The collar looked like a good idea. No buckle, nothing fiddly, just slip it onto the cow. According to a Halter engineering account published by Ubiquity Ventures, it fell off in about five minutes. The team adopted a conventional buckle. For a venture firm devoted to advanced technology, this is a wonderfully useful detail: sometimes the part that needs inventing is smaller than you think.
- Ubiquity backs technical founders putting software into the physical world.
- It leads or co-leads pre-seed and seed rounds with $1 million to $3 million checks.
- Its free founder tools cover the less glamorous machinery: cash, boards and customers.
01 / A farm is a software problem
Ubiquity Ventures, founded by Sunil Nagaraj in 2017, calls its territory “software beyond the screen.” Halter makes that phrase tangible. Its collars and software let farmers manage livestock through virtual fencing. Ubiquity met founder Craig Piggott in a Palo Alto coffee shop in 2018 and co-led the seed round a few days later. The company announced a $220 million financing at a $2 billion valuation in March 2026.
The useful connection is between a physical task and a product that can keep improving. A sensor observes something. Software interprets it. A device acts. Performance feeds the next adjustment. Ubiquity’s thesis explicitly favors that last step: technology that changes what happens outside the computer. A dashboard may describe a problem beautifully while leaving somebody else to solve it.
Measure the result. Improve the software. Repeat.
02 / The rocket inside the argument
Nagaraj’s earlier work at Bessemer Venture Partners helped shape this view. In a 2021 essay, he described Rocket Lab’s software-driven electric turbopumps as central to his investment reasoning. Changing code offered a different development rhythm from repeatedly rebuilding mechanical parts. Rocket Lab belongs to his Bessemer record; it is easy to blur a partner’s career with a younger firm’s portfolio.
Ubiquity’s own investments carry the argument into other settings. WindBorne builds weather balloons. Eyebot offers automated vision-testing kiosks with doctor-written prescriptions. Loft Orbital supplies space infrastructure so customers can deploy payloads and software without building the whole satellite operation themselves. These businesses serve different buyers, but each places software inside work with physical constraints.
03 / Strange ideas, ordinary arithmetic
There is a financial boundary to this enthusiasm. Ubiquity’s $75 million third fund, announced in January 2024, came with an argument for disciplined deeptech: stage the risks instead of raising a mountain of money against a distant promise. Its firm-reported assets under management are approximately $200 million across three funds. That figure measures managed investment assets; it is not annual sales.
The firm manages capital for limited partners and buys stakes in startups. Founders receive financing and help building the company, with the economic trade-off of sharing ownership. Its current advertised checks are $1 million to $3 million. Specific ownership percentages depend on a deal. The distinction from a generalist seed investor is the narrow technical thesis and the willingness to work with tiny teams before conventional traction exists.
For a founder choosing investors, that makes the comparison practical. A generalist may offer a broader network; a technical specialist may recognize an unfamiliar product sooner. The question is whether the investor can help with the next obstacle: a manufacturing decision, a first enterprise pilot or a financing story that people outside the laboratory can understand. Ubiquity’s materials make that early operating work central to its pitch.
“Off the beaten path” but not “off the deep end.”Sunil Nagaraj, announcing Fund III
04 / Five months vanished in a spreadsheet
One of Nagaraj’s more revealing stories involves an unnamed company’s board meeting. The previous financial update showed eight months of runway. Forty-five minutes into the meeting, the team disclosed that only three remained. Digging into the discrepancy revealed a misused tab in a downloaded financial model.
That experience prompted Ubiquity’s simpler model, first shared in 2023 and refreshed in September 2026. It focuses on cash, spending and runway. The lesson travels well beyond venture-backed hardware: a forecast is useful only when the people using it understand the calculation. Complexity had concealed the information the founders most urgently needed.
Five months of apparent breathing room disappeared when the model was understood.
05 / A board meeting you can borrow
Ubiquity turns recurring problems into public tools through Ubiquity University. Its videos and templates are free. The six-slide board deck published in September 2026 gives founders a monthly routine: review highlights and lowlights, product and customers, hiring, cash, and the date the money runs out. Keep the format stable so changes become visible.
Nagaraj recommends spending less than 30 minutes mechanically updating subsequent decks. The purpose is reflection and conversation. A founder can copy this without taking Ubiquity’s money: compare last month’s expectations with events, name the uncomfortable problem, and ask for specific help. A cash-out date is particularly hard to wave away. It sits on the calendar alongside everything else.
06 / The field gets a vote
Halter’s hardware story supplies the necessary caution. Early prototypes leaked. Cast parts tore in the field. A vent that survived laboratory testing could be punctured by grass. The response was practical: retrieve failures, investigate causes, use established components where possible, and change the design. Software updates cannot repair every physical mistake.
That sets the conditions for copying Ubiquity’s approach. A team needs a customer problem worth solving, hardware it can make dependable, and a path to revenue that fits the available capital. Long research cycles or uneconomic manufacturing can overwhelm a seed budget. Calling something physical AI does not alter those constraints.
The network matters because the founder’s problems change. Getting a device to work is one assignment; selling it into a hospital or a farm is another. Ubiquity describes help with recruiting, positioning, enterprise pilots and fundraising. Its public community program also brings CEOs together to exchange experience. For an engineer becoming a chief executive, access to somebody who has already made a particular mistake can be as useful as another technical opinion.
Ubiquity is adapting as the category grows. In August 2026 it added investor Cami Gil, who had previously evaluated the firm from the limited-partner side at Allocate. Nagaraj remains the managing partner, supported by an extended network of more than 50 operators. The proposition is concrete: someone should understand the invention deeply enough to fund it early, then care enough about the business to ask when the cash runs out.