Vantora Raises $100M+ to Build Physical AI Startups

Most startup studios try to build companies that can sell to everyone. Vantora is increasingly building companies that may only ever need one customer.

Vantora, formerly known as UP.Labs, has secured more than $100 million from Silversmith Capital Partners in its first outside financing.

The company works with large industrial businesses to identify operational problems, build dedicated startups around those problems and give the corporate partner the option to eventually bring those ventures fully in-house.

It's an unusual model.

And Vantora thinks physical AI makes it much more valuable.

The startup studio is becoming an M&A pipeline

Traditional venture studios usually create businesses that can eventually serve a broad external market.

Vantora is moving in another direction.

Founder and CEO John Kuolt described the strategy as a proprietary M&A pipeline.

The corporate partner becomes an investor and initial customer.

If the technology solves an important enough internal problem, the company can potentially acquire the startup instead of allowing it to sell the same capability to competitors.

That makes sense for industrial AI.

A manufacturer may want autonomous systems built specifically around its factories.

A logistics company may want AI trained around proprietary fleet data.

An airline may want automation tied deeply to internal operations.

Those companies may have little interest in helping a supplier package the same advantage for competitors.

Physical AI makes proprietary data more valuable

The generative AI boom was built largely around internet-scale information.

Physical AI is different.

Factories, warehouses, transportation networks and industrial equipment generate highly specialized operational data.

That information is often proprietary.

And it may be difficult for an outside startup to understand without working directly inside the enterprise.

Vantora embeds startup teams with corporate partners and combines internal operators, data and domain knowledge.

Its own description of the model says it focuses on problems that can represent roughly $50 million to $100 million in annual EBITDA impact for a partner.

That is a very different startup thesis from building another horizontal AI app.

The company is going deeper into physical AI

Vantora says its new strategy increasingly targets areas where AI interacts with real machines and physical operations.

Autonomy.

Industrial equipment.

Manufacturing.

Logistics.

Energy.

Transportation.

One example Kuolt gave involves a Fortune 100-style industrial company that may need to retrofit machinery for autonomy but does not want that intelligence layer controlled by an outside vendor serving competitors.

That creates an opportunity for a company whose job is not simply to sell software.

Its job is to build an asset the customer can eventually own.

Silversmith is funding a very different type of venture platform

The investment is Vantora's first outside capital since launching in 2022.

Silversmith says the money will support new corporate partnerships, hiring across AI and commercial roles and development of Vantora's proprietary data platform, COSMOS.

COSMOS is designed to organize operating data so AI systems and autonomous agents can use it.

That is important because enterprise AI increasingly runs into a data problem.

The model may be capable.

The company's data is fragmented across old systems.

Connecting those two worlds becomes valuable.

Corporate innovation has historically struggled

Large companies regularly launch internal innovation programs.

Many struggle to produce meaningful new businesses.

The organization may move slowly.

Employees have different incentives.

Projects can get trapped between experimentation and actual deployment.

Startup studios promise a different structure.

Build with startup speed.

Use enterprise knowledge.

Give the business a real customer from day one.

Vantora's new approach goes further by offering the enterprise a path to own the successful company.

The funding reflects a wider physical-AI boom

Robotics, industrial automation and autonomous systems have attracted increasing venture interest as AI capabilities move beyond text and software.

Investors are funding:

robotics data,

industrial intelligence,

autonomous machines,

sensors,

manufacturing software,

and physical infrastructure.

Vantora is effectively betting that large industrial corporations already contain hundreds of valuable physical-AI problems waiting to be turned into companies.

What happens next?

The key question is whether this model creates independent startups with venture-scale economics — or something closer to sophisticated internal product development.

Vantora appears comfortable blurring that line.

For its corporate partners, the goal may not be creating the next public software company.

It may simply be creating a proprietary technology asset worth far more than the cost of building it.

That could make physical AI one of the rare startup categories where the best exit is decided before the company is even created.