In a typically candid assessment of the current artificial intelligence landscape, the outspoken CEO of $134 billion software analytics firm Databricks, Ali Ghodsi, issued a stark warning regarding the ballooning valuations of AI startups that lack fundamental business metrics. Speaking at Fortune Brainstorm AI in San Francisco, Ghodsi blasted the trend of investors pouring capital into unproven companies, stating, “Companies that are worth, you know, billions of dollars with zero revenue, that’s clearly a bubble, right, and it’s, like, insane.” Ghodsi clarified that he sees a “huge bubble in many, many portions of the market.”
The vibes in the Valley are bad, in the opinion of Ghodsi, who holds a PhD in computer science. He said that even the investors fueling this frenzy are aware of the unsustainable nature of the market. In private conversations, he claimed, venture capitalists express exhaustion with the hype cycle, telling him, “Maybe I should just go on a break for, like, six months and come back and it’ll be, like, really financially good for me.”
Ghodsi said he agreed with the critique of circular financing among many players in the AI space, artificially inflating the market. Rather than viewing the bubble as near its popping point, Ghodsi predicts the “circular aspect” of the situation will deteriorate before it corrects. “I think like 12 months from now, it’ll be much, much, much worse.” Current market wobbles are actually a healthy signal for CEOs to “take a step back,” he added.
The IPO question and strategic patience
This skeptical view of the current market hype explains Databricks’ reluctance to rush toward an initial public offering (IPO), despite Ghodsi admitting to “flirting” with the idea. He highlighted that staying private at this point offers a strategic buffer against market volatility. He drew a sharp contrast between Databricks and competitors who rushed to go public during the 2021 boom, only to face severe corrections.
“In 2021, most of my peers, CEOs, they were like we got to IPO,” but by 2022, Ghodsi added, they were suddenly in cost-cutting mode, whereas Databricks was able to hired thousands of people. He emphasized that if a bubble does burst, remaining private would allow the company to continue investing in long-term AI utility rather than reacting to short-term stock fluctuations.
Real hurdles vs. market hype
While the venture market overheats, Ghodsi argued that the reality of enterprise AI adoption is being throttled by corporate inertia, rather than a lack of technology. He identified security concerns and data governance as the primary bottlenecks for large organizations.
Databricks, which per its name has many clients that hire it to sort through their data, has many customers 10 years old and older, and they’re all really held back on cyber concerns.
“The big thing holding you back” in that scenario, Ghodsi said, “is that you can’t actually do anything because you’re so worried about getting hacked.”
He said “AI lawyers,” or lawyers specializing in the emerging field of AI law, are now slowing down operations by scrutinizing regulations and model policies. Furthermore, he described the data architecture within most legacy organizations as “an absolute mess” resulting from 40 years of piling on software from different vendors, leaving data siloed and difficult to access—and a lot of work for Databricks to do.
Where the real value lies
Despite his warnings about the bubble, Ghodsi remained bullish on specific, high-utility AI applications, particularly “AI agents” and “vibe coding.” He revealed a surprising statistic: “For the first time we’re seeing over 80% of the databases that are being launched on Databricks are not being launched by humans but by AI agents.”
He argued that the foundation model layer—the technology provided by companies like OpenAI and Google—is becoming a commodity with low margins due to hyper-competitiveness. Instead, the real revenue potential lies in the application layer where agents perform specific work, such as drug discovery in healthcare or automated research in finance.
Ghodsi advised corporate leaders to cut through the internal politics stalling these advancements. Noting the “tussle” between executives fighting to be the “AI person,” he offered blunt advice: “Pick one person for your company” to lead the strategy, rather than creating a “three-headed monkey” of conflicting leadership.
