Masayoshi Son does not do small bets. The man who turned a $20 million investment in Alibaba into $60 billion just secured a $40 billion bridge loan to pour even more money into OpenAI. Not million. Not a rounding error. Forty. Billion. Dollars.
SoftBank Group announced the loan on Friday, March 27, describing it as fuel for "investments in ChatGPT-maker OpenAI and for general corporate purposes." But let's be real — this is not about "general corporate purposes." This is Masa Son strapping himself to a rocket labeled "Artificial Superintelligence" and lighting the fuse.
The loan comes on top of SoftBank's existing $30 billion commitment to OpenAI's record-shattering $110 billion funding round — itself the largest private funding round in history. Combined with previous investments, SoftBank's total exposure to OpenAI is now approaching a figure that makes venture capitalists' eyes water and risk analysts' blood pressure spike.
SoftBank shares promptly dropped on the news. Investors are nervous. The question on everyone's mind: is this the smartest bet in tech history, or the most expensive lesson since WeWork?
The Numbers: A Masterclass in Leverage
Let's break down the math, because the numbers are genuinely staggering.
SoftBank's OpenAI investment timeline:
- Late 2024: Initial OpenAI investment as part of a broader AI pivot
- January 2025: Co-founded The Stargate Project with Oracle, MGX, and OpenAI — a joint venture to build AI infrastructure across the US
- December 2025: Raced to close a $22.5 billion funding commitment to OpenAI before year-end, reportedly tapping margin loans against its Arm Holdings stake
- February 2026: Booked a $4.2 billion unrealized gain on its OpenAI position
- February 2026: Announced additional follow-on investments, with Son declaring OpenAI a "clear leader" in AI
- March 27, 2026: Secured $40 billion bridge loan from a consortium of top US banks
To finance this AI obsession, Son has been liquidating everything that isn't nailed down:
- Sold entire NVIDIA stake — $5.83 billion, gone
- Dumped $12.73 billion in T-Mobile stock between June and December 2025
- Leveraged Arm Holdings shares as collateral for margin loans
- Trimmed various Vision Fund positions across the portfolio
The man is selling the furniture to buy more chips at the casino. Except the casino is artificial intelligence, and Son believes the house always wins — as long as you are the house.
The Strategy: "Total Offense"
SoftBank's internal language tells you everything you need to know. After years of what Son called a "defensive" posture — licking wounds from the Vision Fund's spectacular WeWork-era losses — the company has pivoted to what it calls "total offense."
And they mean it literally.
Son's AI portfolio now reads like someone asked "What if one person tried to own the entire AI stack?":
- Arm Holdings — The chip architecture company powering most of the world's mobile devices and increasingly its data centers. SoftBank owns a massive stake.
- Graphcore — UK-based AI chip company, acquired to compete with NVIDIA in specialized AI silicon.
- Ampere Computing — ARM-based server chips for cloud and AI workloads.
- Wayve — Self-driving car startup, because autonomous vehicles are just AI agents on wheels.
- Intel — A $2 billion stake, betting on the American chip giant's turnaround.
- OpenAI — The crown jewel. ChatGPT. GPT-5. The company that arguably kicked off the entire AI era.
- The Stargate Project — A joint venture with Oracle and OpenAI to build out AI infrastructure across the United States. The initial commitment? $100 billion over four years, with plans to create 100,000 jobs.
This is not a portfolio. It is a vertically integrated AI empire. Chips (Arm, Graphcore, Ampere), models (OpenAI), infrastructure (Stargate), and applications (Wayve). Son wants to own every layer of the AI stack from silicon to superintelligence.
Time Magazine ran a profile this week titled "Masayoshi Son Is Betting It All on American AI." That is not an exaggeration. It might even be an understatement.
The OpenAI IPO Theory
TechCrunch published an analysis hours after the loan announcement with a provocative headline: "Why SoftBank's New $40B Loan Points to a 2026 OpenAI IPO."
The logic is straightforward. Bridge loans are, by nature, short-term financing instruments. They are designed to "bridge" the gap until permanent financing (or a liquidity event) arrives. You do not take out a $40 billion bridge loan unless you expect a massive payday relatively soon.
That payday? An OpenAI IPO.
OpenAI's $110 billion funding round valued it as one of the most valuable private companies in history. But "private" is the key word. SoftBank's investment is currently illiquid — it cannot sell shares on the open market. A bridge loan lets SoftBank service its debt obligations now, with the expectation that an OpenAI IPO will generate enough liquidity to repay the loan and then some.
If OpenAI goes public in late 2026 at a valuation north of $200 billion (which analysts increasingly consider plausible), SoftBank's position could be worth $60-80 billion — turning the $40 billion loan into one of the most profitable trades in financial history.
If OpenAI doesn't go public? Well, that is when things get interesting. And by "interesting," I mean "potentially catastrophic for SoftBank's balance sheet."
The Bear Case: Why Investors Are Nervous
SoftBank shares retreated on the news, and it is not hard to see why.
The debt load is enormous. SoftBank is already one of the most leveraged companies in Japanese corporate history. Adding $40 billion in bridge financing on top of existing obligations puts the company's debt at levels that make bond traders uncomfortable.
OpenAI is still unprofitable. Despite generating billions in revenue from ChatGPT and API services, OpenAI burns through cash at an extraordinary rate. The company's massive compute costs, talent wars (they are planning to double headcount to 8,000 by end of 2026), and research expenses mean profitability remains elusive.
Bridge loans are expensive. Short-term debt typically carries higher interest rates than long-term financing. SoftBank is paying a premium for urgency.
Concentration risk is real. When your entire strategy depends on one company (OpenAI) in one sector (AI) delivering astronomical returns, you are one black swan event away from disaster. What if a competitor leapfrogs ChatGPT? What if regulation throttles AI deployment? What if the AI bubble deflates?
We have seen this movie before. Son's Vision Fund era produced legendary losses: WeWork ($11.5 billion writedown), Katerra (bankrupt), Greensill (collapsed), and dozens of other startups that turned from unicorns into cautionary tales. The Vision Fund lost $32 billion in a single fiscal year. Son promised he'd learned his lesson. Has he?
The Bull Case: Why Son Might Be Right
But here is the thing about Masa Son — when he is right, he is spectacularly right.
His $20 million bet on Alibaba in 2000? It turned into $60 billion, making it one of the most successful venture investments ever made. His early bet on Yahoo Japan printed money for decades. His acquisition of ARM in 2016 was mocked at the time and is now looking prescient as the AI chip market explodes.
Son doesn't think in quarters. He thinks in decades. His "300-year plan" for SoftBank is the stuff of legend (and occasional mockery). But the man has a pattern: make outsized bets on transformative technologies, endure short-term pain, and collect generational wealth on the other side.
AI might be the biggest technological shift since the internet. If Son is right about that — and increasing evidence suggests he is — then being the largest individual investor in the company that leads the AI revolution is not reckless. It is rational.
OpenAI's moat is deepening. ChatGPT's user base is massive and growing. Enterprise adoption is accelerating. The API business creates ecosystem lock-in. And OpenAI's talent bench is arguably the deepest in AI research.
The Stargate Project gives SoftBank infrastructure leverage. Even if OpenAI stumbles, SoftBank's investment in AI infrastructure through Stargate creates value that persists regardless of which model wins.
🔥 Our Hot Take: This Is Son's Alibaba 2.0
I have watched Masa Son operate for years. The pattern is unmistakable.
In 2000, everyone thought he was crazy for giving $20 million to a little-known Chinese e-commerce company run by an English teacher named Jack Ma. The dot-com bubble had just burst. Tech investing was toxic. Son's timing looked terrible.
Twenty years later, that bet made SoftBank one of the most valuable companies in Japan.
In 2026, people think he is crazy for borrowing $40 billion to go deeper on OpenAI. AI valuations look stretched. The debt load looks dangerous. The timing feels reckless.
But here is what the skeptics keep getting wrong about Son: he doesn't care about the next quarter. He cares about the next quarter-century.
Is $40 billion a lot of money? Absolutely. Is it a lot of money relative to the potential value of being the largest investor in the company that builds artificial general intelligence? That is the question Son is answering.
And his answer is: it is a bargain.
The difference between 2000 and 2026? The stakes are higher, the check is bigger, and the technology is more transformative. If AI delivers even a fraction of what its proponents promise, OpenAI could be worth $1 trillion within five years. Son's position would be worth more than most countries' GDP.
If it doesn't? SoftBank becomes the most expensive lesson in corporate history.
That is the bet. Forty billion dollars says Masa Son is right. Again.
The Bottom Line
SoftBank's $40 billion bridge loan is more than a financial transaction. It is a statement of conviction — a declaration that Masayoshi Son believes artificial intelligence will reshape the global economy, and that OpenAI will be the company that leads that transformation.
The markets are nervous. The debt is enormous. The risks are real.
But Masa Son has been here before. Laughed at, doubted, called reckless. And then proven spectacularly right.
The AI era's biggest bet just got $40 billion bigger. Whether it is genius or madness, we'll know within two years.
Place your bets accordingly.