Matej Zaharia, co-founder and CTO of Databricks, almost missed the email announcing that he was the winner of the 2026 ACM Computing Award. “Yeah, that was a surprise,” he told TechCrunch.
Back in 2009, Databricks brought technology that Zaharia developed during his PhD at the University of California, Berkeley under the guidance of renowned Professor Ion Stoica.
Zaharia has created a way to dramatically speed up the results of slow and unwieldy big data projects and released it as an open source project called Spark. Big data back then was what AI is today, and Spark got the tech industry’s ear. Zaharia, 28, has become a tech celebrity.
Since then, he has led the engineering at Databricks, growing the company into a cloud storage giant and now the data foundation for AI and agents. Along the way, the company raised more than $20 billion (at a $134 billion valuation) and had a revenue run rate of $5.4 billion. Silicon Valley dream.
On Wednesday, the Association for Computing Machinery presented him with an award for his collaborative contributions. The award comes with a cash prize of $250,000, which he plans to donate to a yet-to-be-determined charity.
In addition to his CTO duties, Zaharia is also an associate professor at the University of California, Berkeley, and he’s looking forward, not backward. Like others in the Valley, he sees a future filled with AI.
“AGI is already here, but not in the way we appreciate it,” he told TechCrunch. “I think the bigger point of this issue is that we should stop applying human standards to these AI models.”
tech crunch event
San Francisco, California
|
October 13-15, 2026
For example, a person can pass the bar exam to become a lawyer only if he has integrated a huge amount of knowledge. However, AI can easily ingest vast amounts of facts. Even if you answer a knowledge question correctly, it does not equate to general knowledge.
This tendency to treat AI like humans can have significant negative impacts. He gives the example of the popular AI agent OpenClaw.
“On the one hand, this is great. You can do so many things with it. It just does them for you,” he said. But it’s also a “security nightmare.” That’s because it’s designed to imitate a human assistant that you trust with things like passwords. This leads to the risk of your browser being logged in and being hacked, or your agent fraudulently draining your bank of money.
“Yeah, there’s a little bit of humanity in it,” he says.
As a professor and product engineer, Zaharia is most excited about how AI can help automate research in everything from biological experiments to data editing.
He believes that in the same way that vibecoding made prototyping and programming accessible to everyone, research using accurate, hallucination-free AI will one day become ubiquitous.
“There aren’t a lot of people who need to build applications, but there are a lot of people who need to understand information,” he said. Ultimately, leveraging the strengths of AI can improve its capabilities. That means telling him what every rattle in a car means, scanning text and images to include radio and microwaves, and, as he now sees his students doing, simulating changes at the molecular level and predicting their effects.
“What I’m most excited about is so-called AI for search, but especially AI for research and engineering,” he said.
