Masayoshi Son Wants One Hundred Billion Dollars and Nobody Wants to Ask Why

Masayoshi Son Wants One Hundred Billion Dollars and Nobody Wants to Ask Why

SoftBank wants one hundred billion dollars from the Persian Gulf. Masayoshi Son is back on the fundraising trail, chasing sovereign wealth funds and sovereign entities to finance a massive semiconductor and artificial intelligence manufacturing initiative known internally as Project Izanami. The strategy relies on a simple premise: aggregate unprecedented quantities of capital, corner the supply of advanced silicon, and build infrastructure so large that competitors cannot possibly catch up.

Yet the financial press treats these announcements with a strange kind of amnesia. We watch the same cycle repeat every few years. A charismatic founder outlines a trillion-dollar vision, names it after a mythological deity, and invites oil-rich nations to write checks. But the fundamentals of the technology market have shifted underneath SoftBank's feet.

To understand why this latest maneuver feels different, you have to look past the press releases and examine how Vision Fund investments performed over the past decade.

The Anatomy of a Mega-Bet

Masayoshi Son does not think in quarters or fiscal years. He thinks in epochs. When he launched the first Vision Fund, he convinced the Public Investment Fund of Saudi Arabia and Mubadala of Abu Dhabi to commit nearly one hundred billion dollars in a single stroke. The pitch was straightforward. Technology was eating the world, private capital could scale companies faster than public markets, and SoftBank possessed a proprietary radar for winners.

Reality proved considerably messier. We watched WeWork implode from a high-flying real estate play into a cautionary tale of governance failures. We saw Coupang and Didi suffer brutal public market debuts that wiped out billions in paper value. The core thesis—that venture capital could simply subsidize market share until a monopoly emerged—collided with rising interest rates and skeptical public shareholders.

When the music stopped, SoftBank was forced to liquidate massive stakes in Alibaba and T-Mobile just to shore up its balance sheet. The Vision Fund transformed from an engine of aggressive expansion into a defensive holding pen for distressed assets.

Now, Son wants to try the same trick with semiconductors.

The rationale centers on the physical bottleneck of machine learning. Training large language models requires astronomical amounts of compute power. Nvidia currently dominates the market for accelerators, commanding margins that make traditional software companies envious. Son wants to counter this dominance by financing an alternative chip-design ecosystem that can compete directly with Jensen Huang's empire, while securing manufacturing capacity through partnerships with entities like Taiwan Semiconductor Manufacturing Company or Intel.

The problem is not the ambition. The problem is the math.

The Silicon Bottleneck is Not a Financing Problem

Capital abundance often creates the illusion that any physical limitation can be conquered with a sufficiently large checkbook. If you need a fab, you buy land, pour concrete, and import extreme ultraviolet lithography machines from ASML.

Except you cannot simply buy your way past the talent shortage and the energy crisis.

Building advanced semiconductor fabrication plants requires decades of specialized institutional knowledge. It is an industry built on microscopic tolerances, chemical purity, and chemical engineering expertise that exists in very few pockets of the globe. When Intel attempted to scale up its foundry business rapidly, it discovered that owning the machines is only half the battle; operating them at a yield that makes economic sense is an art form passed down through generations of technicians.

Furthermore, artificial intelligence chips consume power at an industrial scale. A single data center cluster designed for frontier model training can draw as much electricity as a small city. Funding the chips without securing long-term, carbon-neutral baseload power sources is like building a fleet of supertankers without checking if there is enough water in the harbor.

Gulf investors are sophisticated participants in this ecosystem. They are not writing blank checks out of blind enthusiasm anymore. They have watched Western venture capitalists pull back, and they have built their own domestic technology investment arms, such as MGX in Abu Dhabi. These funds understand that partnering with SoftBank means taking on massive capital exposure with uncertain exit paths.

If Project Izanami produces chips that cost more to manufacture than Nvidia's established architecture, or if software developers refuse to rewrite their codebases for a new hardware standard, those billions of dollars will evaporate into silicon dust.

The Pivot to Arm and the Stakes of Control

At the center of SoftBank's strategy sits Arm, the British chip designer whose architecture powers nearly every smartphone on earth. SoftBank took Arm public in late 2023, retaining a dominant stake. Arm represents Son's primary leverage point. By utilizing Arm's intellectual property, SoftBank hopes to design custom processors optimized specifically for inference workloads—the phase where AI models actually run in production, rather than training from scratch.

This is where the strategy grows clever, and where it grows dangerous.

Inference is where the real commercial volume will materialize. While training gets the headlines, running queries for millions of enterprise users requires chips that prioritize energy efficiency over raw brute-force computing power. Arm's reduced instruction set architecture is fundamentally suited for this efficiency.

If SoftBank can bundle Arm's designs with dedicated manufacturing financing and sovereign-backed data center deployments, it creates a vertically integrated alternative to the American big tech monopoly. Microsoft, Google, Amazon, and Meta are currently designing their own custom silicon to reduce their reliance on Nvidia. SoftBank wants to position itself as the merchant supplier for everyone else—the sovereign states, the telecommunications giants, and the mid-tier enterprises that cannot afford their own multi-billion-dollar chip design programs.

Yet this vision assumes that Arm's existing ecosystem of licensees will remain comfortable with SoftBank acting as both an independent licensor and a direct competitor in the silicon market. Qualcomm and MediaTek do not like competing with their landlords. If SoftBank pushes too aggressively into finished chip manufacturing or direct enterprise sales, it risks alienating the very companies that make Arm valuable in the first place.

The Reckoning of Patient Capital

Financial markets often mislabel sovereign wealth funds as "patient capital." They assume that because oil revenues span generations, state-backed investors do not care about short-term returns.

That assumption is fading.

Riyadh and Abu Dhabi are aggressively diversifying their economies away from hydrocarbons. Their investments are tied to strict internal benchmarks for job creation, technology transfer, and domestic economic development. A hundred billion dollars sunk into a speculative chip foundry in a foreign jurisdiction does not build local engineering talent, nor does it guarantee that the intellectual property remains in the Gulf.

When Masayoshi Son sits across the table from Gulf sovereign wealth ministers today, he is not offering a guaranteed ticket to the future. He is offering a high-risk gamble on hardware manufacturing at a time when global supply chains are fracturing under the weight of geopolitical tension.

The United States has made it clear that advanced semiconductor technology is a matter of national security. Export controls on high-end chips to China are tightening, and pressure is mounting on Middle Eastern allies to choose sides in the technological cold war between Washington and Beijing. A project financed by Gulf capital but relying on Western intellectual property and Asian manufacturing is a geopolitical Rubik's cube. One misstep by regulators in Washington or Brussels can freeze operations overnight.

We are watching the twilight of the era of cheap, speculative venture capital. The days when a founder could promise infinite growth by burning investor cash to acquire unprofitable market share are gone. If SoftBank secures its one hundred billion dollars, it will not be because the thesis is foolproof. It will be because the alternative—being left behind in the foundational infrastructure of the next century—is a risk sovereign states are even more terrified of taking.

The money will likely move, because the stakes are simply too high to sit on the sidelines. But the accountability will be absolute. There are no more bailouts coming from a roaring public market, and the tolerance for expensive failures has officially run out.

EB

Eli Baker

Eli Baker approaches each story with intellectual curiosity and a commitment to fairness, earning the trust of readers and sources alike.