SoftBank Is Gambling Its Future On OpenAI

[00:00:03][music] >> Over the past three decades, SoftBank has become the largest venture capital investor in the world. Along the way, its founder, Masayoshi Son, has gained a reputation as both a visionary technology investor and a reckless gambler, depending on who you ask. But both descriptions can be true at the same time. The same instincts that led him to Alibaba also led him to WeWork.

[00:00:27]Now, with OpenAI, Son is making the biggest gamble of his career. SoftBank has poured tens of billions of dollars into the company, funded largely with borrowed money. And Son has staked not just his fortune, but his legacy on the bet paying off. If he's right, he'll be remembered as the man who financed the most important technology in human history. If he's wrong, SoftBank itself could be in jeopardy. In this video, we'll go over how SoftBank got here and why its bounce sheet is far riskier than it appears.

[00:00:56]Before we move on, I want to take a quick break to tell you about today's sponsor, Templafy. Templafy is an AI PowerPoint creation tool, which is trusted by millions of professionals at Fortune 500 companies, including KPMG, BDO, and Adobe. But the crazy thing is they just released a completely free version. Not a free trial, it's actually free with no usage or other limits.

[00:01:17]Here's how it works. Let's ask Templafy to create a PowerPoint presentation about SoftBank's investments into OpenAI.

[00:01:24]Before it starts making the presentation, it asks you a bunch of questions about what you want it to cover. Thus, the final product will be highly customized to what you need, which minimizes any edits you have to make after the fact. It then gives you an outline of what the presentation will look like. Once it looks good, just press the create button. Within minutes, you can download a real PowerPoint file, which you can edit and send. If you have to make PowerPoints for any reason, I can't stress how much time Templafy will save you. It handles all the tedious grunt work of formatting, which gives you more time to be creative. I I to stress again that it's completely free with no credit card required. So, there's no risk in checking it out by clicking the link in the description below.

[00:02:04]SoftBank was founded in 1981 by the Japanese entrepreneur Masayoshi Son.

[00:02:09]While Son was born in Japan, he is ethnically Korean and went to college in the United States. SoftBank started out as a retailer of computer software. This was back when software was sold on physical disks. Despite its name, SoftBank is not a bank. The name means software bank. Basically, a place where you go to buy software. Eventually, SoftBank expanded to become a telecommunications company offering cell phone and broadband internet service to consumers. Today, they are one of the largest telecom companies in Japan. You can think of them as basically the AT&T of Japan. In addition to its core telecommunications business, SoftBank also has many venture investments. This is what they're best known for internationally. In 2000, they made a $20 million investment into a little-known Chinese company called Alibaba. Alibaba eventually became the largest e-commerce company in China. The value of the stake increased to tens of billions of dollars. SoftBank eventually became two separate entities. The core telecom business in Japan and international venture investment firm.

[00:03:10]This split became formalized in 2018 when SoftBank spun off its telecom business into a separate entity called SoftBank Corp. The parent company is called the SoftBank Group, which trades under the ticker symbol 9984 on the Tokyo Stock Exchange. The telecom business is called SoftBank Corp, which trades under the ticker symbol 9434.

[00:03:30]The SoftBank Group still owns 40% of SoftBank Corp and retains voting control.

[00:03:36]In addition to making venture investments with its own balance sheet, the SoftBank Group also manages external money. In 2017, they launched their so-called Vision Fund. The Vision Fund was capitalized with $100 billion.

[00:03:49]About 1/3 of this money came from SoftBank itself, with the remainder coming from outside investors. The single biggest outside investor was a Saudi Arabian sovereign wealth fund. In 2019, the Vision Fund suffered huge losses from its investment in the co-working startup WeWork. In that year, WeWork failed to conduct a planned IPO, and its valuation plummeted. Subsequent to the WeWork debacle, SoftBank tried to raise Vision Fund 2, but it attracted very few external investors. Vision Fund 2 is capitalized almost entirely by SoftBank's own money. This chart shows SoftBank's net asset value, or NAV, over time. In 2014, Alibaba went public. At the time, Alibaba represented about half of SoftBank's NAV, with SoftBank Corp representing the other half. In 2016, SoftBank acquired a semiconductor company called Arm. The dark blue bars represent SoftBank's share of the two Vision Funds. In 2021 and 2022, SoftBank sold all of its Alibaba shares. In 2023, Arm went public, and its valuation skyrocketed.

[00:04:53]In 2024 and 2025, SoftBank invested a total of $34.6 billion into OpenAI in exchange for an equity stake of approximately 10%. This is all included within Vision Fund 2. To fund this investment, SoftBank sold other Vision Fund investments, including stakes in Nvidia. They also sold all their shares in T-Mobile, which they acquired back in 2013. In March of 2026, OpenAI announced a $122 billion private funding round at a post-money valuation of $852 billion.

[00:05:24]While the headline number is $122 billion, not all of that money was paid immediately. SoftBank contributed $30 billion to this funding round. If SoftBank follows through, this would bring their equity stake up to 13%.

[00:05:37]While SoftBank is doubling its investment, its ownership percentage only increases marginally because the valuation is now so high. The $30 billion is split up into three equal tranches that are to be paid in April, July, and October of 2026.

[00:05:52]The reason they split it up into tranches is because SoftBank didn't have 30 billion dollars of cash on hand. They needed time to secure financing. In late March of 2026, SoftBank announced that it had secured a 40 billion dollar bridge loan. This is more than enough to fund all three tranches, but this is a relatively short-term loan maturing in March of 2027. As the name suggests, bridge loans are supposed to be temporary in nature. A company takes on a bridge loan if it needs money fast to conduct a transaction. SoftBank has to secure longer-term financing to repay the bridge loan.

[00:06:26]Let's look at the state of SoftBank as of July 2026. These numbers are pro forma for the 30 billion dollar investment into OpenAI, which is scheduled to be completed this October.

[00:06:36]Keep in mind that there may be slight errors in the numbers due to rounding and currency conversions. The telecom company SoftBank Corp is valued at 68 billion dollars on the Tokyo Stock Exchange. The SoftBank Group's 40% stake in SoftBank Corp is worth 27 billion dollars. The SoftBank Group owns 88% of Arm. This stake is currently valued at 250 billion dollars. They own 13% of OpenAI. In the most recent funding round, OpenAI was valued at 850 billion dollars. So, SoftBank's 13% stake is worth 110 billion dollars. Excluding OpenAI, they also have a bunch of other investments in the two Vision Funds.

[00:07:15]These are collectively worth 49 billion dollars. In total, SoftBank's equity investments are worth 436 billion dollars. Against this, they have 146 billion dollars of net debt. That gives them a debt-to-equity of 33%.

[00:07:30]SoftBank's net asset value is 290 billion dollars. This is significantly higher than their market capitalization of 190 billion dollars. So, the company trades at a significant discount to the value of its assets.

[00:07:43]SoftBank follows a self-imposed policy that loan-to-value should be below 25% in normal times and never exceed 35%.

[00:07:51]The borrowing for the most recent OpenAI investment will push the leverage ratio up to 33% at the high end of the acceptable range. Remember that a large portion of this debt is the $40 billion bridge loan that needs to be refinanced.

[00:08:04]In April 2026, it is reported that SoftBank was seeking a $10 billion margin loan secured by its equity stake in OpenAI. SoftBank would use the proceeds to pay back some of the bridge loan. The margin loan would be non-recourse. In the event that SoftBank defaults, lenders could repossess the OpenAI shares, but they couldn't sue SoftBank itself for the money. In other words, if OpenAI goes bankrupt, the losses would be eaten by the lenders, not SoftBank. SoftBank's 25% leverage target only includes loans that have recourse to SoftBank itself. Thus, replacing the bridge loan with a non-recourse margin loan would decrease SoftBank's leverage ratio. The banks denied SoftBank's request for a $10 billion margin loan. OpenAI is not publicly traded, making its shares difficult to value. While OpenAI is technically worth $850 billion, most of the money they've raised is from complicated circular financing arrangements, so the true value of the company is difficult to ascertain. In June of 2026, SoftBank decreased its margin loan request from $10 billion to $6 billion, but this was still rejected by the banks. In July 2026, it was reported that SoftBank has renewed talks with banks for the $10 billion loan, but this time the proposed loan will have recourse to SoftBank itself. So, it's not really a margin loan. It's a loan to SoftBank. This puts SoftBank in a precarious financial position. They own a $110 billion stake in OpenAI, but banks are not willing to lend money against these shares, rendering them effectively worthless for the purposes of leverage. Banks are much more likely to offer a margin loan against publicly traded stock as opposed to private company stock. The value of a publicly traded company is much more transparent and the shares can be liquidated much more quickly. It was previously expected that OpenAI would IPO by the end of 2026, but in June, The New York Times reported that it will likely delay its IPO until 2027.

[00:09:55]OpenAI's investment banks advised an IPO this year would fail to achieve a $1 trillion valuation.

[00:10:01]Sam Altman has said that any value less than $1 trillion is a non-starter.

[00:10:07]In mid-June of 2026, Elon Musk's SpaceX IPO'd. SpaceX is technically an AI company after Musk merged XAI into it.

[00:10:16]SpaceX was perhaps the most anticipated IPO in history. SpaceX IPO'd at $135, giving the company a market cap of $1.75 trillion.

[00:10:27]In the first 2 days, retail traders pumped the price up to $200, giving a peak valuation of $2.6 trillion. But since then, the price has free-fallen to $120, significantly below the IPO price.

[00:10:40]In August, the first lockup will expire and SpaceX employees will finally be able to dump their own shares. This could send the share price even lower.

[00:10:48]This result was completely predictable.

[00:10:51]Shortly before the IPO, we published a video explaining why the company is extremely overvalued and likely to disappoint. So, what does any of this have to do with OpenAI?

[00:11:02]In a lot of ways, OpenAI's proposed IPO is very similar to SpaceX. A massively hyped-up company whose valuation is completely disconnected from its financial performance. If anything, OpenAI is even worse than SpaceX. At least SpaceX's Starlink business is profitable. No part of OpenAI is anywhere close to profitable. Investors just got burned badly on the SpaceX IPO.

[00:11:25]They'll think twice before buying into OpenAI. When the investment banks advised Altman they wouldn't achieve a $1 trillion valuation, They specifically cited the lackluster performance of the SpaceX IPO. Since then, SpaceX's share price has declined much further. The probability of OpenAI conducting its IPO this year seems very low indeed. This puts SoftBank in a bad position.

[00:11:47]Masayoshi Son is gambling his company's future on OpenAI.

[00:11:52]In May of 2026, Bloomberg published a lengthy exposé about SoftBank's investments into OpenAI. Multiple executives at SoftBank told Bloomberg that they are concerned about the company's massive bet on OpenAI. But Masayoshi Son is starstruck by Sam Altman and refuses to listen to any criticism. Son first met Sam Altman in 2017, back when OpenAI was still pretending to be a nonprofit. Altman told Son a bunch of fairy tales about artificial general intelligence. Altman probably talked about how his AI will cure cancer or whatever. Son was captivated and immediately wanted to invest. Once OpenAI transitioned to a for-profit company, SoftBank happily pumped tens of billions of dollars into the company. Now, OpenAI is burning billions of dollars and there's no sign of the AGI Altman promised. Even if you believe in the future of AI, this doesn't necessarily mean that OpenAI will be a winner. On the enterprise side, they're losing money to Anthropic, which recently surpassed them in terms of revenue. On the consumer side, they're losing market share to Google's Gemini chatbot. In June of 2026, SoftBank held its annual general meeting of shareholders. Masayoshi Son explained that he is no longer motivated by money.

[00:13:02]He instead wants to establish his legacy by developing artificial superintelligence or ASI. He believes that OpenAI will achieve this. As the main financial backer of OpenAI, he will be remembered as the man who enabled this world-changing technology. He originally planned to retire in his 60s.

[00:13:19]He's now 68 years old, but he revised his timeline. Now, he doesn't want to retire for another 10 years.

[00:13:27]So, what's the worst-case scenario for SoftBank? They have $436 billion against $146 billion of debt. Of their assets, $110 billion is their stake in OpenAI.

[00:13:39]In a worst-case scenario, OpenAI fails to conduct an IPO either this year or next year. Eventually, the company runs out of money, goes bankrupt, or its stock falls to zero. In which case, the value of SoftBank's assets would decline to $326 billion, of which $250 billion is its 88% stake in Arm. Arm is a semiconductor company. Historically, they've not manufactured or even designed their own semiconductors. They instead develop CPU architectures. You can think of the architecture as kind of like the blueprints that control how the CPU functions. They license this architecture to other companies that design and manufacture the CPUs. The main selling point of Arm architecture is its power efficiency. Historically, their main end markets have been smartphones and other mobile devices where battery life is a key consideration.

[00:14:29]They also license their architecture for CPUs used in AI data centers. So, they have indeed benefited from the AI data center buildout. SoftBank acquired Arm in 2016 for $31 billion. In 2023, they IPO'd Arm on the Nasdaq at a valuation of $54 billion. Since then, Arm's share price has skyrocketed, and the company is now valued at almost $300 billion. In the most recent fiscal year, Arm generated $5 billion of revenue, an increase of 23% versus the prior year, and they reported a net profit of $900 million. The company is valued at 58 times revenue and more than 300 times its earnings. 23% revenue growth is pretty good, but it doesn't come anywhere close to justifying this huge valuation.

[00:15:16]So, why is Arm so overvalued? Prior to 2023, Arm was a fully owned subsidiary of SoftBank. During the IPO, they only sold 10% of their shares. So, SoftBank owned 90% of the company. Since then, SoftBank's stake has been diluted a little bit as Arm paid stock-based compensation to its employees. Today, SoftBank owns 88% of Arm's outstanding shares. That means that only 12% of the shares are available to the general public. If only a tiny sliver of equity is circulating in the public markets, the trading price ceases to be an accurate reflection of the entire firm's underlying value. When a typical stock becomes overvalued, there is usually a sufficient supply of shares for sellers to bring the price back down to reality.

[00:15:57]But, because of Arm's tiny float, this correction mechanism is broken. There aren't enough shares to satisfy the demand, and borrowing shares to short the stock is often too expensive. This means that even a small amount of buying pressure can skyrocket the share price to levels that would be impossible to maintain if more shares were actually trading. This is particularly true for a company like Arm, which is at the heart of the AI hype cycle. Retail and momentum traders are fighting over a tiny sliver of equity. The crazy price they set gets extrapolated to the 88% of the company that SoftBank still owns. On paper, SoftBank's Arm shares are worth $250 billion. But, if they tried to sell them, the free float would increase, and the trading price would likely collapse.

[00:16:40]On paper, SoftBank's current leverage ratio of 33% doesn't look that bad. But, the vast majority of its assets consist of Arm and OpenAI, both of which have questionable valuations. It's not too difficult to imagine a scenario where SoftBank goes bankrupt. It's certainly possible that OpenAI goes bankrupt within the next few years. Arm is a real company with consistent profits. It's never going to go bankrupt. But, if SoftBank starts fire selling its shares and market sentiment changes, its stock price could easily fall by 60, 70, or even 80%. Arm currently trades at more than 50 times revenue. An 80% decline would put it at 10 times revenue, which is still a pretty rich valuation. If OpenAI goes bankrupt and ARM declines significantly, SoftBank could become insolvent. I'm not saying this is a likely scenario, but it's certainly possible. And this is why SoftBank Zone executives are growing worried by Masayoshi Son's perceived recklessness.

[00:17:37]Masayoshi Son is a gambler by nature. If he wasn't willing to take big risks, he wouldn't have gotten to the position he's in today. He made his original fortune by investing $20 million in Alibaba back in 2000. While this sounds like a no-brainer in hindsight, it was a significant risk at the time. Alibaba was a very new startup that would probably fail. And $20 million was a lot of money for SoftBank at the time. They couldn't afford to make very many failed investments of this size. More recently, SoftBank has had its share of failures.

[00:18:08]Most notably was their $14 billion investment to WeWork. This went to zero when WeWork declared bankruptcy in 2023.

[00:18:15]So far, SoftBank's wins have outweighed its losses, so their net asset value and share price have increased over time.

[00:18:22]But as Masayoshi Son's war chest has grown, so too has his ambition. He knows he's getting old and doesn't have much time left. He wants to end his career with one last spectacular gamble.

[00:18:34]All right, guys. That wraps it up for this video. What do you think about SoftBank? Let us know in the comments section below. As always, thank you so much for watching and we'll see you in the next one. Wall Street Millennial signing out.