iRobot’s sale of its defense division to Arlington Capital indicates that commercial markets will drive innovation in autonomy.

On Politico’s Morning Defense today, Jeremy Herb asked some think-tankers what to look for in today’s budget release. Plenty wondered what spending cuts would offset the emerging Third Offset, and even suggested that our New New Thing would be getting underway more slowly than the hype would have. From the news flow, we might have thought otherwise. On Defense One this morning, Patrick Tucker wrote about the US Army’s contract with Cal-Berkeley to develop robotic cockroaches. In Small Wars Journal last week, Gary Anderson of GWU’s Elliott School wrote that “to beat ISIS, we ought to try robotskrieg”—building a droid army to clean out their mine-infested warrens. And fully a year ago on War On The Rocks, Paul Scharre was musing about the difference “between a Roomba and a Terminator” in the coming swarm. So, are hordes of hackable killer robots about to descend on our enemies—or us—or not?

Not everyone is equally convinced of the speed of the advance. As Marjorie Censer wrote more generally for Inside Defense last week, the strategies of some of the larger defense contractors “have begun diverging, staking out increasingly varied paths forward.” One of those is iRobot, soon no longer a defense contractor. Last Wednesday, its shares (NASDAQ: IRBT) closed at $32.43. That evening, the company announced that it would be selling its military business to Arlington Capital Partners for as much as $45 million. The next morning, IRBT opened at $35.01. The $2.58 gain per share represented an increase of about $75 million in market capitalization. Ceteris paribus, one could take this to mean that investors collectively thought that the company was worth $30 million more without the defense division than with it, even after the cash.

How can one sell for 45 but make 75? We lack the full details of the earn-out, and there was no talk of assumption of debt associated with the sale. Perhaps there was other relevant overnight news flow. The $45 million in cash could be used to buy back stock—highly popular for contractors over the past several years—and the company has approval from its board to repurchase up to one million shares this year. Those buy-backs could also be leveraged with borrowings. But note that the company has been spending heartily on research and development—almost $70 million last year, or about 13% of revenues—and without enough to show for it in defense. The 2014 annual report (the most recent) stated that revenues were up 14% overall, and up 19% in the home robot business. The nascent telepresence segment was only getting started, so the defense & security segment had unspecified but clearly declining sales.

Perhaps CEO Colin Angle and his board have thrown in the towel, concluding that the Bob Work is talking big, but that ein klein roboterkrieg will be less economically attractive. Maybe they’ve long been tired of the nein-saying contracting officers who make their lives difficult. Perhaps they’ve concluded that their technology was getting stale; several years ago, co-founder Helen Greiner left to establish Cyphyworks, a microfilament-tethered drone company still focused on the military. Perhaps they just don’t want to play roboterkrieg at all, figuring that an actual drone army wouldn’t be good for the vacuum cleaner brand. Whatever the motivation, the company’s exit is probably a well-considered strategic change. Angle is well known for his presentation on iRobot’s “Fourteen Failed Business Models”—some of which were great for other companies, just not his, at the time.

We could lament this as a further abandonment of the obtuse military market by commercially-minded firms, but I think that the trend here is deeper, and more technological. Angle and his colleagues may have concluded that there’s more to be done in commercial markets, and investors are telling them that they’re $30 million in the right. Arlington Capital’s new company will likely be a net importer of new ideas, but with those military robots still selling for $20,000 to $600,000 each, it will afford the licenses it needs. The spin-off and spin-on of technological progress between commercial and military markets ebbs and flows; we’re just seeing another cycle play out, and in more than smart phone apps. I might thus recommend that for every pilgrimage Pentagon officials make to California, they might consider one as well to the Commonwealth.

James Hasík is a senior fellow at the Brent Scowcroft Center on International Security.