August 24, 2026
TechCrunch reported that General Intuition, a startup training foundation models to help AI agents understand space and time, is in talks to raise at a $6 billion pre-money valuation, with new backing from Valor Ventures, Point72 Ventures, and Seven Seven Six. That's a staggering number for a company most business owners have never heard of, and it tells you where the smart money thinks AI is headed next: out of the chat window and into the physical world.
The core idea is simple to state and hard to build. Today's large language models are excellent at manipulating text and code, but they have no real sense of how objects move, how a hallway connects to a room, or how a robot arm should adjust when it misjudges a grip. General Intuition is betting that a model trained specifically on spatial and temporal reasoning becomes the missing layer underneath every robotics and embodied-AI product that comes after it -- something closer to an operating system than an application. If that bet pays off, the winners won't just be robotics companies; they'll be every logistics, manufacturing, and field-service business that eventually inherits cheaper, more capable automation because the underlying model got good enough to trust.
I'd flag the obvious caveat: a $6 billion valuation on a foundation model that hasn't shipped a product most people can point to is a valuation on a thesis, not on revenue. We've seen this movie before in AI funding, and not every high-conviction bet from a marquee investor turns into a durable business. Still, the fact that a firm like Point72 Ventures is willing to write that check says the smart-money consensus has shifted from "can AI write better emails" to "can AI physically act in the world," and that shift matters even if this specific company doesn't end up being the one that wins it.
The second story is closer to home for most businesses. TechCrunch's look inside OpenAI's agent strategy describes a company trying to move AI agents beyond the software engineers and power users who adopted them first, and out toward everyday, non-technical users. That's a much harder problem than it sounds. Coders adopted AI agents quickly because the failure mode is cheap: a bad suggestion gets caught in code review or a test suite. For a marketing coordinator, an HR generalist, or a small-business owner, an agent that books the wrong meeting, emails the wrong client, or misfires on a financial task has real consequences and no safety net.
This is the gap we've been pointing at in our own coverage of the enterprise AI loyalty problem -- companies are willing to trial agents, but trust doesn't transfer automatically from one use case to the next. Building an agent that's genuinely useful for "everyone" means it has to work reliably across dozens of different, messy, non-technical workflows, each with its own tolerance for error. That's an integration and reliability problem as much as a model-capability one, which is why we think the near-term winners will be tools where the agent is scoped tightly to a specific job -- inside a CRM, a helpdesk queue, or a project board -- rather than a general-purpose assistant expected to do everything for everybody. It's part of why we built agent-ready workflow automation around specific business processes rather than a single do-everything chatbot.
Put these two stories together and a pattern emerges. Whether it's a robot learning to navigate a warehouse or a software agent trying to handle a non-technical employee's to-do list, the industry's next fight is over autonomous action, not conversation. That's a bigger ask than anything chatbots have had to prove so far, because acting in the world -- physical or digital -- means owning the consequences of mistakes. Businesses evaluating any of this should ask vendors a blunt question: what happens when the agent gets it wrong, and who is accountable when it does? Teams weighing whether to build custom internal tools versus bolting on a general-purpose agent should treat that question as the real deciding factor, not the demo.
Which of these two futures do you think arrives first for your business: an AI agent you'd trust to act on your behalf inside your existing tools, or a robot that reliably handles physical tasks in your workplace?
Sources