Vantora, the startup builder formerly known as UP.Labs, has raised $100 million from Silversmith Capital Partners and changed its operating model around a central enterprise premise: some of the most valuable AI systems cannot be commercialized broadly.
The company builds new ventures alongside corporate partners, which invest in the resulting startups and become their first customers. Its new approach gives those partners an option to bring a venture into their own business rather than leave it as an independent company selling to the wider market.
Founder and CEO John Kuolt describes the model as a proprietary M&A pipeline. The practical effect is that Vantora can pursue products that a large customer considers too strategically sensitive to share with competitors.
What changed
UP.Labs previously built companies intended to address a corporate partner’s problem while also serving an external market. That is a familiar venture-building logic: a committed design partner helps validate demand, then the startup seeks a broader customer base.
Vantora is narrowing that mandate. It will continue to work with corporate customers, including Porsche, Alaska Airlines, J.B. Hunt, Wabash and TDG, the parent company of Ashley Furniture. But it is increasingly building startups specifically for those customers, with a path for the customer to acquire or absorb the business.
The $100 million investment is Vantora’s first outside funding, according to the company. It remains separate from venture firm Up.Partners, despite sharing office space and an earlier association.
Why physical AI fits this model
The company’s focus is physical AI: AI software and intelligence layers tied to machinery, fleets, industrial hardware and real-world workflows. These deployments can be harder to standardize than office software because the data, equipment configuration, operating procedures and safety requirements are specific to a company.
For an industrial enterprise retrofitting hardware for autonomous operation, the strategic value may lie less in buying a generic tool than in controlling the models, integrations and operational knowledge built around its assets. That concern is especially acute when the resulting capability could affect productivity, maintenance, routing, manufacturing processes or other competitive operations.
Vantora argues that its previous model caused it to abandon some of those opportunities because corporate partners would not allow a solution to be sold to rivals. Kuolt cited an AI concept developed with J.B. Hunt that was not pursued under the old approach because it could not be taken to the broader market.
The new structure is intended to turn that restriction into the business case.
The trade-off for enterprise buyers
For corporate partners, a purpose-built venture may offer a faster route than developing everything internally while preserving more control than a conventional software procurement relationship. The startup can recruit a dedicated team, operate with more autonomy than an internal program, and begin with a committed customer and domain access.
But ownership changes the economics and the burden. A company that intends to fold the resulting startup into its operations must be ready to govern the technology, retain the talent, manage cybersecurity and data rights, and support the product after acquisition. It also gives up some of the cost-sharing and product-market learning that comes from a vendor serving many customers.
For Vantora, the model similarly trades the possibility of a large horizontal software company for a more concentrated, customer-specific outcome. Its ability to identify problems that are large enough to justify forming a company — but sufficiently bounded for eventual integration — will be the core test.
What to watch next
The important signal will be execution, not the rename or financing alone. Watch for the first ventures built under the proprietary model, whether corporate partners exercise their acquisition option, and how Vantora defines IP ownership and operating control before a deal closes.
More broadly, Vantora’s move reflects a growing distinction in enterprise AI: some tools will remain broadly sold software, while others — particularly systems connected to physical assets and proprietary operating data — may be built as strategic capabilities that companies prefer to own outright.
