SoftBank's Masayoshi Son Bets Another $290 Billion

On October 28, OpenAI officially announced the completion of its capital structure restructuring. The plot of SoftBank, led by Masayoshi Son, betting approximately $41 billion (about 290 billion RMB) on the future of AI has surfaced.

Not long ago, at a shareholders' meeting in the summer of 2025, Masayoshi Son shook off the haze of investment setbacks in previous years and talked eloquently about his latest ambition: to lead the global "Artificial Superintelligence (ASI)" era in the next decade. The head of SoftBank painted an ambitious blueprint for shareholders - SoftBank is no longer content to be a spectator investor but wants to get directly involved and become an "organizer" in the AI industry. He seemed to have returned to the peak moment when he invested in Alibaba and reaped huge dividends in the Internet era.

During his speech, Masayoshi Son repeatedly mentioned a name: OpenAI. He firmly stated that OpenAI will eventually enter the capital market, and an IPO is just a matter of time. This investment guru has quietly been playing a big "game" regarding AI - the scale of his bet has once again shocked the entire industry.

Restructuring Completed,OpenAI's Equity Undergoes a Major Shake - Up

The core of OpenAI's restructuring this time is to transform the originally non - profit OpenAI Inc into a for - profit entity, while being controlled by a newly established non - profit foundation. After the restructuring, the new for - profit company is named "OpenAI Group PBC" and is still regulated by the OpenAI Foundation.

According to the restructuring plan, the non - profit OpenAI Foundation will obtain 26% of the equity. Based on OpenAI's current valuation of approximately $500 billion, this part of the equity held by the foundation is worth about $130 billion. The foundation has promised to use part of its equity for public - welfare scientific research, such as accelerating breakthroughs in medical and health fields, to fulfill OpenAI's original intention of "benefiting humanity".

It is worth noting that Sam Altman, the co - founder and CEO of OpenAI, does not hold any equity after the restructuring. This unconventional arrangement not only eliminates potential conflicts of interest but also triggers various interpretations of his motives from the outside world.

One of the biggest beneficiaries of OpenAI's restructuring is undoubtedly Microsoft, its long - term financial backer. After months of negotiations, Microsoft and OpenAI reached a new agreement: Microsoft will obtain approximately 27% of the equity. Calculated based on the value of about $135 billion corresponding to 27%, it indirectly confirms OpenAI's astonishing valuation of $500 billion. This means that Microsoft's initial investment of approximately $13.8 billion has increased nearly tenfold on paper.

The new agreement also greatly extends their cooperation - until 2032, Microsoft will continue to have the priority to use OpenAI's latest AI models and products. However, given the particularity of AGI, the agreement adds an "AGI verification mechanism": if OpenAI claims to have achieved AGI in the future, it needs to be objectively verified by an independent expert panel. Once AGI is developed and passes the verification, Microsoft will stop receiving revenue shares from OpenAI. This is regarded as Microsoft's concession to the principle of "AI benefiting the public" and a gesture towards regulatory authorities.

In terms of cloud computing cooperation, Microsoft has voluntarily given up its priority to purchase OpenAI's new cloud business. Previously, Microsoft Azure was OpenAI's exclusive cloud service provider, but now Microsoft has loosened this restriction.

Behind this concession, there is actually a big deal: OpenAI has promised to purchase an additional $250 billion worth of Azure cloud services in the future. OpenAI uses part of its cloud revenue to lock in Microsoft in exchange for more autonomy. With the right to choose cloud services independently, OpenAI immediately reached a sky - high cooperation with Oracle. According to The Wall Street Journal, OpenAI will purchase up to $300 billion worth of cloud computing services from Oracle in installments over five years starting from 2027 for the training and deployment of its AI models. This marks OpenAI's shift from a single - cloud architecture to a "dual - cloud" strategy: Azure and Oracle will jointly support its computing power needs.

On the surface, Microsoft has lost some exclusive rights, but it is not a loser. Azure will still receive huge orders in the next few years, and OpenAI's introduction of Oracle Cloud can also relieve the pressure on Azure's expansion, which can be regarded as a win - win situation in a roundabout way. In addition, Microsoft has specifically clarified that while allowing OpenAI to freely cooperate with third - parties, it will not claim any rights to OpenAI's hardware products.

On the day when the restructuring news was announced, Microsoft's stock price soared 2% at the opening, and its market value returned to the $4 trillion mark. Wall Street analysts generally breathed a sigh of relief - the relationship between OpenAI and Microsoft was full of uncertainties before, and now Microsoft's key layout in the AI landscape has finally been settled.

It is worth mentioning that payment giant PayPal also announced that it will cooperate with OpenAI to launch an AI - driven payment wallet. Stimulated by this positive news, PayPal's stock price soared 10% before the market opened.

"Tacit Agreement and Open Competition"

Let's first understand the profound meaning behind the new agreement between Microsoft and OpenAI.

The restructuring negotiations lasted for a long time, and the focus was on the balance between interest - binding and development autonomy. When Microsoft bet on OpenAI in 2019, it signed strict and self - beneficial terms: exclusive cloud service provider rights, technology exclusivity period, revenue sharing, and a revenue redistribution mechanism once OpenAI achieves AGI. After ChatGPT became popular and OpenAI grew rapidly, these terms were proven to be both Microsoft's moat and a shackle for OpenAI to seek external financing.

It is reported that Sam Altman, the CEO of OpenAI, lobbied Microsoft's senior management several times to modify the cooperation terms so that OpenAI could raise funds and purchase computing power more freely. Microsoft was initially tough, fearing to undermine the guarantee of its early investment returns. However, OpenAI's development momentum made Microsoft not dare to act arbitrarily - if OpenAI was restricted by resources, Microsoft's long - term interests would also be damaged. So, the two sides engaged in a long - term game.

The new agreement this time can be described as "both sides making concessions": Microsoft still locks in important rights and interests until 2032 but also gives up some exclusive rights. More intriguingly, Microsoft has begun to show a little "caution" and a sense of distance towards OpenAI. Since this year, on the one hand, Microsoft has continued to deeply integrate OpenAI models into its core products such as Office 365 Copilot; on the other hand, it has also been developing its own large models and is no longer putting "all its eggs in OpenAI's basket".

Some reports even point out that Microsoft may replace OpenAI's models with those of its competitor Anthropic in some scenarios to avoid over - dependence. This means that Microsoft and OpenAI are entering a new stage of being "both friends and foes": cooperation remains the main theme, but both sides are quietly paving the way for possible future competition.

$30 Billion,Masayoshi Son's Calculated Move

While Microsoft is carefully balancing its own profits and OpenAI's growth, Masayoshi Son has chosen to "go all in" on OpenAI.

SoftBank's investment in OpenAI's restructuring is astonishing - approximately $30 billion, a unprecedented single - investment plan. Even more amazingly, SoftBank's funds are not transferred in one go but are designed in a rather dramatic "staged betting" way: in April this year, SoftBank first promised to invest $10 billion as part of OpenAI's financing, and the remaining $22.5 billion will only be unlocked after OpenAI completes its restructuring. In other words, Masayoshi Son has written an IOU - as long as OpenAI transforms into a for - profit structure by the end of the year, SoftBank will complete the $30 billion investment; otherwise, the total investment will be reduced or even cancelled.

Now that the restructuring has been successfully completed, OpenAI has achieved its goal: SoftBank's previous $10 billion commitment has already been fulfilled, and with the newly approved $22.5 billion "balance" by the board, a total of approximately $30 billion has been officially invested in OpenAI. According to sources, the total scale of this round of financing led by SoftBank is as high as $41 billion. In addition to SoftBank, investment institutions such as Dragoneer and Thrive have invested approximately $11 billion.

What is Masayoshi Son's game plan? First of all, while SoftBank acquires approximately 5 - 10% of OpenAI's shares, it has also won a strategic position deeply bound to OpenAI. As Masayoshi Son said in his speech in June, SoftBank not only wants to invest money but also "get deeply involved in the technical and infrastructure aspects" and participate in OpenAI's full - link layout from R & D to deployment.

As early as the beginning of 2025, Donald Trump announced at the White House that OpenAI would join hands with SoftBank, Microsoft, Oracle, etc. to promote a huge project codenamed "Stargate": a total investment of $500 billion in the next four years to build next - generation AI super - computing infrastructure across the United States. SB Energy, a new - energy company under SoftBank, will be responsible for the construction of part of the power infrastructure in the park. It is not only a core investor but also undertakes the role of infrastructure builder.

OpenAI's current valuation of $500 billion represents the capital market's bet on the prospects of general artificial intelligence. SoftBank's huge investment at this time aims directly at the super IPO dividends in the next few years. Once OpenAI can enter the capital market before achieving AGI or large - scale profitability, SoftBank's book returns will increase several times.

After the restructuring is completed, OpenAI still needs a period of performance and regulatory adjustment, which could be as short as 1 - 2 years or as long as 3 - 5 years, and is expected to meet the IPO conditions. The value of the equity obtained by SoftBank's $30 billion investment at that time is unimaginable. It can be said that OpenAI's restructuring is clearing the way for its future IPO, and SoftBank is seizing the opportunity early, determined to be the big winner at the end.

However, there are also many uncertainties behind this big bet. Masayoshi Son once fell from grace due to his aggressive bets on companies such as WeWork. SoftBank's Vision Fund once suffered huge losses and was forced to downsize to survive. After learning from the pain, Masayoshi Son has resumed the "high - risk, high - return" strategy in the past two years and has placed all his bets on the AI track.

Some people also question the timing of SoftBank's heavy investment in OpenAI when its valuation is already high: why not invest earlier and buy in at a lower price? SoftBank's executives' answer is that they "only make a move when everything is ready". However, the risk of buying at a high price cannot be ignored: if the growth of the AI industry fails to meet expectations and OpenAI's profit cannot keep up with its burning rate, the bubble of its $500 billion valuation may burst first, and SoftBank's huge investment may be in great danger.

Masayoshi Son clearly understands this. Therefore, on the one hand, he is promoting OpenAI's rapid expansion in terms of computing power and territory; on the other hand, he is also actively planning the commercial prospects of AI. In short, for Masayoshi Son, this is a bet that he must and can only win, with no way out.

OpenAI's Double - Edged Paradox

Since its establishment, OpenAI has carried an idealistic color: non - profit, advocating open sharing, and benefiting all of humanity.

However, in just a few years, the script of fate has taken a sharp turn - huge amounts of capital have poured in, closed - source models have emerged, and commercialization has been in full swing. OpenAI has seemingly transformed into a profit - seeking enterprise. This restructuring is a concentrated manifestation of this identity transformation. Although OpenAI has repeatedly emphasized that "the non - profit organization still holds control, and safety and public welfare are still the top priorities", many observers are skeptical.

Some AI ethics organizations are worried that after the restructuring, the profit motive will override the public - welfare mission, and OpenAI may deviate from its original intention of "developing AGI for the benefit of humanity". For example, the GPT - 4 model is closed - source, and details are not disclosed, which has been criticized for being profit - driven and opaque; another example is the high - priced API service, which is accused of going against the original intention of "making AI accessible to the public".

More dramatically, Elon Musk, a co - founder of OpenAI, has also stepped forward to "question". After this maverick billionaire withdrew from OpenAI's board of directors in 2018, he has now sued OpenAI for deviating from its mission and suspected investment fraud. Musk has publicly criticized OpenAI for becoming a vassal of Microsoft and a "money - making machine" and has clashed with Altman on social media several times.

When the "Stargate" project was announced, Musk sarcastically commented that "they actually don't have that much money", implying that OpenAI's $500 billion AI plan is an exaggeration. Altman rarely showed toughness this time and countered that Musk was spreading false information and invited him to go to Texas to see the construction site of the data center in person. The confrontation between these two Silicon Valley leaders reflects the collision between ideal and reality: Musk adheres to the principles of open - source and safety first, while Altman firmly pursues the commercialization path.

Within OpenAI, the senior management also needs to delicately balance the mission and profits. The attitude of the US regulatory authorities is crucial for the smooth passage of the restructuring negotiations. The attorneys general of California and Delaware once jointly intervened in the review and pressured OpenAI to ensure that the non - profit organization has sufficient control and safety supervision rights in the new structure. The final compromise solution is to give the OpenAI Foundation 26% of the equity and special voting rights to retain the "public - welfare moat".

When OpenAI has billions of dollars in financing on its back, investors naturally demand performance and returns. Although safety and ethics need to be considered, can they still be uncompromising in the face of commercial pressure?

Altman himself plays an intriguing role in this: he is both an "idealist" and a shrewd entrepreneur. On the one hand, he still doesn't hold a single share of OpenAI and claims that he doesn't pursue wealth; on the other hand, he actively promotes various commercial cooperations and product launches and is enthusiastic about profit - making models.

This "duality" makes it difficult for the outside world to figure out whether he is sticking to the mission or is full of ambition. In a recent interview, Altman admitted: "Our mission has always been to benefit humanity, but achieving the mission requires resources. I don't think making money and being altruistic are contradictory." However, the oath of "not seeking profit" made back then has ultimately been broken by reality. OpenAI is walking on a tightrope between public welfare and profit, and every step is like walking on thin ice.

Source: Bloomberg News Reporting

The Closed - Loop AI Ecosystem

Behind OpenAI's restructuring, it reflects an "ecosystem approach" being formed in the US AI industry - led by giants, with multiple parties collaborating, clear division of labor, and shared interests.

The biggest feature of the US AI ecosystem is not only technological leadership but also the mutual nesting of capital, hardware, software, and data. This AI boom has evolved into a huge closed - loop where giants invest in each other and are bound together.

At the bottom layer, computing power is the starting point and also a barrier. Chip companies such as NVIDIA and AMD play the role of "energy suppliers", continuously providing GPU computing power for the entire ecosystem. NVIDIA's market value has exceeded $4 trillion, making it the biggest beneficiary of the AI boom. It not only makes money from selling chips but also relies on in - depth cooperation with downstream companies. Companies such as OpenAI, Microsoft, and Google have locked in GPU supply contracts in advance to ensure stable computing power through long - term procurement and even allow NVIDIA to participate in the investment of some projects. This reverse binding has made NVIDIA not just a chip manufacturer but a core capital player. At the same time, AMD is also seeking the same path - jointly developing customized chips with OpenAI and allowing customers to