Ali Ghodsi on AGI
From Databricks CEO: We Don't Need AI To Get Smarter · · Bloomberg Technology
“I believe we actually already have AGI. Artificial general intelligence already arrived. We already have... We don't need AI to get smarter. It just is lacking context. If we could capture all the context of all the conversations we're having and everything that's happening in the processes... If you could feed that to the AI, they already could be extremely productive. But no one's really focused on that problem. Everybody's focused on superintelligence, and can we, like, continue the scaling laws?”
On , Ali Ghodsi, Cofounder and CEO at Databricks, spoke about AGI during Databricks CEO: We Don't Need AI To Get Smarter on Bloomberg Technology.
About Ali Ghodsi
Ali Ghodsi, cofounder and CEO of Databricks, has appeared at several events in mid-2026 discussing the company's strategy, fundraising, and views on artificial intelligence. In June, he addressed Databricks' recent $5 billion equity raise at a $134 billion valuation, stating that the company is seeing "acceleration in all of our business" and that he does not rule out raising additional private capital before a potential IPO. He described 2026 as "a terrible year" to go public due to macroeconomic uncertainty and the scale of other large IPOs, but said Databricks will eventually become a public company to provide liquidity for its approximately 14,000 current and former employees. Ghodsi emphasized that Databricks is free cash flow positive and does not need to burn capital, allowing the company to choose its own timing for an IPO. Ghodsi has repeatedly argued that artificial intelligence has reached the level of artificial general intelligence but lacks context, not intelligence. He stated, "We don't need AI to get smarter. It just is lacking context," and described the problem as "how do we feed it that context? And that context is in the data." He promoted Databricks' "Lakehouse" as a system of record for AI agents and highlighted the product "Genie," a conversational AI tool that uses an ontology graph to answer quantitative questions. At the RSA Conference in April, Ghodsi introduced the "open security lakehouse" architectural pattern, advocating for organizations to store data in open formats in their own cloud accounts to avoid vendor lock-in. He also noted that open-source and Chinese AI models are "absolutely dominating" and that the lifespan of frontier AI models has shortened to less than a quarter.