Vantora raises $100M+ to build physical AI startups industrial companies can own

Why it matters

Vantora is wiring customer data, startup formation and a potential acquisition into one process, giving industrial companies a route to own AI built around their most sensitive operations.

Vantora, the venture builder formerly known as UP.Labs, announced on September 16th that it raised more than $100 million from Silversmith Capital Partners to build AI companies around the operational problems of large industrial customers.

Founder and CEO John Kuolt (@Kuolt) started UP.Labs in 2022 after building corporate ventures at BCG Digital Ventures. His path into venture building ran through management information systems at Santa Clara University, an executive MBA at Oxford and an earlier mobile software startup, IQzone. The thread running through that career is a belief that large corporations possess useful data and expensive problems, while startup founders possess the incentives and speed to turn those assets into products.

The rebrand formalizes Kuolt's sharper version of that thesis. Vantora increasingly plans to build companies specifically for its corporate partners, which invest at formation, become the first customers and receive the option to absorb the resulting startup or technology. Kuolt described the structure to TechCrunch as a "proprietary M&A pipeline."

That acquisition option changes which projects Vantora can pursue. Under UP.Labs' earlier model, a startup was generally expected to sell beyond its original corporate partner. Kuolt told TechCrunch that some of the most valuable ideas were abandoned because a customer would never permit technology built from its operations and data to be sold to competitors. Vantora can now pursue those sensitive projects with ownership for the customer built into the arrangement.

A venture studio with the buyer already inside

Vantora embeds founders, product leaders and AI engineers inside a corporate partner to work with its operators, facilities and operational data. The teams identify a costly problem, build a standalone venture around it and test the product with the partner as the anchor customer.

Kuolt's insight is straightforward: industrial corporations do not lack problems worth solving. They struggle to recruit startup-grade teams around narrow internal workflows, then give those teams enough independence and economic upside to move quickly. Vantora uses equity to align the outside founders with the corporate customer while avoiding the usual pilot process, where a software vendor arrives with a general product and spends months adapting it to legacy systems.

Vantora calls the ownership concept "Sovereign AI." The label is convenient marketing, though the underlying issue is concrete. A manufacturer automating machinery, an airline rebuilding maintenance planning or a freight carrier training agents on proprietary workflows may treat the resulting intelligence as part of its core operating advantage. Those customers have little reason to fund a product that will later be sold across their industry.

Vantora's model gives them another route: build the technology with an independent founding team, establish its value in live operations and retain a path to bring it in-house. The corporate partner gets a custom AI business without forcing the project through an internal innovation group. The founder starts with data, distribution and a paying customer, though the potential market may be bounded by the partner that helped create it.

The projects Vantora accepts typically represent a claimed $50 million to $100 million in potential annual EBITDA contribution for the partner, according to the investment announcement. That range is Vantora's estimate, rather than a disclosed record of customer results.

From Porsche vehicle data to freight operations

Porsche became UP.Labs' first corporate partner in 2022, with an original plan to create six companies over three years. The first was Pull Systems, a Los Angeles startup created to analyze electric-vehicle data, including battery and power-electronics performance. A second Porsche venture, Sensigo, applied AI to vehicle diagnostics and repair workflows.

The initial Porsche structure expected the resulting products to reach a broader market. Vantora's new approach accommodates systems that a partner considers too strategically valuable to share.

That issue surfaced in logistics. J.B. Hunt and UP.Labs established a venture lab in October 2024 to create as many as six companies focused on freight operations. Kuolt told TechCrunch that Vantora had previously passed on one J.B. Hunt project after the carrier considered it too proprietary for outside commercialization.

The partnership has since produced Overroute, an AI freight startup developed within J.B. Hunt's operations. J.B. Hunt said this month that Overroute automates manual load-execution work, giving Vantora a current example of software trained around the workflows of an anchor customer rather than a generic enterprise AI pitch.

Vantora also names Alaska Airlines, Wabash and TDG, the parent of Ashley Furniture, as partners. TechCrunch reported that Vantora is working with additional industrial manufacturing and oil and gas customers.

Silversmith backs the ownership model

Vantora says the Silversmith financing is its first outside investment and that it was profitable before the deal. The financing announcement did not specify the amount above $100 million, Vantora's valuation or Silversmith's ownership percentage.

The capital will fund additional corporate partnerships, hiring in AI and commercial roles, and development of COSMOS, Vantora's proprietary data ontology product. Vantora says COSMOS organizes a customer's operating data for use by AI processes and autonomous agents. That product could become an important source of repeatability for a business whose basic service remains labor-intensive and customized for each partner.

Silversmith is also taking governance positions. Todd MacLean, Danielle Waldman and Annie Cory will join Vantora's board. The growth equity firm manages more than $5 billion, according to its announcement, and is making a sizable first institutional bet on a founder who has spent years refining the mechanics of corporate venture creation.

Vantora says it has launched 17 ventures and expects to reach 20 by the end of 2026. It also reports 79% year-over-year revenue growth, without publishing revenue or annual recurring revenue. Those figures show throughput, though Vantora's longer-term performance will depend on whether the ventures create measurable operating gains and whether corporate partners continue funding or acquiring them.

Kuolt has moved the acquisition question to the beginning of the startup-building process. That makes Vantora a different proposition from a conventional venture studio chasing independent companies and broad markets. Its founders enter with a customer, privileged data and a possible buyer already at the table. The trade is equally clear: some of these startups are being built to become indispensable to one industrial operator, rather than universal software companies.