Planet Labs has reached profitability on an adjusted basis for the first time, driven by governments buying satellite imagery faster than any commercial customer segment. The Earth observation company reported revenue of $116.1M for its second quarter, up 58% from a year earlier, with adjusted earnings before interest, tax, depreciation and amortisation of $13.9M against a loss in the same quarter last year.
Defence and intelligence revenue grew by more than 90%. Backlog stands at $815.4M, the company describes a satellite services pipeline of around $4B, and it raised full-year guidance to between $430M and $441M. Non-GAAP gross margin reached 59%.
The business behind those numbers is straightforward to describe and was difficult to make pay. Planet operates a fleet of imaging satellites, including its SuperDove, SkySat and higher-resolution Pelican lines, scanning roughly 150M square kilometres of land and 20M square kilometres of ocean every day, then selling access to the resulting archive by subscription.
Defence Buying Changed The Business
The contract list from the past eighteen months explains the shift. Planet signed a multi-year agreement worth €240M with the German government covering dedicated Pelican capacity, direct downlink over Europe, SkySat imagery and AI-assisted surveillance tools, with revenue recognition starting in January 2026.
Around it sit a cluster of smaller but strategically useful deals: a seven-figure NATO contract for space-based surveillance and maritime awareness, a US Navy agreement covering Pacific maritime domain awareness, an expanded arrangement with the US Defense Innovation Unit for Indo-Pacific monitoring, and an $8M award from the National Geospatial-Intelligence Agency.
Not all of it is military. The quarter also brought a contract with the Rwanda Space Agency, described as the first national satellite programme in Africa, and a renewal with the New Mexico State Land Office covering 9M acres.
What The AI Layer Actually Is
Several of these agreements are described as AI-enhanced, and the descriptions stop short of explaining what that means. Planet sells analysis built on top of its archive, detecting change between images of the same place over time, which is the sort of task that suits automation because the volume defeats human review.
What the company has not detailed publicly is how those detections are verified before reaching a customer, or what the error rates are. For a buyer tracking ship movements or construction at a specific site, the difference between a flagged change and a confirmed one is the whole value of the product.
Imaging Once And Selling Repeatedly
The economics work because the same picture serves many buyers. Planet images the whole planet whether or not anyone has ordered a particular scene, which means an additional subscriber costs almost nothing to serve once the satellites are flying.
That shows in the margin profile, with gross margin near 59% and a combined growth and profitability measure the company puts at 70%, against 29% a year earlier. It is closer to software economics than to the project-by-project tasking model that older imagery businesses ran on, and it took a decade of capital spending to arrive at.
The Market Prices Launch Above Data
Planet trades at roughly 14.6 times sales on a market value near $6B. Rocket Lab trades at around 55 times, and SpaceX has been marked in private transactions at something closer to 96 times revenue ahead of its expected listing.
Those are not like-for-like comparisons. Private marks reflect scarcity and access as much as fundamentals, and SpaceX's launch and Starlink businesses are structurally different from selling imagery subscriptions. The pattern they describe is consistent, though: investors are currently paying the highest premiums for launch capacity and vertically integrated infrastructure, and considerably less for the data that infrastructure produces.
Planet Also Builds For Other People
A second line of business ran through the same rocket that carried Planet's own satellites this month. On SpaceX's Transporter-18 rideshare on 1 October, Planet flew 18 SuperDoves and its second Tanager hyperspectral satellite, and also built the prototype spacecraft carrying four of Google's AI processors for Project Suncatcher, the search company's experiment in running machine learning hardware in orbit.
Building and operating spacecraft for other organisations is a different revenue line from selling imagery, and one with obvious appeal when technology companies want something in orbit without becoming satellite manufacturers. It sits alongside a broader movement of capital into space hardware, where commercial constellations and defence programmes are now competing for the same launch slots and engineers.
A Decade Of Spending, Then A Change In Buyers
Earth observation spent years as an industry that could photograph anything and struggled to sell it. The imagery was abundant, the customers were mostly agricultural and governmental, and the budgets were small relative to the cost of building and replacing satellites.
What changed is not the technology but who is buying. European governments rearming, NATO expanding surveillance, and defence agencies treating commercial imagery as an operational input rather than a supplement have turned a thin market into a growing one, and Planet has been positioned to receive that spending.
The figures from this quarter describe a company at the point where a long infrastructure investment starts returning cash, with its largest single contract only beginning to contribute in January. Whether the defence demand behind it holds at this rate is a question about government budgets rather than about satellites.
