Databricks co-founder Matei Zaharia wins ACM's top award, says AGI has already arrived

Matei Zaharia, co-founder of Databricks and creator of Apache Spark, one of the most influential computing frameworks in history, has received the highest award from the Association for Computing Machinery (ACM). On accepting the prize, he made a statement that would sound like marketing narrative coming from most people, but not from him: AGI is already here, and people just misunderstand it.

Zaharia is not a startup founder making bold predictions for hype. In 2009, he created Apache Spark as part of his doctoral dissertation at UC Berkeley. The distributed data processing framework quickly became an industry standard: today Amazon, Microsoft, Netflix, Uber and thousands of other companies around the world use it. It is for this contribution to computer science that the ACM presented him with its main computing award, one often compared to a Nobel Prize in data science.

In 2013, Zaharia and colleagues from Berkeley founded Databricks, which turned Spark into a commercial enterprise platform and grew to a valuation of $62 billion. Today Databricks competes with AWS, Google and Snowflake for the enterprise data and ML infrastructure market. The company is also investing actively in open language models, as shown by its DBRX series and its investment in Mistral.

But his view of the current state of AI is more interesting than the awards. In Zaharia's opinion, the main mistake in most AGI discussions lies in how the question is framed. People wait for an "AGI moment" as a single event: a model crosses some threshold, and everything changes. But that is the wrong frame. AGI is not a switch but a process. Modern language models and agentic systems already outperform humans at writing code, analyzing complex documents and synthesizing scientific papers. The question is not whether "we have reached AGI" but how to use what already exists.

Zaharia is now focused on applying AI to scientific research. He considers the AI for Science field, which covers accelerating drug development, automating laboratory experiments and analyzing genomic data, the most important practical use of large language models. In his words, the most promising use of LLMs is not chatbots and assistants but research agents: systems that can read thousands of scientific papers, build hypotheses and propose concrete experiments to test them.

Zaharia's ACM win is not just an honor for the creator of Spark. It is a reminder that the current AI revolution stands on the shoulders of the previous one, the distributed computing revolution. Spark gave the industry infrastructure for processing data at scale; now that infrastructure trains and runs language models. When the person who built that foundation says "AGI is already here," it deserves more attention than yet another debate about terminology.