In a typically frank assessment of the current state of artificial intelligence, Ali Ghodshi, the outspoken CEO of $134 billion software analytics company Databricks, issued a stark warning about the inflated valuations of AI startups that lack fundamental business metrics. Speaking at Fortune Brainstorming AI in San Francisco, Godoshi slammed the tendency of investors to pour capital into unproven companies, saying, “A company that's worth billions of dollars and has zero revenue is clearly a bubble. I don't think it's insane.” Gody declared that he sees “a huge bubble in many parts of the market.”
Godoshi, who has a PhD in computer science, says the atmosphere in the valley is bad. He said even the investors fueling the frenzy recognized the unsustainable nature of the market. In a private conversation, he claimed, the venture capitalist indicated that he was tired of the hype cycle and told him, “Maybe you should take a six-month break and come back. You'd be much better off financially if you did that.”
Godi said he agrees with the criticism of circular finance among many players in the AI space, which artificially inflates the market. Rather than seeing a bubble approaching the point of bursting, Godoshi predicts that the “cyclical aspects” of the situation will get worse before it is corrected. “I think 12 months from now, things are going to be much, much worse.” He added that the current market volatility is actually a healthy signal for CEOs to “take a step back.”
IPO doubts and strategic patience
This skepticism about the current market hype explains Databrix's reluctance to rush toward an initial public offering (IPO), even though Godoshi admitted he was “floating” with the idea. He emphasized that keeping it private for now is a strategic buffer against market volatility. He drew a sharp contrast between Databricks and competitors that rushed to go public during the 2021 boom but faced a tough correction.
“In 2021, most of my colleagues and CEOs felt as if they had achieved an IPO,” Godi added. But in 2022, they suddenly went into cost-cutting mode, while Databricks was able to hire thousands of people. He emphasized that if the bubble bursts, remaining private would allow the company to continue investing in long-term AI utilities rather than reacting to short-term stock price fluctuations.
Real hurdles and market hype
Godoshi argued that while the venture market is heating up, actual enterprise AI adoption is being held back by corporate inertia, not a lack of technology. He cited security concerns and data governance as the main bottlenecks for large organizations.
Databricks, as its name suggests, has many clients who hire it to organize their data, but many of them are over 10 years old, and all of them have strong cyber concerns.
“The big hindrance in this scenario is that you're so worried about getting hacked that you can't actually do anything about it,” Godoshi said.
He said “AI lawyers,” or lawyers specializing in the emerging field of AI law, are currently slowing down their work by scrutinizing regulations and model policies. Additionally, he said the data architecture within most legacy organizations is a “complete mess” after 40 years of piling on software from different vendors, making data siled and difficult to access, leaving Databricks with a lot of work to do.
where is the real value?
Despite warnings about the bubble, Godi remained bullish on certain high-utility AI applications, particularly “AI agents” and “vibe coding.” He revealed some surprising statistics. “For the first time, we found that more than 80% of databases launched on Databricks are launched by AI agents rather than humans.”
He argued that the underlying model layer (technology provided by companies such as OpenAI and Google) is becoming a low-margin product due to increased competition. Rather, the real revenue potential lies in the application layer, where agents perform specific tasks, such as drug discovery in healthcare or automated research in finance.
Mr. Godey advised corporate leaders to cut through the office politics that are holding back such progress. He noted the “battle” among executives fighting to become “AI guys,” and offered candid advice: “Pick one person in your company” to lead strategy, rather than creating a “three-headed monkey” of conflicting leadership.
