Vision: Four Decades of Changing the World

We have lived through every major platform revolution of the last four decades. At every step, the pattern has been the same: find great people, work on problems that matter, have fun doing it, and change the world in the process.

The PC Era: A Computer on Every Desk

In the 1980s, the Personal Computer was dismissed as a toy. Mainframe vendors said no serious business would trust a desktop machine. They were wrong.

At Microsoft, Rich Tong was part of the team that proved them wrong. As the fourth product manager on Excel for Windows, he helped push Excel’s market share from 5% to 25% in two years. Later, as Vice President on the Executive Staff for Windows NT, SQL Server, and Exchange, his team grew enterprise revenue from $49M to $1.2B in 18 months. Paul Maritz, as a member of Microsoft’s Office of the President, led all Systems and Tools products including Windows, Windows NT, and Visual Studio — the platforms that put computing power on every desk in every office.

The mission was “a computer on every desk and in every home.” That sounded impossible when Bill Gates first said it. Together with thousands of colleagues, this team put computing power in the hands of billions.

The PC revolution was 10x. A spreadsheet made an analyst 10x more productive than a calculator and ledger paper. The companies that adopted early won their industries.

The Mobile Era: A Computer in Every Pocket

In the 2000s, the smartphone was dismissed as a phone with a bad browser. Carriers said nobody would do real work on a tiny screen. They were wrong.

Satoshi Nakajima, who had designed the architecture for Windows 95 Explorer and Internet Explorer 3.0 and 4.0 at Microsoft, co-founded Ignition Partners to invest in the companies that saw this revolution coming. He developed one of the first photo-sharing applications that debuted in the top 200 iTunes apps and remained the top sharing application for three years. He led UIEvolution to become the only Tier-1 software-only automotive supplier in the world — bringing computing into every car, every pocket.

The mobile revolution was another 10x. Companies like Uber, Instagram, and WhatsApp built entirely new categories of business on top of mobile platforms. Companies that thought “our customers don’t use phones for work” are footnotes.

The Cloud Era: Computing Moves Up

Cloud computing inverted the model. Instead of pushing compute down to every desk and pocket, it pulled compute up into massive data centers. Paul Maritz pioneered this transition — starting Pi Corp to build cloud infrastructure before the term existed, then leading VMware as CEO (creating billions in value), and then leading Pivotal, a leader in hybrid cloud computing.

At Amazon, Apple, Google, Nordstrom, and General Motors, our team saw how technology transforms industries — and learned what works, what fails, and why.

Cloud was another 10x in scale. But notice something important: through all three revolutions — PC, mobile, cloud — the fundamental job stayed the same. Hardware changed. Software evolved. But at the core, you still had to write code. Humans still had to translate business logic into programming languages, debug it, test it, deploy it, and maintain it. The tools got better, but the work was the same work.

Why AI Is Fundamentally Different

AI breaks that pattern.

For the first time, the fundamental act of creating software is changing. You do not write code line by line. You describe what you want, and the machine builds it. You do not debug with breakpoints. You evaluate outputs and refine prompts. The entire development model is different.

We are already seeing up to 10x productivity gains — fewer people per project, faster turnaround, better outputs. And we believe this is still the early innings.

Up to 100x in the Next 12–18 Months

The mechanism is self-learning agents. Today’s AI systems are largely reactive — they respond to prompts. Tomorrow’s systems are already being built to plan, execute, evaluate their own output, and improve. When AI can run its own improvement loop, the compounding rate accelerates dramatically.

Up to 100x is possible in 12–18 months. That means your team of 10 doing the work of 1,000. Not by working harder — by directing agents that work around the clock, in parallel, learning as they go.

If your competitor gets there and you do not, the gap compounds. Every week of delay is not a linear setback. It is an exponential one.

Up to 1000x in the Next 24–48 Months

A team of 10 outperforming a team of 10,000? Seems impossible.

Today, maybe. But think about what changes in a world of autonomous agents that create and learn for themselves. Agents that spawn other agents. Systems that improve their own code, find their own training data, design their own experiments. At that point, the constraint is not compute or headcount — it is the quality of your strategy and the sovereignty of your data.

The organizations that win will be the ones that have been building their AI moat for years — training on proprietary data, capturing institutional knowledge, compounding their advantage. The organizations that were waiting for the technology to “mature” will be a decade behind.

We are not predicting this with certainty. We are saying the trajectory is visible, the mechanism is real, and the time to position is now.

But this power brings a fundamental question being decided right now. The vast majority of the world’s valuable data is private — behind firewalls, inside enterprise systems, never touching the public internet. The publicly searchable web represents less than 5% of all online content¹, and over half of enterprise data is never even analyzed². AI is hungry for data. Will 1 to 3 proprietary model providers and 3 to 4 global hyperscalers control the world’s data? Or will every person and every organization have their own AI?

We believe in the scale-out world. And that belief is what drives our mission.