How SoftBank’s Investment Model Turned Velocity Into a Strategy

How a Japanese conglomerate applied industrial logic to venture capital – and built a portfolio designed more like a platform than a fund.

Venture capital is usually a human-scale business. Small teams back early founders. Growth is uneven. Strategy shifts. Bets are scattered to survive the flux.

Masayoshi Son, SoftBank’s founder and architect of the Vision Fund, rewrote the model. His thesis was blunt: pick a winner, flood it with capital, and let scale decide the market. The firm became a machine built to execute that belief.

SoftBank raised capital on a scale venture investing hadn’t seen before. Vision Fund 1 pulled in $98.6 billion. Vision Fund 2 followed with $56 billion. That money helped expand the reach of companies now embedded in daily life: Uber, DoorDash, Grab, Paytm, ByteDance, Slack. At one point, Son claimed he raised $45 billion in under an hour—roughly a billion dollars a minute—during a pitch meeting with Saudi investors.

What began as bets turned into infrastructure quietly powering how people move, eat, pay, work, and scroll. The system behind it was one of the most unusual attempts at engineered investing in the history of modern capital.

Capital architecture: Scale as a design input

SoftBank’s key move was structural. It built a mechanism, not just a single winning firm. The first Vision Fund, backed by Saudi Arabia’s PIF and Abu Dhabi’s Mubadala, let SoftBank concentrate capital in late-stage companies and move faster than any rival.

Checks were often oversized and front-loaded. Startups received more money than they could absorb. The goal was speed. Careful execution came later, if at all.

When it worked, the returns were spectacular. DoorDash IPO’d at a $68 billion valuation, giving SoftBank a reported $11 billion paper gain. Coupang, armed with $3 billion in SoftBank capital, built a fully integrated e-commerce and logistics system in Korea. When it listed on the NYSE in 2021, SoftBank’s stake was worth close to $30 billion.

But capital at scale brought exposure. OYO expanded into China and Europe with over $1 billion in backing. When the pandemic hit, its valuation dropped from $10 billion to around $2.7 billion. The model assumed growth would iron out friction. It didn’t always.

This is the core of venture investing: most bets don’t work. If a few do, the model holds. Son once illustrated this with a cartoon goose and a pile of eggs during a shareholder meeting. A few eggs would hatch. One might be gold. The job was to be the goose.

SoftBank treated capital as a lever for momentum. Few other investors tried to direct markets by volume alone.

Platform logic: Beyond portfolio theory

SoftBank didn’t behave like a traditional fund. It wanted to build a network. Uber, Ola, Grab, ByteDance, Slack, Paytm. These weren’t siloed bets. Executives were introduced. Playbooks were shared. A portfolio-wide rhythm emerged.

Son described himself less as a gatekeeper,and more like a conductor. He set tempo and let companies move in coordinated bursts—not uniform, but attuned.

Sometimes that worked. Grab absorbed Uber’s Southeast Asia business. Flipkart, backed after SoftBank’s Snapdeal play failed, beat Amazon to the punch and sold a controlling stake to Walmart. The payoff: a $2.5 billion investment flipped into a $4 billion exit in under a year.

But the system didn’t filter for readiness. Katerra, backed with over $2 billion, promised to industrialize construction. It collapsed in 2021 under the weight of its own complexity. WeWork became SoftBank’s cautionary tale—$14 billion invested, billions written off, and a public collapse that forced Son to take control and restructure the firm.

Geographies as modules

SoftBank went global without building a global operating company. It deployed capital into regional champions and let them scale: Coupang in Korea, Grab in Southeast Asia, Flipkart and Paytm in India, Rappi in Latin America.

These firms began as category leaders and were pushed to expand into infrastructure. Grab added payments. Paytm built a full-stack fintech. Rappi moved into credit, delivery, and dark stores. Capital enabled reach. The model: national champion plus capital equals soft control.

But that surface area came with risk. Didi went public in the U.S. against Chinese regulatory advice. SoftBank’s $12 billion stake cratered. Snapdeal lost to Flipkart. OneWeb filed for bankruptcy. Presence didn’t guarantee leverage.

Still, Son’s operating posture remained: reset, recalibrate, stay in the game. As the Japanese proverb goes—nana korobi ya oki—fall seven times, stand up eight.

Governance after the fact

SoftBank’s early investment cycle favored speed. Son moved quickly, gave founders room, and let systems emerge later. WeWork broke that model.

The failed IPO in 2019 triggered a shift. SoftBank took losses, replaced WeWork’s leadership, and formalized oversight: investment committees, board controls, slower capital deployment. Vision Fund 2, though smaller, operated under stricter review.

The shift wasn’t symbolic. SoftBank moved away from hyper-growth platforms and toward upstream infrastructure—semiconductors, automation, and logistics systems. Fewer bets on apps, more on rails.

In Japanese industry, a defect isn’t just a flaw—it’s a system failure. Son absorbed that lesson. He publicly accepted responsibility for WeWork. He trimmed exposure, slowed pace, and pivoted from what he once called “blitzscaling” to a more deliberate approach.

A system stretched, then reset

The Vision Fund wasn’t elegant. It was forceful, structured, and legible. It showed what happens when venture capital is industrialized—when growth is manufactured, and capital treated as infrastructure.

The second cycle is slower. The thesis is narrower. But the system endures. ARM went public. DoorDash delivered. Alibaba’s $72 billion windfall remains unmatched. Some wins covered the losses. Others still sit unrealized, like ByteDance.

SoftBank remains a rare experiment in financial systems thinking. Not because it avoided failure—but because it kept building through it. Kaizen, the Japanese practice of continuous improvement, doesn’t require perfection. It demands momentum. And SoftBank, for all its misfires and resets, has never run out of that.