Forget AGI…The Prize is Enterprise AGI

**Dave Vellante** (0:00)
Much of the AI industry is chasing the wrong prize. Frontier model vendors like Anthropic and OpenAI, you know, they may have shifted their commercial focus toward enterprise customers, but they've not changed their fundamental architecture. Specifically, they're still trying to concentrate ever more intelligence inside a generalized model. We agree with Ali Ghotzi that the practical definition of AGI has actually been achieved. Now, moving the goalpost to super intelligence or what we've called Messiah AGI in a prior Breaking Analysis does little to create differentiation for enterprise customers.
The real prize as we see it is what we call Enterprise AGI.
Now, what do we mean by that? Specifically, we're talking about intelligence that is unique to and owned by each enterprise. Enterprise AGI is all about harmonizing proprietary data, business processes, policies, and this tacit human knowledge that we talk about a lot on Breaking Analysis, and then turning those artifacts into persistent assets over which models and agents can reason and ultimately act.
The frontier model is definitely an important ingredient in all this equation, but the enterprise system of intelligence, what we call the SOI, we often refer to as its ontology or digital twin, is the linchpin of achieving what we're calling enterprise AGI, and that's where true business value comes from for enterprises. Now, this is distinct from data communism. We've used that term before. Data communism is where everyone consumes essentially the same embedded intelligence. We're advocates of data capitalism, data capitalists, if you will, in which each company controls, governs, and advances its own differentiated intelligence. So welcome to this week's Breaking Analysis. In this episode, George Gilbert and I are going to build on our previous works and further discuss the implications of what we heard at Databricks Data and AI Summit, and put these findings in the context of our framework, and we're going to discuss it in detail, the point of this opening visual here, Alex, if you bring up the first slide. Sam Altman, this is George's wonderful creation, and George, I'm going to ask you to chime in. Sam Altman here is Wile E. Coyote, racing past the enterprise opportunity, and he's off the super intelligence cliff. Meanwhile, Ali Ghotzi and Satya Nadella, they're like the road runner. They're watching from the edge of the cliff and they're smirking, and this joke should send a message to the entire industry, the Databricks, Microsoft, Google, Amazon, Snowflake, and other enterprise SaaS players, who could also be on that cliff. They're competing in a different race, and they're trying to own the system of intelligence layer that connects enterprise data, business meeting, and agent actions. In this Breaking Analysis, George and I are going to define what we mean by enterprise AGI. We're going to look at Databricks' latest moves through the lens of our system of intelligence, that framework that we've been talking about for weeks and months and over a year now. We're going to assess the maturity of the Databricks' genie ontology, which they furthered this past week, and explain how this emerging architecture could reshape the enterprise software market. George, welcome. Great job conceptualizing the opening graphic. Where did you come up with that idea and the LLMs?

**George Gilbert** (3:35)
Yeah. I mean, everyone grows up with Wile E. Coyote, the Road Runner, and the rest of the cartoons.
And this was just meant to capture the idea that racing off for the super intelligence Messiah at EGI is like, at some point, they're going to realize they're way out over their skis. And that the real prize is doing the hard work that, frankly, Palantir pioneered, which is connecting the proprietary data and processes and asset knowledge, and turning that into an asset that agents can reason over and act, and that all these assets become guardrails that allow agents to take actions, but that they're also, they're the orchestration platform that allows agents and humans to achieve collective outcomes. That you can't do that with just siloed models alone.

**Dave Vellante** (4:39)
Well, in our opening, we talked about data communism. So I want to sort of go back to that idea, Alex, if you bring up slide two of George, this is something you introduced to our audience last year. So here we show Sam Altman, Alex Wong, and Karl Marx, and the ideas that the world's best intelligence should be absorbed into ever smarter models and then distributed back to everyone. So this common intelligence layer, well, sounds great and would be very powerful.
Why, in your opinion, that concentrating more intelligence in the model might be the wrong architecture for enterprise AGI?

**George Gilbert** (5:45)
So, okay, let me, it was just to explain where this came from. The germ of the philosophy came out in a comment Alex Wang made on the No Priors Podcast in late 2024, where he said, I think the world's smartest people should contribute their reasoning traces to the frontier models for the benefit of mankind, which I read as Karl Marx saying, from each according to his ability to each according to his needs. And that sounded like data communism. But there's a germ of truth in there, which is when you look at what Alex Wang was referring to, reasoning traces, the first couple generations of these frontier models were trained just on the on internet data.