Cyanogen's Founder Is Back. He Only Wants 50M Windows Users.

In 2015, Kirt McMaster told Forbes his startup was "putting a bullet through Google's head." The startup was Cyanogen. It had raised about $115 million to build a better Android than Google's.

By the end of 2016, Cyanogen had shut down its operating system.

On November 10, 2026, McMaster came back with a new company and a new giant to poke. Kora Systems unveiled a Linux-based desktop operating system built around AI agents, and it's aimed squarely at Windows.

You'd expect the same swagger. Instead, the plan is surprisingly modest. And the two changes he made are the most useful things in the story for anyone going up against a bigger company.

TechCrunch reports that Kora Systems, based in Palo Alto, showed a desktop OS that lets AI agents work across applications, open files and remember context between tasks. It runs on Intel, AMD and Qualcomm Snapdragon machines.

Hardware Busters adds detail. An engine called Cortex spreads AI work across a computer's processors and runs models on the device by default, calling cloud models only for heavier jobs. Agents work in sandboxed workspaces with permissions the user controls. There's also a console-style game mode, a leftover from the company's past life.

That past life matters. Kora used to be Playtron, a gaming OS startup McMaster led. TechCrunch says it rebranded in July 2025 to focus on AI, building on nearly three years of earlier work.

Launch is planned for 2027. The OS will be free, and revenue is supposed to come from subscriptions that bundle AI models and agent services. An enterprise version called Humain OS, built with Saudi AI company Humain, is due on Snapdragon laptops the same year.

Cyanogen grew out of CyanogenMod, a community-built version of Android that dates to 2009. Android Authority traces the company's founding to September 2013, when it took venture money to turn the hobby project into a commercial product.

It got one big break. The OnePlus One shipped with Cyanogen's software in 2014. Then the partnership fell apart in public after one device.

TechCrunch's 2016 post-mortem lists the damage: layoffs, a CEO change, a closed Seattle office, and an end to its OS services by December 31, 2016. The core difficulty was the pitch itself. Handset makers had to give up Google's services for Cyanogen's alternatives, and almost none would.

McMaster's own summary, ten years later, is sharper. Cyanogen still relied on Google Mobile Services, the apps and plumbing that make an Android phone useful.

"If you are trying to compete with an incumbent but are still reliant on them, you have no chance," he told TechCrunch.

Two things stand out.

He's building on ground the rival doesn't own. Kora sits on Linux and Nix, both open source. Microsoft can't switch them off, change the license terms or withhold a key service. The parts Kora thinks are valuable, its AI engine and desktop interface, stay proprietary.

He's refusing to chase everyone. Statcounter data cited by TechCrunch gives Windows about 77% of the desktop market. McMaster isn't claiming he'll take it. He estimates that 45 to 50 million people account for much of today's AI usage and spending, and those are the only users he wants at first.

"Microsoft can keep its billion users," he said.

That's a very different sentence from the bullet quote. It's also a far better go-to-market plan. People who already pay for AI tools feel the limits of their current setup every day. They'll tolerate a rough first version if it gives agents real access to the machine.

Picture two pitches. One says, "There are over a billion desktop users, and if we get 1%..." The other says, "There are about 45 million people who already pay for heavy AI use, here is where they hang out, and here is the one thing they can't do today."

The first sounds bigger. The second is the one you can act on Monday morning.

A beachhead works when four things are true:

• The group has a problem that's painful right now.
• They already spend money trying to solve it.
• They talk to each other, so one win leads to the next.
• You can reach them without a giant ad budget.

Kora's bet is that heavy AI users pass all four. Whether they'll switch operating systems for it is unproven. But the logic is sound, and it's the same logic you'd use for a bookkeeping app or a meal-prep service.

Most first-time founders don't compete with Google. Plenty still build on a platform that could become a competitor, or already is one.

Run this quick audit.

1. List every outside service your product can't run without. App stores, APIs, marketplaces, model providers, ad networks.
2. Mark any that are owned by a company you compete with. Or by one that could copy you in a quarter.
3. For each, ask what happens if the price doubles or access ends. Would you be down for a day, a month, or for good?
4. Find one substitute for the scariest item. You don't have to switch today. You do need to know the exit exists.
5. Decide what you own. Customer relationships, data and brand are hard for a platform to take.

Write the result into your plan next to your target customer. You can do that in a spreadsheet, a Notion page, or a planning tool like Foundra that walks first-time founders through go-to-market and competitive analysis in one workspace.

Some dependence is fine. Nobody builds their own payment network. The danger is depending on the one company that benefits when you fail.

Kora has about 50 engineers. McMaster told TechCrunch that AI coding tools sped up development at least fivefold, mostly in the past six months. CTO Franck De Girolami put it this way in the launch materials: "We believe 50 superpowered engineers can take on Windows."

That's a company claim, and nobody outside has measured it. Still, the direction is hard to argue with. Projects that once needed hundreds of people are being attempted by dozens.

Notice the hiring choice as well. None of Kora's engineers came from Cyanogen. McMaster recruited from the Linux gaming community, people who had already spent years making software run on hardware it wasn't designed for. He hired for the hard part of the product.

Quite a lot, and it's worth being clear about it.

There's no download yet, no named laptop, and no price for the subscription. Funding isn't disclosed either. TechCrunch lists backers including Samsung Next, Square Enix, Alumni Ventures, Polychain, VanEck and Circle, and McMaster says the company can operate for several years.

The market estimates are his own. Reports even differ on the forecast: TechCrunch quotes more than 200 million heavy AI users by 2033, while Hardware Busters cites more than 150 million.

People rarely change operating systems. Their apps, files and habits live there. Kora says it has a layer that runs selected Windows apps on Snapdragon, though "selected" is doing a lot of work in that sentence.

And Microsoft isn't standing still. It's adding AI agents to Windows too.

• Kirt McMaster, who co-founded Cyanogen, unveiled Kora's AI desktop OS on November 10, 2026, with a launch planned for 2027.
• Cyanogen raised about $115M and shut its OS in 2016. McMaster blames its reliance on Google's own services.
• Kora is built on open source foundations that Microsoft doesn't control.
• It targets roughly 45 to 50 million heavy AI users, not every Windows user.
• A good beachhead is a group in pain, already paying, connected to each other and cheap to reach.
• Audit your own dependencies and know your exit from the riskiest one.

What is Kora? A Linux-based desktop operating system from Kora Systems that builds AI models and agents into the OS. It was unveiled November 10, 2026.

Who is Kirt McMaster? He co-founded Cyanogen, which tried to build a commercial alternative to Google's Android, and now leads Kora Systems.

Why did Cyanogen fail? Reports point to a hard sell to phone makers, a public split with OnePlus and leadership conflict. McMaster says the root issue was relying on Google's services.

How will Kora make money? The OS is planned to be free. Revenue would come from subscriptions for AI models and agent services.

What is a beachhead market? It's a small, specific group of customers you win first, before expanding to a wider market.

Is it risky to build on a big platform? It can be. The risk is highest when the platform owner competes with you or could easily start.

1. After challenging Google's Android with Cyanogen, its co-founder takes aim at Windows for AI era, TechCrunch, November 10, 2026
2. KORA Desktop OS Builds AI Into Linux, Heads to Snapdragon X2 Elite PCs in 2027, Hardware Busters, November 10, 2026
3. Cyanogen failed to kill Android, now it is shuttering its services and OS as part of a pivot, TechCrunch, December 24, 2016
4. Cyanogen: where did it go wrong?, Android Authority