Cofounder and CEO, Databricks
Databricks CEO: We Don't Need AI To Get Smarter
Ali Ghodsi, Databricks co-founder & CEO, says we have already reached artificial general intelligence, but the AI is just lacking context to be more productive. He argues that is the problem everyone should focus on right now with Caroline Hyde and Ed Ludlow on the sidelines of the Bloomberg Tech event in San Francisco.
--------
Like this video? Subscribe to Bloomberg Technology on YouTube:
/ @bloombergtechnology
Watch the latest full episodes of "Bloomberg Technology" with Caroline Hyde and Ed Ludlow here:
• "Bloomberg Tech" With Caroline Hyde and Ed...
Get the latest in...
About Ali Ghodsi
Ali Ghodsi, cofounder and CEO of Databricks, has been speaking publicly about the company’s recent fundraising, product launches, and his views on the state of artificial intelligence. In July 2026, Ghodsi said on CNBC that Databricks was fundraising at a $188 billion valuation, driven by customer demand for open-source AI models such as Kimi and a resulting shortage of GPUs. He stated that the company’s Unity AI Gateway product, which allows customers to use both open-source and proprietary models, had seen explosive demand. At the company’s Data + AI Summit in June, Ghodsi announced the open-sourcing of a new project called Open Sharing, which he described as a superset of Delta Sharing that enables sharing of data, AI agents, skills, and models. He also noted that over 100,000 people had signed up for the event, making it the largest data and AI conference in the world. Ghodsi has repeatedly argued that artificial general intelligence (AGI) has already been achieved, stating that the current constraint on AI productivity is not intelligence but context. He said that AI models are already smart enough but lack the enterprise context—such as data, processes, and organizational knowledge—to be fully useful in the workplace. He described Databricks’ Lakehouse as a “system of record for agents” and highlighted new products including Customer Lake for marketing and Lake Watch for security. On the topic of an initial public offering, Ghodsi said he believes 2026 is a “terrible year” to go public due to macroeconomic uncertainty and the presence of several large IPOs, and that Databricks, which he said does not burn capital, can afford to wait for more stable conditions. He also noted that the company may raise additional private capital before going public.
Transcript (20 segments)
I
Interviewer0:00
Ali Ghodsi is with us, the CEO and cofounder of Databricks, one of the most valuable private companies that has yet to file confidentially with the SEC. As far as we know. We're gonna hold back on IPO chat for a minute and ask you about whether this moment of all of these companies looking to access capital in the public markets, is that an anxiety for you? Could there be oxygen sucked out the room in some ways for back companies such as yourselves?
A
Ali Ghodsi0:24
Well, we're a private company. Yeah. So private investors, they're soon not gonna have all these companies to deploy their capital on. And so there's gonna be a lot of capital available in private markets. So, no, that's not gonna be a problem at all for us.
I
Interviewer0:35
Ali, I'm really interested in what life's like for you right now as a CEO of a company of Databricks' valuation, but also, like, you know, I don't mean this sort of flippantly. Like, okay. Great. Big private valuation. The company is growing incredibly fast, you know, revenue run rate. Carol set up at the top everything that's happening in software right now. What is it that you are dealing with most often? What is it that you are sort of, I don't know, waking up and going, okay. Wow. This is today.
A
Ali Ghodsi1:04
Yeah. My view is that we're in a very special moment because I'm a contrarian. I believe we actually already have AGI. Artificial general intelligence already arrived. We already have...
I
Interviewer1:12
What's your evidence?
A
Ali Ghodsi1:13
Yeah. I ask groups all the time. I ask this question. I ask, how many of you believe we have AGI? Roughly 10% of the audience usually raises their hands. Then I ask, how many of you think that the AIs that you use, the frontier models, are smarter than most of the people you work with most of the time? And 90% of any crowd, I've done this, like, twenty, thirty times, raise their hand. Okay? So it's already really plenty smart. 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...
I
Interviewer1:44
Databricks. That's what you do.
A
Ali Ghodsi1:45
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? And the problem is, if we just could have that context, we would have huge breakthrough. So that's what Databricks is focused on. You know, that's our only focus is how do we infuse that data context into this product we have called Genie, which then can help answer and automate questions. I think this is the most important question that people should focus on right now.
I
Interviewer2:15
And we've got plenty new products. Like, as you're getting ever higher valuations, you're backing it up with new offerings. I think about Lake Base. I think about some of the M&A that you've been doing in the market to beef up the product offerings. What's hitting home at the moment? What are the customers demanding of you?
A
Ali Ghodsi2:29
Yeah. As you said, Lake Base is, you know, everyone wants these databases. People are not paying attention to databases. Everybody's paying attention to software is being commoditized because AI can produce software. But what they don't think about is that all software needs a database. It's a fact. You cannot have a piece of software, and it's not using a database. And Lake Base is really, really, really tailoring to those agents. So, you know, for instance, Stripe uses Lake Base, and they're using, they put all their KPIs in it. And that way, the agents can actually start answering questions on all the KPIs that they have because, you know, everybody's using agents now.
I
Interviewer3:05
Are they? Is everyone using an agent now? Is adoption, is productivity ready then?
A
Ali Ghodsi3:09
Yeah. Okay. So that depends. So back to that context question, if we could just get the data to the agents. People are using agents, but they're not using it in its full potential. Maybe we've, like, tapped into 1% of it because it's not like autonomous agents are running around, you know, working together, collaborating with humans and, you know, so that's not happening yet. Why is that? Because they don't have the context. So you have to infuse that context into them. So to give you another example, Novo Nordisk. You know, they make Ozempic, you know, big customer that actually leverages Genie, and they're able to infuse a lot of the, all the trials that they run. They infuse it to the AIs, and now you can ask questions from the AIs about, you know, how's that obesity study going? And they were able to shorten the time it takes to understand these studies using Genie and agents from, you know, weeks down to a few minutes. So it is happening, but we're just at the beginning.
I
Interviewer3:59
Case study is real. Right? And I'm not questioning if it is, but then Jensen Huang's equation of more compute, more tokens, more revenues, it means that that should put some on a trajectory to more profitability. Right? In to many people, it's really simple. The other thing that gentleman outlined at the beginning of this week, Computex, yes, was that it's complete nonsense that software is in trouble because his thesis is that the agents themselves become the ultimate users of that software. I've been finding that concept very hard to understand. You probably are in a place where you can evaluate it, Ali.
A
Ali Ghodsi4:37
Yeah. So I think Jensen is spot on with both of those. You know, we actually did the calculation. We think in the next nine months, well, let's say, year round, one year or two years, we're gonna have more software being written than in all history of mankind. You know, no matter how you back into it, this kinda checks out. Okay. With so much software being written, there's gonna be so much demand for all the other things. Like, there's a big ecosystem around software. We happen to do Lake Base, which is a database, and, you know, we're seeing huge demand in these databases, but there are other ones around software. So, yes, this is just continuing.
I
Interviewer5:11
And, therefore, I go back to how we started this conversation. As you keep on scaling, as you keep driving innovation, do you need more private capital? Would you look to go for public capital at this moment?
A
Ali Ghodsi5:21
Yeah. I mean, we are always interested, and we're always talking to investors. And, you know, I'm sure we're gonna do more of that. We will also be a public company. I just think...
I
Interviewer5:28
But why also, why? Like, if the private markets will keep going with you, what's your impetus for going with...
A
Ali Ghodsi5:35
Because, you know, at our scale, you know, I think we raised $20,000,000,000. At these scales, we're almost trying to create a market transaction mechanism for our employees, which, you know, it's 10,000 right now. But if you take all of them, the prior employees, it's probably 14,000 or so. You know? So there is a thing that can do that. It's called the public markets. So I just think this is a terrible year for, you know, to go public. There's so much happening.
I
Interviewer5:58
Super quickly, how closely did you read the SpaceX S-1, and how closely will you track an Anthropic IPO to evaluate the future market that you want to go into?
A
Ali Ghodsi6:09
Yeah. Look. We're not really trying to time the market. We're trying to really win the market in the long run. That's really our focus. So, no, I did not go through that one filing word by word.