Will This Be a Triumph or a Disaster?

Compiled by Peilin

Masayoshi Son once again stands on the edge of a cliff, but this time, his bet is not on Alibaba or Arm—it’s on OpenAI.

Over the past year, SoftBank’s total investment in OpenAI has surpassed $60 billion. To fund this AI gamble, SoftBank has sold off some core assets, including stakes in NVIDIA and T-Mobile US, and secured record-breaking loans in an attempt to reestablish itself at the center of global tech investment.

But problems have followed.

OpenAI remains one of the brightest stars in the AI era. ChatGPT has rewritten the imagination of the global tech industry, and Altman is seen as the most charismatic AI entrepreneur of the moment. However, competitors like Anthropic and Google Gemini are closing in fast, and OpenAI faces multiple uncertainties in legal, governance, and business partnerships.

What worries some inside SoftBank even more is that Masayoshi Son seems to be betting more and more of the company’s fate on Altman alone.

This inevitably brings to mind WeWork. Back then, Son was similarly infatuated with a charismatic founder, only to end up with write-downs exceeding $14 billion. Now, the OpenAI story is even grander, and the stakes are even more astonishing.

Is Masayoshi Son replicating the Alibaba miracle this time, or reliving the WeWork nightmare?

A $60 Billion Bet on Altman

Questions about OpenAI have been circulating inside SoftBank for some time.

Last year, when SoftBank began pouring billions into OpenAI, executives asked Son a blunt question: What if OpenAI fails?

The question wasn’t absurd. Even though OpenAI is one of the most central companies in the AI industry, the competition for AI models is far from over. Large models burn money, computing costs are high, commercialization paths are still being validated, and competitors are catching up.

But Son’s stance was crystal clear.

In his view, Altman is leading the most important technological transformation of this century. OpenAI is not just another startup—it’s the key gateway to AGI, or artificial general intelligence.

How strong is this belief?

According to reports, whenever internal concerns were raised, Son would dismiss them brusquely, to the point where subordinates eventually stopped bringing it up.

This is the source of unease inside SoftBank.

They don’t entirely dismiss OpenAI, but they worry that SoftBank is concentrating too much capital into a single private company—one that has been far from stable lately.

OpenAI has weathered governance turmoil, including Altman’s ouster by the board and his swift return; Elon Musk has sued OpenAI, accusing it of straying from its original nonprofit mission; and tensions have flared between OpenAI and partners like Microsoft and Apple.

Critically, SoftBank holds more than 10% of OpenAI’s shares but has no board seat—not even an observer role.

This means SoftBank has placed a massive bet but doesn’t truly hold the steering wheel.

For Masayoshi Son, who is used to being in control, this is a very delicate situation.

His past investments in Alibaba, acquisition of Arm, and operation of telecom businesses all followed a core logic: secure a key position to influence industry trends. But with OpenAI, SoftBank is more like a super financier than a rule-setter.

This also makes some insiders worry: Son’s trust in Altman may be one-sided.

From the Alibaba Miracle to the WeWork Shadow

Son’s willingness to gamble stems from a career built on high-stakes bets.

In 2000, he invested $20 million in Jack Ma, then a relatively unknown entrepreneur. Fourteen years later, Alibaba went public in New York with a valuation exceeding $160 billion, and SoftBank held roughly a third of the shares. That deal made Son a legend.

The massive returns from Alibaba allowed SoftBank to acquire Arm for $32 billion in 2016 and launch the $100 billion Vision Fund in 2017.

At that time, Son was practically the most sought-after figure among global tech entrepreneurs.

But there were also painful lessons after the highs.

WeWork was the most glaring example of a collapse. Son had been extremely impressed by founder Adam Neumann, believing he could reshape office spaces. SoftBank kept doubling down, driving WeWork’s valuation to dizzying heights.

Then the bubble burst. WeWork’s business model, governance structure, and founder issues all surfaced at once, leading SoftBank to write down more than $14 billion.

Later, Son admitted he had ignored too many warning signs.

So naturally, when he once again poured massive funds into a charismatic founder today, tension arose inside SoftBank.

Similarities do exist: both involve grand visions, highly charismatic founders, aggressive expansion, and SoftBank using huge sums to accelerate the narrative.

But the differences are also clear.

WeWork was ultimately a real estate and office-leasing business, while OpenAI stands at the center of the AI technological revolution. It’s not just a business model (business model) innovation—it’s a foundational platform that could reshape software, hardware, data centers, energy, robotics, and even the entire productivity structure.

That’s why Son can’t let go.

In his eyes, missing out on NVIDIA was already a huge regret. SoftBank sold nearly 5% of its NVIDIA stake in 2019 for $3.3 billion, only to watch NVIDIA’s market value soar and become one of the biggest winners in the AI era.

Son doesn’t want to miss out again.

So when OpenAI launched ChatGPT in 2022 and ignited a global AI wave, Son rediscovered his sense of mission. He believed that this time, he couldn’t just be a bystander—he had to be at the center of the action.

The Bigger the AI Gamble, the Harder for SoftBank to Turn Back

The problem is, OpenAI is getting increasingly expensive.

In 2024, SoftBank finally got the chance to invest in OpenAI, starting with $500 million. At the time, OpenAI was valued at roughly $157 billion after the funding round.

By 2025, SoftBank announced a $30 billion investment in OpenAI—its largest-ever investment in a single private company.

In February this year, SoftBank announced an additional $30 billion in three tranches, bringing its total investment to over $60 billion and pushing OpenAI’s valuation to $730 billion.

This means that in less than two years, OpenAI’s valuation has multiplied several times over, and SoftBank has doubled down repeatedly.

On paper, the investment looks brilliant so far. SoftBank’s annual profit, announced last week, more than quadrupled to a record ¥5 trillion, with a significant portion coming from OpenAI’s rising valuation.

But the capital markets aren’t fully convinced.

After hitting an all-time high in October last year, SoftBank’s stock price has fallen more than 20%. S&P Global also downgraded SoftBank’s rating outlook, citing concerns that the massive bet on OpenAI could drain liquidity and affect asset credit quality.

To support its AI investments, SoftBank needs to keep raising funds, selling assets, and forgoing other investment opportunities. The Vision Fund has also laid off staff due to reduced dealmaking.

This has made SoftBank increasingly dependent on OpenAI.

Complicating matters further, OpenAI’s competitive pressures are rising.

Anthropic’s Claude model continues to challenge OpenAI’s lead, with recent reports suggesting its valuation could surge. Google Gemini is also catching up in scientific research and reasoning capabilities. The AI model industry is not a winner-takes-all market—it’s still a rapidly evolving battlefield.

If OpenAI succeeds in going public at a valuation exceeding $1 trillion, Son will have his second legend after Alibaba.

But if OpenAI’s growth slows or the competitive landscape shifts, the pressure on SoftBank will multiply.

Moreover, OpenAI isn’t the only AI direction Son wants to bet on.

He also hopes Arm can participate in the AI chip revolution, breaking NVIDIA’s dominance in AI chips; he’s involved in the “Stargate” initiative to drive AI data center construction across the U.S.; SoftBank has also acquired robotics businesses and laid out data center plans, attempting to extend AI from models to hardware, infrastructure, and the real world.

In other words, Son isn’t just betting on OpenAI—he’s betting on an entire AGI worldview.

This is also where the highest concentration of risk lies.

Because companies can hedge, industries can diversify, but worldviews are hard to hedge.

Conclusion

Masayoshi Son’s investment career has always been an extreme pendulum swing.

He can earn mythical returns on Alibaba but also suffer painful losses on WeWork. He can prove himself amid skepticism but also overlook risks due to overconfidence.

Now, OpenAI has become the ultimate gamble of his career.

For supporters, this is Son’s chance to reclaim his place at the center of the AI era. The AI revolution is just beginning, and OpenAI remains one of the most critical players—not going all-in means no chance of winning.

But for skeptics, SoftBank is concentrating too much capital, assets, and strategic imagination into a single company, a single founder, and a single future narrative.

The real question is: As the AI era enters a more brutal competitive phase, will Son’s bet this time be on the next Alibaba—or a much larger version of WeWork?

Article Source: SoftBank Founder’s Starstruck Bet on OpenAI Raises Concern, Bloomberg

THE END