The Robot Company That Bought Guards

Knightscope was founded in 2013 to build autonomous security robots. Self-driving machines that patrol parking garages and industrial sites, watch for intruders, and alert a monitoring center. The company calls them the Autonomous Security Force. At the GSX security conference this week, it will present a unified platform combining robots, AI software, and human agents under one contract.

Then look at the financials.

The robot business brought in $11.3 million over all of 2025, a 5% increase from 2024. With a $33.8 million annual loss. In February 2026, Knightscope spent roughly $18 million to acquire Event Risk, a traditional security guard firm with over 400 licensed agents. The quarter ending June 2026 showed revenue of $9.0 million, up 228% year over year. The headline says record growth. The footnote says the guard company is doing the growing.

The more interesting question than whether the robots work is what the acquisition reveals about the robots as a business. You don't buy 400 human guards if you're confident your machines will replace them. You buy them if you need revenue now and the machines haven't delivered enough. That doesn't mean the thesis was wrong. But it does mean the path changed.

Let's look at what Knightscope ships. The K5 autonomous security robot has logged over 4 million patrol hours across the United States. There are 434 clients in 42 states. The company also sells emergency communication devices — wall-mounted panic stations with cameras and two-way video. The next-generation K7 robot is expected to deploy to select clients in Q4 2026. This is not vaporware. The product works, people use it, and it generates revenue.

The revenue is just very small. $11.3 million in a year, against operating expenses of $29.1 million. A business that costs three times more to run than it earns is not a business yet. It's an expensive research program that happens to charge customers. The robots are real. The economics are not there.

The Event Risk acquisition was priced at approximately $18 million: $5 million in cash, $7.3 million in newly issued stock, $4 million in deferred payments through 2028, plus earn-outs and revenue-sharing that could push the total higher. It closed on February 27, 2026.

The revenue impact was immediate. Q1 2026 brought in $6.0 million with only 32 days of Event Risk contribution. Q2 was $9.0 million with a full quarter. The first half of 2026 totaled $15.0 million, nearly triple the $5.7 million in the first half of 2025. Gross margin turned positive for the first time in recent memory, at $0.7 million in Q2, compared to a gross loss of $0.9 million a year earlier. The guard business carries better margins than the robot business does at this scale.

The losses got worse at the same time. Q2 2026 net loss was $14.1 million, up from $6.3 million the year before. Operating expenses jumped from $5.4 million to $13.8 million. The company absorbed 400+ employees, expanded R&D, and carried integration costs. You can't triple revenue and halve your operating expense at the same time.

Cash is the number that tells you how long the experiment lasts. Knightscope held $20.6 million at the end of 2025. By the end of June 2026, it was down to $8.2 million. At the Q2 burn rate, the remaining cash would last less than a quarter. The company has been selling shares gradually into the market through an "at-the-market" equity facility. Share count has climbed to 23.6 million from roughly 20 million a year ago. More stock sales mean each share gets smaller.

This is the constraint every small unprofitable company faces. Grow to profitability fast enough, or keep diluting shareholders to fund the gap. Knightscope has been doing the latter for years. The $42.2 million it raised from equity issuances in 2025 covered the $30.3 million it burned in operations. The question isn't whether it can raise more. It's whether each round of dilution is buying something that compounds.

The GSX pitch is a legitimate business model. The physical security market is enormous and deeply fragmented. AI and robots augmenting human guards to create a lower-cost, higher-quality service is not fantasy. It's just unproven at Knightscope's scale. The company claims 92–93% of industry security alerts are false positives, and that a human-in-the-loop system with AI orchestration solves the problem. If true, that's a real efficiency gain. If the robot can handle the false positives, you need fewer humans per site. The math works if the technology actually delivers.

The problem is that the two businesses still report as separate segments in the filings: "Core Technology Development" and "Acquired Security Force". There's no evidence yet that the robots are being deployed alongside the guards to reduce per-site costs. There's only the promise that they will be.

The contradiction is obvious once you see it. Knightscope's story for 13 years has been about reducing or replacing human guards through autonomous technology. Then it bought a human guard company. You can reconcile the two only if the end state is a hybrid model — robots augment, not replace. That's plausible. It's also not the story that got the company here.

The way to test whether this pivot works is not to look at revenue. Revenue is going to grow, because the guard business adds revenue immediately regardless of the robots. The test is unit economics. Do sites with both robots and guards cost less than sites with guards alone? Does the combination improve retention or contract size? The answer will show up in the Q3 earnings report, scheduled for November 12. Or it won't.

The stock trades at about $1.34, with a market cap of roughly $33 million. That's close to what Knightscope paid for the guard business alone. The robot business — 13 years of development, a product with real users — is priced into this stock at essentially zero.

That's not necessarily a bad thing for a new buyer. It means the market has priced in the risk that the robot story is over. The upside exists if the hybrid model actually produces better economics than either piece alone. The downside is that the company runs out of cash before it proves that, and shareholders get diluted away.

The right thing to watch is whether the next two quarters show the two businesses working together or just sitting next to each other. Revenue growth alone won't tell you. Margin improvement on combined contracts will. If the combined service costs less than the guard-only service did before the acquisition, the thesis is working. If margins stay flat and the robot business keeps burning cash as its own cost center, the acquisition was just a way to make the revenue number bigger.

Either way, the conference booth in Atlanta won't be the answer. The earnings call in November will be.