Cantos Ventures closes $70M Fund V with decade-long thesis on hard tech and defense startups

Cantos Ventures closes $70M Fund V

Cantos Ventures has closed its fifth fund at $70M, up from $50M for its prior vehicle, staying deliberately small after a decade of early-stage hard tech and defense investing that started with a $4M fund zero.

Ian Rountree founded Cantos around a concentrated, founder-first thesis — writing checks small enough that a $1.5M investment represents a meaningful share of the firm's deployed capital. That constraint is intentional. At $70M, Cantos can credibly tell a founder that their company matters to the fund, something that becomes harder to say as fund sizes grow. The firm claims a 90%-plus success rate in helping portfolio companies close their next round.

“New fund we just announced is $70,000,000. Last fund was $50M. We feel like the constraints breed creativity. We want to be able to look a founder in the eye when we write a $1.5M check and say, your company at this stage is a meaningful percentage of my net worth. We have a no PhD policy at Cantos — I learned this the hard way.”

Portfolio construction

Cantos's portfolio includes Niros, Castellan, Radiant, and Shenk (a fish-farming robotics company), across defense, energy, and ag tech. The common thread isn't sector so much as iteration velocity. Rountree and Grant Gregory argue that the most expensive input in hardware companies is labor, not hardware itself, so the firms they back prioritize getting into the field fast — testing, breaking, and refining before the burn compounds.

Shenk is a useful case study. When Cantos invested, the company was licensing robots to fish farms. Rountree pushed founder Safa to stop selling the robots and instead use them to sell better fish directly — a vertically integrated model the founder already believed in but had been talked out of by other investors.

No PhD policy

Cantos has a firm no-PhD policy. The concern isn't technical depth but orientation — researchers trained to defend a thesis are wired to hedge and caveat, which is the wrong reflex for a startup trying to move fast. Rountree acknowledges exceptions (he cites John Gedmark and Astranis as a counterexample of a business-minded founder with a technical background), but the default posture is skepticism toward founders trying to commercialize their doctoral work.

Geography

The firm operates out of San Francisco and Los Angeles, but Texas is increasingly relevant. Cantos is invested in SoluGen and Venus Aerospace in Houston and Resurgent Industries in Austin. The pattern they observe is consistent: companies start in the LA defense corridor, graduate to Torrance or Hawthorne for manufacturing, then move production to Texas, Tennessee, or New Mexico once testing at scale becomes necessary. Seattle is emerging as another node, driven by former Starlink team members.

The AI pull

Rountree is watching AI demand reshape the addressable market for energy and infrastructure portfolio companies without requiring those companies to pivot. Radiant, originally positioned around military microreactors, now fields calls about behind-the-meter power for data centers. The view is that AI demand hands energy companies a growth tailwind they didn't have to engineer — and that the technology will eventually commoditize into broader applications regardless.

Cantos is sitting on a general robotics investment still in stealth. Where to commercialize first is unsettled, but data center infrastructure is on the table alongside more conventional industrial applications.