Decart, the AI startup that stunned the industry last year with a real-time, fully interactive version of Minecraft generated entirely by a neural network, is now weighing exit options that could value the company at up to $5 billion. According to Globes reporting, the company has been in discussions that span both acquisition and IPO scenarios — a remarkable position for a startup that has barely been public-facing for more than a year. For a company operating at the bleeding edge of generative world models, that kind of valuation signals just how hot the market for real-time AI infrastructure has become.
Decart’s core technology isn’t a chatbot or a code assistant — it’s something structurally different. The company built a system capable of generating interactive, dynamic environments in real time using large-scale neural models, with its Oasis demo serving as the most visible proof point. That demo, which let users navigate a fully AI-generated Minecraft-like world at playable frame rates, went viral and repositioned the company overnight from a relatively obscure deep-tech lab to one of the most-watched names in generative AI. The implications go well beyond gaming — AI general capability benchmarks are being redrawn by exactly this kind of architecture, one that models not just language but physical, spatial, and temporal dynamics simultaneously.
What Decart Actually Built — and Why Buyers Are Paying Attention
The Oasis model demonstrated something that major labs had been working toward for years: a generative system fast enough and coherent enough to serve as a real-time interactive environment. Unlike video generation tools that render clips offline, Decart’s approach processes user input and generates the next frame of a simulated world on the fly. The technical lift required to do that at playable latency — without pre-rendering or caching the world state — is substantial, and it represents a category of capability that has obvious applications in simulation, robotics training, autonomous vehicle testing, and entertainment.
That breadth of application is likely what’s driving the $5 billion figure. Acquirers from gaming giants to defense-adjacent simulation firms to cloud infrastructure players would each have legitimate strategic reasons to absorb this kind of capability rather than build it from scratch. The company has been tight-lipped about its revenue and customer base, but the valuation being floated in exit discussions suggests that potential buyers are pricing in future platform value, not just current contracts. In a market where AI agent infrastructure is commanding premium multiples, a foundational world-modeling stack is a credible strategic asset.
Exit Timing, Market Context, and What Comes Next
The timing of these discussions is notable. AI valuations have remained elevated even as broader tech markets have experienced volatility, and several high-profile AI exits — through both acquisition and public offering — have validated the idea that foundational model companies can command outlier prices when their technology is genuinely differentiated. Decart appears to be leaning into that window deliberately, exploring options while buyer appetite remains strong rather than waiting for a later funding round that would dilute existing stakeholders further.
Whether this leads to a clean acquisition by a strategic buyer, a financial sponsor-backed deal, or a longer runway toward an IPO remains unclear. What is clear is that Decart has moved from a viral demo moment to a credible exit process in an unusually compressed timeframe — a trajectory that reflects both the quality of the underlying technology and the broader industry hunger for anything that can model the physical world in real time. The next few months will determine whether the $5 billion figure holds, or whether competitive dynamics and due diligence reshape the terms of whatever deal ultimately closes.
