5 startups that caught VCs’ attention at the latest PearX demo day

I spent the last week looking through the latest crop of companies coming out of PearX. If you aren't familiar, this is Pear VC's accelerator program. It has a reputation for being early, aggressive, and usually right about where the money is flowing in the Bay Area. The latest demo day just wrapped, and the signal was loud: the market is tired of thin wrappers. The focus has shifted to the foundation.

For the last eighteen months, the startup world was flooded with people building features on top of OpenAI's API. We saw a thousand variants of "AI for legal" or "AI for marketing." Most of those are dead or dying because they lacked a moat. The VCs I talked to at the event are looking for something harder to replicate. They want hardware, spatial intelligence, and local execution. They want the stuff that is actually hard to build.

The Shift to Local Compute

One of the loudest themes at this demo day was the move away from the cloud. For builders, this is a massive shift in strategy. We have spent a decade being told that the cloud is the only way to scale. But AI has changed the math. The latency and cost of running everything through a central server are becoming bottlenecks for real-world applications.

There is a specific focus now on chips designed for local AI. We are seeing founders build hardware that allows devices to process complex models without pinging a server in Virginia. This isn't just about privacy, though that is a nice side effect. It is about speed and reliability. If you are building a robot or an autonomous system, you cannot wait 500 milliseconds for a cloud handshake. You need the intelligence to live on the edge.

For founders, this means the barrier to entry is getting higher. You can't just be a good coder anymore. You need to understand silicon, thermal management, and low-level optimization. The "move fast and break things" era is being replaced by the "engineer things that actually work" era.

Spatial Intelligence is the New Frontier

Another area catching eyes is spatial modeling. We have spent plenty of time teaching AI to understand text and pixels on a 2D screen. The next frontier is teaching it to understand the three-dimensional world. This isn't about the metaverse or some VR gimmick. It is about giving machines the ability to navigate physical space with the same intuition humans have.

The startups getting attention here are working on spatial models that help AI understand depth, physics, and movement. This is the missing link for widespread robotics. If an AI can't tell the difference between a shadow and a hole in the ground, it isn't useful in a warehouse or a home. The VCs are betting on the teams building the data sets and the architectures that solve this specific problem.

Why Founders Should Care

If you are a founder trying to figure out your next move, look at the complexity of these projects. The low-hanging fruit has been picked. The investors at PearX weren't hunting for another chat interface. They were looking for deep technical breakthroughs. This suggests that the venture market is returning to its roots: funding high-risk, high-reward technical innovation rather than just growth-hacking tools.

• Infrastructure is the moat: If your startup can be replaced by a ChatGPT update, you don't have a business.
• Hardware is no longer a four-letter word: Investors are warming up to physical products because they provide a level of defensibility that software lacks.
• Vertical integration matters: The most successful teams are controlling more of their stack, from the custom chip to the end-user application.

The Skeptics View

I have to stay honest here. Even with all the buzz, we are still in a bubble of sorts. Just because a startup gets VC attention at a demo day doesn't mean they have a viable business model. A lot of these hardware plays require massive amounts of capital before they ever see a dime of revenue. The burn rates will be astronomical.

We also have to ask if the market is ready for this level of local compute. Just because we can build a chip that runs a local LLM doesn't mean every refrigerator needs one. Founders need to be careful not to build solutions for problems that don't exist yet. The excitement at PearX is a leading indicator of where technology is going, but it is not a guarantee of market fit.

The most dangerous thing for a founder right now is to confuse VC hype with customer demand.

What Happens Next

The takeaway for the builder community is clear: get deeper into the stack. If you are a software engineer, start learning how your code interacts with the hardware. If you are a designer, start thinking about 3D environments instead of just flat UI. The next wave of value creation is happening at the intersection of the physical and digital worlds.

We are moving past the era of the AI wrapper. The companies that survive the next three years will be the ones that own their infrastructure, solve hard physics problems, and provide value that can't be replicated with a simple API call. It is a harder path, but it is the only one that leads to a real company.

Read the original at TechCrunch Startups →