Venture capitalist Chamath Palihapitiya has raised $135 million for 8090, an AI software startup he founded in early 2024, and is now running the company as CEO, according to TechCrunch.

8090 is built on a simple idea: use AI to rebuild business software that most companies already use, but make it cheaper and simpler. The company's name reflects its goal — build software with 80% of the features at 90% of the cost. Here is the insight: big companies that use enterprise software (the programs that manage payroll, sales, inventory, and finances) often pay for features they never actually use. 8090 bets it can strip away the unused complexity and sell a leaner version for much less, per Forbes.

Building software with AI is expensive. 8090 expects to spend more than $10 million per year just on the cost of running its AI systems — a figure reported by Business Insider in March 2026. That is a huge cost for a startup. As the company grows and sells to more customers, it needs to make sure those AI costs do not eat into the savings it promises customers.

A Familiar Pattern, New Tools

The idea of beating established software makers with a simpler, cheaper alternative is not new. Salesforce did it to an older system called Siebel; newer companies have carved out corners of the market from giants like SAP and Oracle. The difference now is how fast AI can make it happen. Instead of hiring hundreds of engineers over years, AI code generation tools can help a small team build software faster. A project that might once have taken 200 people five years could, in theory, be done by a much smaller group.

Palihapitiya is running 8090 himself rather than hiring a separate CEO. That sends a signal: he personally believes in this bet, not just investing from the sidelines.

The $135 million is a big bet on this idea. It comes at a time when AI coding tools are attracting serious money — tools like GitHub Copilot and others that help developers write code faster. But 8090 is different: it is not selling a tool to coders. It is selling finished software directly to businesses.

Here is a real concern: the claim that 80% of features is enough has not been tested in the real world at large companies, especially in industries with strict rules. Banks, hospitals, and government agencies need software that tracks every action and proves compliance with regulations. The missing 20% of features might contain exactly what those companies need. Only when 8090 tries to sell to these businesses will it know whether the math really works.

Another thing to watch: if the cost of running AI keeps falling — because the technology improves and companies compete on price — then 8090's high costs today might not be a problem tomorrow. But if the company builds its entire business around today's AI costs, it could hit problems if those costs do not fall as expected.

For years, Palihapitiya has argued publicly that enterprise software is bloated and overpriced. With 8090, he is putting his money and his time where his mouth is.