi Son bets big, it either marks the dawn of an era or the dusk of a tech cycle.

The chairman of SoftBank Group Corp. has enjoyed spectacular wins. Spotting Alibaba Group Holding Ltd. early handed SoftBank $58 billion when the Chinese e-commerce firm went public a decade ago. A $32 billion take-private deal of ARM Holdings Plc in 2016, followed by a timely listing in 2023 when the AI theme took off, gave Son another windfall. ARM now accounts for roughly half of SoftBank’s portfolio value.

But there were breathtaking failures. WeWork Cos.’s decline — a botched listing followed by bankruptcy — cost Son $11.5 billion. The collapse of Greensill brought SoftBank to court. India’s Oyo Hotels & Homes postponed its third IPO attempt after Son demanded stronger earnings. In the two years ending March 2023, his Vision Funds incurred almost $70 billion in losses.

So how should we read the tea leaves now that Son is “all in” on Sam Altman’s OpenAI Inc.? In April, SoftBank led a funding round that valued the unicorn at $300 billion, promising to invest as much as $30 billion by year-end.

Call me cautious, but Son’s latest fascination feels eerily familiar, especially recalling the WeWork days. From his funding pattern to the income-liability mismatch in both firms’ models, there are many parallels.

In both cases, SoftBank was late to the game, forcing Son to shell out billions just to sit at the table. WeWork had been raising money since 2011, but SoftBank did not come in until 2017, after the first $100 billion Vision Fund. Similarly, OpenAI’s record-breaking round this year was late-stage, limiting upside. It also raises concentration risk: OpenAI would be SoftBank’s second-largest asset, after ARM, if Son invests the full amount. By June, he had already put in $9.7 billion.

A sharp jump in valuation in a short time is another common trait. SoftBank first took a WeWork stake in August 2017 at a $21 billion valuation. Just over a year later, it invested another $3 billion, doubling the valuation to $45 billion. Likewise, SoftBank is now buying OpenAI shares in a secondary sale valuing the company at $500 billion, nearly twice what Son was paying six months ago. Is he overpaying?

Both unicorns projected rosy futures without enduring a down cycle. WeWork failed partly because Covid-19 altered office demand. Generative AI could likewise be disrupted by external shocks.

And then there’s the cash burn. In late 2019, fearing WeWork would run out of money, SoftBank, already a one-third stakeholder, had to rescue it. In 2025, concerns persist about OpenAI’s cash flow, with projections of $115 billion burned through 2029. How it plans to pay the $300 billion cloud contract with Oracle Corp. is unclear. That deal spans five years from 2027, when OpenAI projects just $60 billion in sales.

Both startups share fixed expenses but uncertain income. WeWork signed long leases without securing commitments from tenants. OpenAI is bound to massive cloud costs while consumers can cancel subscriptions anytime. That model means VC money must keep flowing, or liquidity dries up.

For now, Son is riding high. SoftBank shares hit a record last week after Oracle disclosed nearly half a trillion dollars of future contracts, with a sizeable share from OpenAI. But is this really good news for SoftBank, and can OpenAI ever achieve positive free cash flow with such spending?

Once again, Son, late to the game, may be staring into a bottomless pit.

[Abridged]

Courtesy Bloomberg/Shuli Ren