Planet has just printed the best quarter in its history: revenue up 58%, positive adjusted EBITDA, positive free cash flow and $865m of cash. The business is not the problem. The price is. Even after a 65% drawdown from May’s high, $18.12 still capitalises roughly 30% compound revenue growth for a decade at margins the company has never earned — the far right tail of the distribution rather than its centre. Our band is $6.20–10.60.
The operating quarter was genuinely excellent and the balance sheet is a fortress. Q2 FY2027 revenue of $116.1m grew 58% year on year, 8–11% above consensus, at a 59% non-GAAP gross margin, with $13.9m of adjusted EBITDA, positive first-half free cash flow and $865m of cash and short-term investments. Growth plus EBITDA margin summed to roughly seventy. Whatever the conclusion on the shares, the company itself is in the strongest position of its life, and nothing in this report should be read as disputing that.
The price capitalises the tail even against a base case built to be generous. Our central case credits management’s satellite-services pipeline converting broadly on schedule: an 18.5% compound revenue growth rate for ten years to $2.39bn, with adjusted EBITDA reaching 31%. That is worth $7.75 a share. Our bull case — 22.8% compounded to $3.41bn at 34% margins — is worth $12.95. Run the model backwards and $18.12 requires a 30.5% compound rate for ten years to $6.3bn of revenue, fourteen times today’s scale. The shares have fallen 65% from their peak and still sit above our bull case.
Watch bookings, not the growth print. Remaining performance obligations fell from $852m in January to $753m in July; backlog fell from $900m to $815m. Implied first-half bookings of roughly $125m against $210m of revenue is a book-to-bill near 0.6 — the company is converting last year’s landmark wins into revenue faster than it is replacing them. Management’s answer is a $4bn identified pipeline, over $1bn of it described as near-term. That claim is the single most important variable in this valuation and it is testable within two or three quarters.
Planet publishes an unusually complete ladder from prospect to revenue, and the whole investment case turns on how far down it you are willing to underwrite. At the top is a $4bn identified satellite-services pipeline, disclosed on the second-quarter call, with over $1bn of it described as near-term — quarters, not years. Below that sits $815m of company-defined backlog. Below that, $753m of remaining performance obligations, the number that appears in the accounts. And at the bottom, $437m of revenue guided for the year ending 31 January 2027.
| Rung | Amount | What it actually is |
|---|---|---|
| Identified pipeline | $4.0bn | Opportunities management has identified across EMEA, APAC and North America. Not contracted, not bid-adjusted, and disclosed for the first time this quarter. Binds nobody |
| “Near-term” pipeline | >$1.0bn | The subset management expects to convert in quarters rather than years. Also uncontracted. This is the number the base case below credits |
| Backlog | $815m | Company-defined. Includes termination-for-convenience clauses and optional value; $61.8m is cancellable or optional. Up 11% year on year but down from $900m in January |
| Remaining performance obligations | $753m | The GAAP measure, in the Form 10-Q. Up 9% year on year and down 12% from $852m in January. Roughly half converts within twelve months |
| FY2027 revenue, guided | $437m | Midpoint of $430–441m. Up 40–43%. The only figure here that is nearly certain |
The gap between the top rung and the bottom is the entire disagreement. At $18.12 the enterprise is capitalised at $7.01bn, which is 16.0 times this year’s guided revenue, 8.6 times the backlog and 1.8 times the whole identified pipeline. Put the other way round: the market is paying roughly nine years of current revenue, and the contracted book covers about two of them. Everything else has to be won.
That would be an ordinary growth-stock proposition were bookings running ahead of billings. They are not. Remaining performance obligations fell $99m in six months while revenue grew 50%, which implies first-half bookings of roughly $125m against $210m of revenue, a book-to-bill near 0.6. Management’s explanation is that the company is burning last year’s landmark sovereign contracts into revenue faster than it is replacing them, and that the replacements are in the pipeline. That explanation is entirely plausible. It is also, as of the balance-sheet date, unevidenced, and three August wins announced after that date — an $8m NGA award, a German civil tender worth up to €25m over five years, and a seven-figure European defence agreement — are useful but an order of magnitude too small to reverse it.
A growth company guided at +40% cannot run a sub-1 book-to-bill for long without the growth rate following it down. The backlog and RPO lines in December’s third-quarter release are the cleanest early-warning indicator in this entire story, and they are two numbers on one page. If backlog resumes growing sequentially and a nine-figure sovereign award lands, the base case below migrates toward the bull case and this rating gets revisited. If it does not, the pipeline was a description of hope.
One more piece of context on the price itself. Planet rose 474% in the twelve months to late May and peaked at $51.76. It has since fallen 65%. That drawdown is severe enough to feel like a valuation reset, and it is not one: the shares still trade above our bull case. The stock being far cheaper than it was is a statement about the past, not about the price.
Ten-year revenue CAGR 18.5% · cumulative deferred-revenue float $420M · terminal value is 83% of enterprise value.
This is an unlevered discounted cash flow with a float leg, and the float is not a decoration. Ten explicit years run from FY2028 to FY2037. Free cash flow to the firm is adjusted EBITDA, less stock-based compensation charged in full, less capital expenditure, less cash tax, plus the change in deferred revenue. That last line is why Planet generated $68.4m of operating cash flow in the first half against $12.9m of adjusted EBITDA: governments pay satellite-services milestones and multi-year subscriptions up front, and the deferred balance has reached $298.6m, about 68% of this year’s guided revenue. A model that ignores customer prepayment cannot explain why this company is cash-generative at a 1.5% EBITDA margin, and most naive discounted cash flows do ignore it.
The two revenue sliders define the entire path. Growth starts at the FY2028 setting and decays geometrically by whatever factor lands the path on the FY2037 endpoint, so the base case runs 34%, 29%, 25%, 22%, 19% and on down to 9% in the final year. Adjusted EBITDA margin rises from the 1.5% implied by this year’s guidance onto the terminal setting, tracking log-scale rather than the calendar — margin is earned by getting bigger, not by waiting. On base settings that puts FY2028 at 13.7%, which is below the 12.0% Planet already printed in the second quarter only in the sense that the full-year guide is depressed by a deliberate second-half investment ramp. Stock-based compensation and capex ramp linearly from their observed FY2027 levels of 13.5% and 20% of revenue to the terminal settings.
Where each risk is priced, once. Execution, bookings, customer concentration and competition live in the scenario paths and their weights, never in the discount rate. The discount rate carries market risk only. That is why the base case uses 10.5% rather than the roughly 15% a capital-asset-pricing calculation on the shares’ ~2.1 beta would produce: that beta measures the flow-regime volatility of a space-sector basket, not the risk of largely government-contracted cash flows sitting behind a net cash balance. Readers who think flow volatility is a permanent feature of this asset should use the 11.5% row of the grid in Section 03, which is why it is printed. Dilution is priced once too, in the share count, and the slider goes to 560m so that the remaining $1.38bn at-the-market programme can be modelled rather than argued about.
Reading the inputs. Terminal margin, terminal SBC and terminal capex are the three settings that decide the answer, because 83% of enterprise value sits in the terminal year on base settings. That concentration is not a modelling artefact to be apologised for — it is the honest shape of a company whose explicit decade is spent building the thing that eventually earns. It does mean the sensitivity grids matter more than the centre cell, which is why the rating is anchored to a band and not to $7.75.
The number above is not a fact about Planet. It is what these assumptions produce, and the most consequential one is terminal cash conversion: the mature shape of a business that is part data subscription and part satellite manufacturing is genuinely unknown, and nobody — including us — has seen one at scale. Four percentage points of terminal free cash flow margin moves the base case about 20%. The rating survives that range and the band does not narrow it. If you think the mature Planet looks more like software than like a prime, move the terminal margin slider up and use what falls out.
Each card loads into the model above and prints that path’s discounted value: $1.50, $7.75, $12.96 and $21.99. Two notes on construction. First, the bear card is the pure cash-flow answer, and on a path where enterprise value goes slightly negative it is close to meaningless — a business that never covers its cost base is worth its balance sheet. We therefore carry the bear scenario at $2.50, which is net cash of about $0.93 a share plus strategic asset value at roughly 1.9 times forward sales, against the ~4x Maxar fetched when taken private in 2023 with a declining business. Second, because the central case already assumes the pipeline converts, the bear carries 30% weight: a base that credits the upside has to hand its failure branch more probability, not less. Weighting the four at 30 / 45 / 20 / 5 gives $7.90 — close to the base case, because with the optimism moved into the centre little skew remains to the right of it.
| Assumption | Bear | Base | Bull | Superbull |
|---|---|---|---|---|
| FY2028 revenue growth | +14% | +34% | +40% | +46% |
| FY2037 revenue | $0.67bn | $2.39bn | $3.41bn | $4.65bn |
| Ten-year revenue CAGR | 4.4% | 18.5% | 22.8% | 26.7% |
| Terminal adj. EBITDA margin | 20.0% | 31.0% | 34.0% | 38.0% |
| Terminal SBC, % of revenue | 8.0% | 5.8% | 5.5% | 5.0% |
| Terminal capex, % of revenue | 8.5% | 10.5% | 10.5% | 10.0% |
| Terminal deferred revenue, % of revenue | 30% | 30% | 30% | 30% |
| Discount rate | 10.50% | 10.50% | 10.50% | 10.50% |
| Terminal growth | 2.00% | 3.50% | 4.00% | 4.50% |
| Diluted shares | 434.5m in every path — 363.7m basic at 31 July, plus 33m equity awards, plus 38.5m convert shares if-converted at $11.95 | |||
| Cash tax | Zero for four years against the $1.6bn accumulated deficit, then 21% | |||
| Discounted value per share | $1.50 | $7.75 | $12.96 | $21.99 |
| Published scenario value | $2.50 | $7.75 | $12.95 | $22.00 |
The base growth rate is deliberately above the historical base rate, not below it. That is the reverse of our usual posture and it is intentional. Companies at roughly $400m of revenue growing 40% have historically delivered median ten-year sales compounding nearer 10%; our base assumes 18.5%. We set it there because we want the rating to survive crediting management’s own story. If the base case were set at the base rate, the conclusion would be far more negative and far less interesting. A bookings-anchored alternative — growth fading from the current run-rate to the mid single digits, a 13.5% rate to $1.6bn — is worth about $5.70, and readers who weight the balance sheet over the call commentary should anchor there. The rating is the same from either anchor.
A 31% terminal EBITDA margin is above what management has guided to. The stated ambition is 25%-plus. We use 31% because the data-subscription book is genuinely software-like — 98% of annual contract value recurring, 94% annual or multi-year, net dollar retention 109% — while scale in this model comes partly from satellite services, which carries hardware margins. Those two mix effects pull in opposite directions and 31% is our estimate of where they settle. It is an estimate, and it is on a slider.
Stock-based compensation is charged in full and never added back. At $33.5m in the first half it is roughly 16% of revenue and it exceeds adjusted EBITDA. Every non-GAAP profit metric the company reports excludes its single largest real cost of talent. We fade it to 5.8% of revenue by FY2037, which already assumes considerable discipline; holding it at 8% instead costs about a dollar a share.
The deferred-revenue float fades from 68% of revenue to 30%. This is the input we are least sure of and it does real work: each ten points of terminal float is worth roughly $0.40 a share. The fade is deliberate. Today’s balance is inflated by satellite-services milestone prepayments on a small number of large sovereign contracts, and as the revenue base grows prepayment should become a smaller proportion of it. A mature subscription book collecting annually in advance would sit nearer a quarter of revenue than two-thirds. If you think governments keep pre-funding the fleet at today’s intensity, move it up.
Both grids hold the base operating path fixed and vary two inputs at a time. The first is the discount rate against terminal growth, and its corners are the published fair-value band.
| WACC ↓ / terminal growth → | 2.50% | 3.50% | 4.50% |
|---|---|---|---|
| 9.50% bull-market rate | $8.05 | $9.12 | $10.63 |
| 10.50% base | $6.99 | $7.75 | $8.77 |
| 11.50% flow-volatility premium | $6.19 | $6.75 | $7.47 |
The second grid shifts the whole growth path up or down five points a year against the terminal margin. It is wider, because those are the inputs that actually decide this company.
| Growth shift ↓ / terminal EBITDA → | 27% | 31% | 35% |
|---|---|---|---|
| −5pp per year 13.5% CAGR | $4.50 | $5.60 | $6.68 |
| Base path 18.5% CAGR | $6.18 | $7.75 | $9.31 |
| +5pp per year 23.5% CAGR | $8.67 | $10.92 | $13.15 |
Base case $7.75. Fair-value band $6.20–10.60, the corners of the discount-rate grid, which is what anchors the rating. Probability-weighted value $7.90 across the four scenarios at our own weights. Note what the second grid says: even the +5pp growth path at a 35% terminal margin — 23.5% compounded for a decade at software-adjacent margins — produces $13.15, still below the price. There is no cell in either grid that reaches $18.12.
Hold this report’s own base margin structure — 31% terminal EBITDA, SBC fading to 5.8%, a 10.5% discount rate — and solve for the growth that makes the model equal the market. The answer is a 30.5% compound revenue rate for ten years, reaching roughly $6.3bn of FY2037 revenue, fourteen times today’s scale. Grant the bull case’s richer margins and the requirement is still a ten-year rate in the high twenties. Discount the entire bull case at a bull-market 9.5% and it is worth $15.64, which is still below the price.
The burden of proof is therefore on us to explain why the market is wrong, and there is a readable answer. The sell side ratcheted targets up with the price through a sector mania in which one bank moved from $8 to $12 to $53 in nine months, and consensus targets sat below the market price for much of the run — the classic signature of anchoring rather than analysis. The whole space complex re-rated as a single basket around the SpaceX listing and is still unwinding. And the issuer itself registered $1.5bn of stock for sale on the day it reported first-quarter results, then sold $120m of it at an average of $31.95. The best-informed seller in this market has been a seller.
Cheap against other space stocks, expensive against every cash-flow yardstick. Every name at the same 4 September close, on share counts taken from each company’s own reporting rather than estimated.
| Company | Price | Market cap | Current-FY revenue | Latest qtr growth | Cap / sales |
|---|---|---|---|---|---|
| Planet Labs (PL) | $18.12 | $7.87bn | $430–441m | +58% | 18.0x |
| BlackSky (BKSY) | $20.50 | $0.84bn | $130–150m | +50% | 6.0x |
| Rocket Lab (RKLB) | $64.26 | $41.09bn | ~$900m | +62% | 45.7x |
Market capitalisation is the 4 September close times shares outstanding — 434.5m fully diluted for Planet, and weighted counts of 40.9m and 639.4m for BlackSky and Rocket Lab. Planet’s enterprise value, net of $865m of cash and short-term investments, is $7.01bn, or 16.0 times the guidance midpoint. We do not publish enterprise values for the peers: their net-cash positions were not verified on the same basis, and an unverified denominator is worse than a missing one. Peer revenue guidance, growth and backlog are as each company reported them for its most recent quarter.
Read that three ways. Within the space basket Planet at eighteen times sits between BlackSky at six and Rocket Lab at forty-six. Some premium to BlackSky is deserved — three times the revenue, a unique daily-scan asset, positive company-level free cash flow, an $815m backlog against $378m — but three times the multiple for a business growing at a similar underlying rate is generous. Against its own history the re-rating is the whole story: Planet traded at two to four times sales for most of 2023 to 2025, and the move to the high teens happened in nine months on the defence-space narrative and the SpaceX listing wave. Against cash flow, none of these multiples survives contact with a discount rate, which is what the grids above show.
Intrinsic value is not a price target and this report does not publish one. But the difference is worth being explicit about, because it explains why we expect this rating to be early rather than immediately vindicated. Apply scenario-consistent forward multiples — the regimes this sector has actually traded in this year — and the twelve-month trading ranges are roughly $7–9.5 in the bear, $17–19.5 in the base, $23–27.5 in the bull and $32–38 in the superbull. Weighted, that is about $17.5, close to the current price. Near term the stock is roughly fairly priced as a bet within the prevailing space-sector multiple regime. The overvaluation is a statement about cash flows, and it shows up if and when the regime normalises. We publish the rating anyway, because a rating that only fires once it is already obvious is not worth publishing.
The base case is built to credit management. It assumes the “over $1bn near-term” pipeline converts broadly on schedule, which sits ahead of the current bookings evidence — a first-half book-to-bill near 0.6 and RPO down 12% since January — and above historical growth-persistence base rates for companies at this scale. Terminal cash conversion dominates the answer, and the mature shape of a data-software and satellite-hardware hybrid is genuinely unknown; four points of terminal margin moves the base case about 20% either way. FY2028 consensus was not directly observable at writing, so the +34% base is an assumption rather than a consensus anchor. The bear floor rests on private-market transaction evidence, the 2023 Maxar take-private, not on any Planet-specific bid. Consensus estimates for the second quarter differed across sources between $104.5m and $107.2m, which is why the beat is described as a range. And we discount below the rate the shares’ own beta implies, deliberately; the 11.5% row exists for readers who disagree.
The trend is unambiguously down. Price sits below every moving average and the averages are themselves inverted: 20-day $22.1, 50-day $23.7, 200-day $26.9, 100-day $30.8. The 14-day relative strength index near 29 is oversold, so bounces from here are normal and should not be mistaken for a change of regime. Thirty-day realised volatility is around 60% annualised. Support sits at $17.73, the 4 September low, then $16.00, the 78.6% retracement of the entire $6.26 to $51.76 advance, then an unfilled gap toward $13–15 from last December’s results. Resistance runs $20.4–21.6, the August breakdown shelf, then $23.6, $24.75, $29 and the $32–34 June ledge. For anyone tempted by the bounce, the burden is a reclaim of $21.6 on volume; below that, rallies are distribution. The chart and the model agree, which is unusual enough to note and is not evidence of anything on its own.
Planet reported after the close on 3 September, with the call the same evening. The shares fell 8.2% into the print and recovered roughly 4–6% after hours, then closed the following session at $18.12, marginally below the pre-results price. That is the tape’s verdict on a quarter that beat on every line.
| Metric | Q2 FY27 | Q2 FY26 | Change | Why it matters |
|---|---|---|---|---|
| Revenue | $116.1m | $73.4m | +58% | Fastest growth since listing, and 8–11% above consensus |
| Point-in-time share of revenue | 12% | 1% | +11pp | Quantifies the satellite-services lump. Underlying ratable growth was ~41% |
| Non-GAAP gross margin | 59% | 61% | −2pp | Hardware mix dilutes, and it still cleared guidance of 52–55% |
| Adjusted EBITDA | $13.9m | $6.4m | +117% | A 12% margin while growing 58%. Real operating leverage |
| GAAP net loss | ($9.4m) | ($22.6m) | +58% | $17.1m of stock compensation is the remaining GAAP wedge |
| Remaining performance obligations | $753m | ~$691m | +9% | But down from $852m in January — see Section 01 |
| Backlog | $815m | ~$734m | +11% | Half converts within twelve months, so over $400m of next-twelve-month revenue is visible |
| Net dollar retention | 109% | ~99–100% | +~9pp | Existing customers now expand meaningfully. A genuine inflection |
| Cash & short-term investments | $865m | $271m | +219% | Converts, warrants, the at-the-market programme, and positive free cash flow |
| H1 adjusted free cash flow | +$28.8m | +$54.3m | −47% | Positive while capex nearly doubled — prepayments fund the fleet |
Prior-year RPO, backlog and net dollar retention are derived from the year-on-year percentages the company stated rather than separately reported. Source: press release, Form 10-Q and earnings call, all 3 September 2026.
One: a clean double beat with record everything. Revenue of $116.1m came in 8–11% above consensus, with sources placing that consensus between $104.5m and $107.2m. Non-GAAP earnings of two cents reversed an expected two-cent loss. Adjusted EBITDA more than doubled. This was a fourth consecutive quarter at or above the Rule of 40, this one at roughly seventy, which the chief executive celebrated on the call in exactly the terms you would expect.
Two: part of the beat was timing, and management said so unprompted. Commissioning of the Swedish Armed Forces satellite finished faster than planned, so point-in-time handover revenue landed in the second quarter instead of the third. Point-in-time revenue was $14.1m, 12% of the quarter, against $0.8m a year earlier. Strip it from both sides and underlying ratable revenue grew 40.6% — still exceptional, and the right base for extrapolation. It is also why the third-quarter guide of $101–105m, up about 27% at the midpoint, looks soft without being a demand signal. Screeners will misread this metric in both directions as satellite services scales.
Three: defence and intelligence is now the company. That line grew 93% to $81.0m, roughly 70% of revenue. EMEA grew 135% to $55.2m, making Europe Planet’s largest region ahead of North America’s $39.0m. Commercial grew 17% and civil government 7%. A business that described itself as a broad commercial data platform three years ago is now, on the numbers, a European-weighted defence contractor with a commercial side.
Four: the low end of guidance went up and capex went up a lot. Full-year revenue moved to $430–441m, lifting the floor by $5m; non-GAAP gross margin to 55–57%, above the prior range’s top; adjusted EBITDA to $3–10m. Capex jumped from $80–95m to $100–115m for advanced procurement of long-lead Pelican and Owl components and the San Francisco and Berlin factory build-outs. The chief financial officer reaffirmed positive adjusted free cash flow for the full year. The negative surprises in the quarter were the third-quarter EBITDA loss guide of −$6m to −$1m and the size of that capex step-up, both of which are investment decisions rather than demand problems, and both of which are real cash.
Five: the strategic pitch got specific. A $4bn identified satellite-services pipeline with over $1bn near-term. The Owl next-generation constellation — 1-metre class, roughly ten times the data, ten times faster, latency down to an hour — being accelerated at customers’ request, with management explicit that customers expect prices to rise. And the August NGA Global Monitoring Service award disclosed as sole-source, with Planet the only vendor considered. Sole-source is the single most valuable disclosure in the quarter, because it is a moat statement priced by procurement officers rather than by promoters.
On tone: the chief executive was confident bordering on exuberant, while the chief financial officer’s sections were disciplined — repeatedly re-anchoring to annual rather than quarterly free cash flow, calling out the second-to-third-quarter timing shift before anyone asked, and describing at-the-market usage as balancing market dynamics against a desire to minimise dilution. Analyst questions concentrated on five things: whether commoditising AI models erode Planet’s edge, the credibility and geography of the $4bn pipeline, the RPO step-down, the halved-margin implication of second-half guidance, and the mid-year capex raise. The answers were direct on all five. What to check in the third-quarter release, in order: the backlog line, the RPO line, and the at-the-market disclosure in the Form 10-Q.
Planet Labs was founded in San Francisco in 2010 by three NASA scientists and has been listed since a December 2021 merger with a special-purpose acquisition company. Its core asset is a daily scan: roughly 200 shoebox-sized SuperDove satellites photograph essentially all of Earth’s land mass every day at about three-metre resolution. No other operator, Western or otherwise, does this. On top sit around ten Pelican high-resolution tasking satellites in the sub-metre class with 30-centimetre capability in development, the Tanager hyperspectral line built with Carbon Mapper and JPL for methane and carbon-dioxide detection, and a ten-year archive the company says holds thousands of images of every point on land — the raw material for AI-based change detection.
Customers are governments first. Defence and intelligence agencies — the US National Geospatial-Intelligence Agency, the US Navy, NATO, and a lengthening list of European and Asian ministries — took roughly 70% of second-quarter revenue. Civil government customers, including agriculture ministries, land offices and space agencies, make up part of the rest; Rwanda became Planet’s first national programme in Africa this quarter. Commercial customers run from agriculture platforms in the John Deere ecosystem and FarmQA to insurers, mappers and, lately, a hyperscaler AI developer paying to monitor data-centre construction.
| Engine | What is sold | Revenue character |
|---|---|---|
| Data subscriptions | Licences to the daily scan, tasking capacity and the archive | Recurring and ratable. 98% of the annual contract value book is recurring, 94% annual or multi-year, net dollar retention 109% |
| AI-enabled solutions | Analytics on top of the data: Global Monitoring Service pattern-of-life warning, Maritime Domain Awareness, agricultural yield models, and a natural-language application now in open beta | Recurring and higher value. The vector by which Planet moves up-market |
| Satellite services | Building, launching and operating sovereign satellites and dedicated capacity for governments — Sweden’s armed forces, JSAT, the German government | Large and lumpy, partly recognised at a point in time on handover. Prepaid, and the main driver of the $299m deferred-revenue balance |
The third engine is the newest and the one repricing the story. Planet delivered Sweden’s first sovereign reconnaissance satellite roughly four months after contract signature, in an industry that has historically taken years, and handed it over in the second quarter — which produced the quarter’s beat. Management now discloses a $4bn identified pipeline, over a quarter of it near-term, spread across EMEA, APAC and North America, and is doubling manufacturing capacity with a Berlin factory plus a launch partnership with Germany’s Isar Aerospace. In defence procurement, speed is currently the scarcest good, and Planet has it.
Management says Planet holds under 5% of today’s Earth-observation market, excluding satellite services. That claim deserves scrutiny, because it is doing quiet work in the bull case. Planet’s trailing revenue is $378m. For that to be under 5% of the market, the market would have to be at least $8bn today. Novaspace, the Euroconsult successor and the industry’s standard reference, put commercial Earth-observation data and services at about $5bn as at its late-2024 edition, and projected it to exceed $8bn only by 2033, with the services segment growing from $3.1bn to $4.9bn over that period. On those numbers Planet holds something nearer 7–8% of the market, not under 5%, and management’s implied denominator is roughly the market Novaspace expects seven years from now.
That is not an accusation of bad faith — management may be counting categories Novaspace excludes, and the definition of this market is genuinely contested. It is a caution about the headroom argument. The commonly cited market-research estimates for this sector range from under $4bn to nearly $17bn for what is nominally the same thing, an eightfold spread, which is why none of them appear in our model. Nothing in the valuation above depends on a total-addressable-market figure; the revenue path is built from the company’s own pipeline disclosure and its own bookings. The separate figure management cited on the call, that investment managers alone may buy some $23bn of alternative data annually by 2030, is a third-party forecast repeated by the company and we have not verified it.
Three currents dominate. Rearmament and sovereignty: European governments want national eyes in space and are increasingly buying commercial rather than building. Germany has committed well over €250m to Planet across defence and now civil programmes, Sweden bought a sovereign satellite, NATO buys integrated optical-plus-radar products, and the current US administration is leaning into commercial geospatial procurement. AI, which raises the value of proprietary, calibrated, historical real-world data — the chief executive’s answer to the commoditisation question was that cheap models make the data the scarce asset, and he is probably right. And the SpaceX effect: its mid-2026 listing first inflated every space equity and then drained them as capital rotated into the leader. The sector now trades as one high-beta basket, which cuts both ways.
In daily broad-area monitoring Planet is effectively unopposed, which is the basis of its sole-source NGA and US Navy awards. In high-resolution tasking it competes with Vantor, the former Maxar, now private; with Airbus; with BlackSky, whose Gen-3 constellation is scaling fast; and with national systems. The coming Owl constellation deliberately walks into that contested market, where daily-scan logic does not apply and price competition is real. In synthetic-aperture radar, ICEYE and Umbra lead and Planet partners rather than builds. In sovereign satellite services its rivals are traditional primes — Airbus, OHB, IAI — that deliver in years against Planet’s months. The systemic threat is verticalisation: a SpaceX or Starshield entry, or a Chinese state constellation, deciding to give away good-enough imagery and collapsing the price umbrella under everyone.
| Fiscal year, ends 31 January | FY2025 | FY2026 | FY2027E, guided |
|---|---|---|---|
| Revenue | $244.4m | $307.7m | $430–441m |
| Growth | +11% | +26% | +40–43% |
| Adjusted EBITDA | ~($9m) | modestly positive | +$3 to $10m |
The cash story deserves emphasis because it is unusual for this sector. First-half operating cash flow was $68.4m against adjusted EBITDA of $12.9m. The wedge is customer prepayment: deferred revenue grew $50m in the half to $298.6m, plus interest on the cash pile. Add $107.8m of warrant proceeds and roughly $120m of at-the-market share sales at an average $31.95, and cash plus short-term investments reached $865.4m against $460m of 0.50% convertible notes due 2030. Those notes convert at $11.95, with capped calls lifting the effective conversion price to $18.04. Shareholders’ equity is $564m. The accumulated deficit of $1.6bn records what it cost to build the fleet and the archive — and, in this model, shelters the first four years of cash tax.
| Dimension | Score | Assessment |
|---|---|---|
| Proprietary data and IP | 5 | The daily scan and the ten-year calibrated archive are unreplicated. Bringing telescope production in house, a multi-year effort, removed a supply-chain choke point |
| Scale | 5 | 688 satellites launched on 42 rockets of ten types; around 200 imaging satellites operating. Cost per usable image is far below rivals’, and spreading failure risk across many cheap assets is a structural advantage, not a compromise |
| Switching costs | 4 | Programmes of record embed Planet in customer workflows — NGA’s Global Monitoring Service out of a DIU pilot, Navy maritime awareness, national agriculture-payment systems in Czechia and Scotland. Archives do not transfer |
| Cost advantage | 4 | Capex around 24% of revenue in a build year, against incumbents that historically spent multiples of that per unit of capacity. SuperDoves are consumables, not cathedrals |
| Customer loyalty | 4 | Net dollar retention 109%, positive winback rate, renewals sole-sourced. Government churn is procedural rather than preferential |
| Regulatory position | 3 | US commercial remote-sensing licensing and growing sovereign-industrial status in Germany both help. But export rules cap what can be sold where, and sovereignty logic can eventually favour local champions over a US operator with a European factory |
| Distribution | 3 | Direct government sales are strong; the partner and commercial channel is young. The natural-language application is an attempt to make distribution self-serve |
| Brand | 3 | Excellent inside the geospatial and defence world, thin outside it |
| Network effects | 2 | Mostly absent. More customers do not improve the product, though more archive queries marginally improve the AI layer |
The verdict is directional rather than static. The moat is widening where Planet already wins — broad-area daily monitoring, change detection, maritime awareness — because AI multiplies the value of the archive and because sole-source awards are formal evidence that procurement officers went looking for a substitute and did not find one. It is contested where Planet is heading: Owl pushes into 1-metre tasking against Vantor, Airbus, BlackSky and national systems, a market where the daily-scan logic does not apply and price competition is real. A company can have an excellent moat around the business it has and no moat around the business it is buying growth in, and that is roughly the situation here.
Biggest structural threats, in order. A SpaceX or Starshield entry into commercial imagery, where the capital and launch-cost advantages dwarf everyone’s. Chinese constellations compressing international pricing. And European sovereignty preferences eventually demanding European owners rather than merely European factories — the same political current that is currently Planet’s largest tailwind could reverse on it.
| Concern | Severity | Evidence and why it matters |
|---|---|---|
| Bookings behind billings | HIGH | RPO fell from $852m in January to $753m in July; backlog from $900m to $815m. Implied first-half bookings of about $125m against $210m of revenue is a book-to-bill near 0.6. Backlog is still up 11% year on year and the August wins post-date the balance-sheet date, but a growth company guided at +40% cannot run a sub-1 book-to-bill for long. The cleanest early-warning indicator in the story |
| Dilution | HIGH | Weighted shares rose from 300m to 360m in five quarters, about +20%, through warrants, stock compensation and the at-the-market programme. A $1.5bn shelf was registered in June — the event that broke the stock, down 26% on 5 June — with roughly $1.38bn still available. Management sold $120m at an average $31.95, i.e. issued stock 76% above today’s price. Excellent treasury management, and also the issuer’s own read on value. In fairness, the Form 10-Q cover shows the count rose only ~0.2m between 1 and 27 August, so there has been no meaningful issuance at post-crash prices |
| Revenue quality and point-in-time lumps | MED-HIGH | 12% of second-quarter revenue was recognised at a point in time on satellite handover, against 1% a year earlier. Headline growth of 58% overstates the roughly 41% underlying run-rate, and this metric will whipsaw quarter to quarter as satellite services scales |
| Customer concentration | MED-HIGH | Defence and intelligence is about 70% of revenue. Three customers were 17%, 11% and 10% of the quarter, and two held 36% of receivables. US federal spending faces continuing-resolution risk every October, many contracts carry termination-for-convenience clauses, and $61.8m of backlog is cancellable or optional. Sovereign deals concentrate further — Germany alone spans multiple large programmes |
| Stock-based compensation | MEDIUM | $33.5m in the first half, roughly 16% of revenue, which exceeds adjusted EBITDA. Every non-GAAP profit metric the company reports excludes its largest real cost of talent. Our valuation charges it in full |
| Litigation | MEDIUM | A Delaware class action and an acquisition-related dispute produced $6.2m of first-quarter expense and $7.5m of settlement payments in the half, both excluded from adjusted free cash flow. Manageable in size; the direction is worth watching |
| GAAP optics | LOW | FY2026’s headline net loss of $246.9m and the first quarter’s $138.9m were dominated by non-cash warrant revaluations that ended with the warrant redemption. Alarming on screeners, economically close to noise |
| Insider selling | LOW | The chief executive filed intent to sell 200k shares in July, immaterial against his holding. No pattern of heavy insider distribution identified |
Overall red-flag score: 5 out of 10. Nothing here suggests accounting misrepresentation. The flags are structural — dilution, concentration, lumpy recognition — rather than forensic, and $865m of cash against $460m of cheap convertible debt removes any solvency dimension from the analysis. That matters for how the bear case is constructed: this is not a company that can go to zero on a bad year, which is precisely why the bear scenario is floored at asset value rather than at the cash-flow model’s answer.
| Strength | Weight | Evidence |
|---|---|---|
| Sole-source positions | HIGH VALUE | The August NGA Global Monitoring Service award and the US Navy maritime renewal were both sole-sourced. The US government formally checked for alternatives to the daily scan and found none. That is a moat statement priced by procurement officers, not by promoters |
| Speed as a weapon | HIGH VALUE | First satellites delivered two and four months after contract signature on the two latest sovereign deals, against an industry norm of years. In defence procurement today, speed is the scarcest good |
| Prepaid growth | HIGH VALUE | Deferred revenue of $299m, about 68% of guided full-year revenue, means customers finance the fleet build. First-half operating cash flow of $68m on $13m of EBITDA is the visible result — a working-capital engine most models, including naive discounted cash flows, miss entirely. Ours does not |
| Retention inflection | MEDIUM | Net dollar retention at 109% against roughly flat historically, 98% recurring annual contract value, 94% annual or multi-year. The data business is quietly becoming a normal, good subscription book |
| Calibration and archive barrier | MEDIUM | Fleets cross-calibrated to Landsat, Sentinel and MODIS with a decade of backwards-compatible data. A rival launching a daily scan tomorrow would still lack the archive that makes change detection work — and AI raises the value of exactly this asset |
| Zero-priced optionality | MEDIUM | Owl pricing uplift, AI data licensing to model builders, Berlin as a European industrial-policy magnet, and a hyperscaler already paying to watch data-centre construction. None of it is in backlog and all of it is free at the right price. The problem is that today’s price charges for it several times over |
| Dimension | Score | Assessment |
|---|---|---|
| CEO track record | 3 | Will Marshall, co-founder, physicist, ex-NASA, built the category and the fleet, which is a genuine technical achievement. But the 2021 merger-era projections badly missed — the deal deck’s mid-decade revenue ambitions of around $700m arrived years late and hundreds of millions short. That scar is relevant precisely because a $4bn pipeline is now the valuation’s load-bearing wall. The last four quarters have been rebuilding that credibility, beat by beat |
| CFO credibility | 4 | Ashley Johnson, president and chief financial officer, is the disciplined voice: she pre-empted the timing question, anchors to annual rather than quarterly free cash flow, and executed the financing sequence well. Docked a point for the non-GAAP architecture that adjusts away stock compensation and litigation |
| Guidance accuracy | 4 | Every quarter of FY2026 and FY2027 to date has met or beaten guidance, and full-year revenue guidance has only moved up: $415–440m, then $425–441m, then $430–441m. The second quarter beat its own guide’s top end by $9m. Guiding low and clearing is now an established habit |
| Capital allocation | 4 | Objectively excellent sequencing. $460m of converts at a 0.50% coupon with capped calls to $18.04, struck in September 2025 with the stock around $9. Warrants redeemed for $107.8m near the highs. At-the-market sales averaging $31.95. The company sold its own stock far better than most of its shareholders did. Buying nothing back and stockpiling $865m is defensible given the pipeline |
| Dilution discipline | 2 | The flip side of the same facts: roughly +20% weighted share count in five quarters and a $1.5bn shelf still largely undrawn. Good treasury for the company is a tax on holders. Minimising dilution is asserted more than it is demonstrated |
| Compensation | 2 | Stock compensation at 15–16% of revenue is high even for growth software, and it is excluded from every headline profit metric management celebrates |
| Transparency | 4 | Point-in-time disclosure, book-of-business metrics, net dollar retention, a backlog-versus-RPO reconciliation, and sector and regional growth all volunteered. The call handled the hard questions — the RPO decline and second-half margins — directly rather than deflecting |
| Board and structure | 3 | Founder supervoting Class B shares entrench control. Public-benefit-corporation status aligns with the mission and adds a non-shareholder objective. Insider selling is de minimis |
Composite: 3.3 out of 5. They act like long-term owners of the mission and increasingly like disciplined operators of the business. The residual doubt is whether they treat the share count as something to be protected or as a funding instrument. So far it has been a funding instrument, timed brilliantly — which is a compliment to them and a cost to anyone who held through it.
| Window | Event | Upside | Downside | Confidence |
|---|---|---|---|---|
| Autumn 2026 | Transporter-18: Tanager-2 plus 18 SuperDoves, third launch of the year, satellites already at Vandenberg | Fleet expansion on schedule; Carbon Mapper capacity doubles | Launch slip or anomaly in a tight rideshare market | High |
| 1 Oct 2026 | US federal fiscal year begins; appropriations or continuing-resolution question for NGA, NRO and Defense commercial-imagery lines | Commercial-first budget language expands programmes | A continuing resolution freezes new starts and leaves options unexercised | High on timing, low on outcome |
| Oct–Dec 2026 | Satellite-services awards from the near-term pipeline; German civil ramp; possible further European wins | A nine-figure sovereign award would validate the whole thesis and likely re-rate the stock | Another sub-1 book-to-bill quarter confirms the central concern | Medium |
| Early Dec 2026 | Q3 FY2027 results, guided to $101–105m and −$6m to −$1m of EBITDA | The beat pattern continues to five; bookings inflect | A first revenue guide miss in the new era would be punished severely at this multiple | High |
| By Jan 2027 | Berlin factory begins building satellites; London hub ramps NATO engagement | European industrial-policy orders follow the factory | Cost creep — the capex raise partly funds this | Medium-high |
| Quarterly | At-the-market disclosures in the Form 10-Q, with roughly $1.38bn of the $1.5bn programme remaining | Restraint at low prices signals valuation discipline | Persistent supply caps every rally | High |
| ~Mar 2027 | Q4 FY2027 results and first FY2028 guidance — the year’s single biggest catalyst | A guide at or above +30% would break our base case upward and force a revisit | A guide in the low twenties collapses the growth premium | High |
| 2027 | Isar Aerospace launches a German-built Pelican, a national first; Owl programme milestones and possible acceleration | Owl pre-commitments at higher prices, which management says customers expect | New-launcher risk; Owl schedule slip against incumbent high-resolution rivals | Medium |
| Ongoing | Delaware class action and acquisition dispute; NGA commercial-imagery recompetes | Favourable resolution removes an overhang | Adverse ruling or unfavourable recompete terms | Low |
Confidence reflects the timing of the event, not the direction of its outcome. The recompete timing is our assumption. Sources: company releases and the Q2 FY2027 call of 3 September 2026, and the published US budget calendar.
BULLYou are early, not late. Growth accelerated across five quarters — 10%, 21%, 42%, 58% — with net dollar retention inflecting to 109%. Businesses at this point on the S-curve do not obey decelerating linear extrapolations, and every model in this report is a decelerating linear extrapolation.
BEARPrice is the thesis, and the price needs roughly 30% compounded for a decade at margins the company has never earned. Not good execution — near-perfection, sustained to 2037. And the base case that produces $7.75 already treats management’s own 25%-plus margin target generously. That is not pessimism, it is arithmetic.
BULLThe pipeline is not a promise, it is a queue. Sweden went from signature to orbit in four months. Germany came back for a second and a third programme. The constraint is factory capacity, which is precisely why Berlin exists. Over $1bn near-term on a $437m base means bookings can double revenue’s trajectory inside two years.
BEARBookings already blinked. RPO down 12% and backlog down 9% in six months, while revenue grew 50%. The bull asks to be judged on future bookings precisely because current bookings failed the test. Judge me on the December print and I will judge you on the one we already have.
BULLSole-source is the moat made audible. The US government went looking for an alternative to the daily scan and certified in writing that there is none. That is pricing power in waiting — and Owl is the price rise, with customers already telling management they expect to pay more.
BEARThe growth print is flattered. Underlying ratable growth was 41%, guided to about 27% next quarter. The 58% headline is the Swedish satellite, once. And Owl walks into the one market where Planet has no structural advantage, against Vantor, Airbus and BlackSky, where price competition is real.
BULLAI turns the archive into an annuity. Every foundation-model builder needs calibrated real-world data and Planet has a decade of it, with a hyperscaler already paying to watch data-centre construction. One licensing deal rewrites this model, and none of it is in backlog or in the bear’s numbers.
BEARManagement is on my side of the trade. They registered $1.5bn of stock the day the price peaked on fundamentals and sold $120m at $31.95. The most informed seller in this market has $1.38bn left to go, and the share count is already up 20% in five quarters with compensation running at 15% of revenue.
BULLThe financials de-risked completely. $865m of cash, free-cash-flow positive, customers prepaying $299m. The 2021-era “will they make it” discount should be gone. And the bear is quoting a discounted cash flow while strategic buyers paid roughly four times sales for a declining Maxar.
BEARConcentration cuts when it cuts. Seventy percent defence and intelligence, a US budget process in permanent continuing-resolution risk, termination-for-convenience clauses, and a German pillar exposed to political rotation. One procurement winter erases a year of pipeline talk, and the multiple has no room to absorb it.
On the business, the bull wins comfortably. Acceleration, retention, sole-source validation and prepaid cash generation are real, and the bear did not seriously contest any of them. On the security at $18.12, the bear wins. The bull’s own best numbers, run through the model on this page, arrive between $13 and $16 — and the bull’s strongest rebuttals, the S-curve and the licensing optionality, are claims about the tail that the price already owns. What is genuinely uncertain, and decides who is right inside a year, is bookings: whether backlog resumes growing sequentially and whether any nine-figure satellite-services award lands. Verify next: the Q3 backlog and RPO print in early December, at-the-market activity in the Form 10-Q, and FY2028 guidance in March. If backlog re-inflects and FY2028 is guided above +30%, the base case migrates toward the bull and this rating gets revisited. That is a dated, testable exit from this view, not a permanent one.
Planet owns about two hundred small satellites that photograph the entire Earth every single day, plus a smaller number of sharp-eyed satellites for close-ups. It sells subscriptions to the pictures, and to software that watches the pictures and raises its hand when something changes: a ship where no ship should be, a runway being extended, a forest disappearing.
Lately it has added a third trade: building and flying whole satellites for countries that want their own. Sweden already flies one. Germany is buying capacity. This is the part that is changing the story, because it turns a subscription business into something that also builds hardware to order, and hardware orders are large, lumpy and paid up front.
Governments are in a hurry to see the world from space. Artificial intelligence makes a ten-year photo archive suddenly very valuable, because you need real-world data to train on and Planet has more of it than anyone. And Planet is the only company on Earth with the everyday, everywhere picture. Sales grew 58% last quarter, the company finally makes a small operating profit on its preferred measure, and it sits on $865m of cash.
Planet tells investors it has identified $4bn of future satellite deals. Its accounts say customers are currently obliged to pay it $753m. Neither number is dishonest — the big one counts opportunities the sales team has found, the small one counts signed commitments. But the small number has fallen 12% since January while revenue grew 50%, which means new orders have lately been arriving more slowly than old orders are being used up. Management says the pipeline will fix that. It might. If you take one thing from this report, take that this is the number to check, and that it is published every quarter.
Here is the whole disagreement. After falling 65% from May’s peak, the stock still costs about sixteen times this year’s sales. Work the cash flows honestly — even crediting the company’s big pipeline as likely to land — and the business is worth around $6–11 a share to us, perhaps $13 if nearly everything goes right. At $18.12 you are paying today for the version of Planet that wins almost everything for a decade. Wonderful company; the shares are priced beyond even that compliment.
Right: a large country signs a nine-figure satellite deal from that $4bn list, the new higher-resolution Owl satellites let Planet raise prices, and an AI company pays to train on the archive. Any one of those would force us to revisit this.
Wrong: new orders keep arriving slower than old ones are consumed, the US budget process freezes defence spending in October, or the company keeps issuing shares — it has permission on file to sell $1.4bn more, and the share count already grew about 20% in fifteen months.
The backlog line in December’s third-quarter report, and next year’s revenue guidance in March. Everything else in this report is context for those two.
| Status | What it covers |
|---|---|
| Verified from Planet filings | All revenue, gross margin, adjusted EBITDA, GAAP net loss, cash and short-term investments, convertible notes and their terms, shareholders’ equity, accumulated deficit, deferred revenue, remaining performance obligations, backlog and its cancellable portion, share counts, customer-concentration percentages, revenue disaggregation by sector and region, point-in-time revenue, stock-based compensation, litigation expense and settlement payments, and every guidance range — taken from the Q2 FY2027 press release and condensed financial statements and the Form 10-Q for the quarter ended 31 July 2026, both of 3 September 2026, together with the Q1 FY2027 release of 4 June 2026 and the convertible-note pricing and closing releases of September 2025. |
| Verified from market data | The $18.12 closing price of 4 September 2026 and the daily bar series behind it, including the $17.73 session low and the $51.76 May peak; the moving averages, relative-strength reading and realised volatility in Section 03; and the 4 September closing prices and reported share counts for BlackSky and Rocket Lab used in the peer table. Consensus, analyst-target and June-selloff context is from financial-press coverage between 5 June and 4 September 2026 and is attributed as such. |
| Third-party, attributed | Market sizing for commercial Earth-observation data and services is Novaspace’s, from the 17th edition of its Earth Observation Data and Services Market report as reported in November 2024: approximately $5bn at that date, exceeding $8bn by 2033, with the services segment growing from $3.1bn to $4.9bn. We use it only to test management’s market-share claim and nothing in the valuation depends on it. |
| Our estimate or judgement | Enterprise value and every multiple in Sections 01 and 03; the implied first-half bookings figure of roughly $125m and the 0.6 book-to-bill derived from it; the entire ten-year cash-flow path, the margin-ramp shape, the deferred-revenue fade, the four scenario settings, both sensitivity grids, the reverse test, the $6.20–10.60 band and the $7.75 central estimate; the 30 / 45 / 20 / 5 scenario weights and the $2.50 bear floor; the twelve-month trading ranges; and all scores in Sections 06, 07 and 08. The model on this page is the model that produced those figures — the scenario cards load the published settings and the grid cells are reproducible from them. |
| Still open | The $4bn pipeline is a company disclosure with no published methodology, no bid-adjustment and no geographic or customer breakdown beyond three named regions, so it cannot be independently sized; the same applies to the “over $1bn near-term” subset that our base case credits. FY2028 consensus revenue was not directly observable at writing, so the +34% base growth rate is an assumption rather than a consensus anchor. Second-quarter consensus differed across sources between $104.5m and $107.2m, and the beat is described as a range for that reason. Peer net-cash positions were not verified on a common basis, so peer enterprise values are not published here at all — the peer table shows market capitalisation only. Published market-research estimates for the Earth-observation market range from under $4bn to nearly $17bn for nominally the same thing, and we could not reconcile them; only the Novaspace figure is cited. Section 382 limitations on the accumulated deficit are unverified and the model’s four-year tax shield assumes they do not bind. Segment profitability is not disclosed, so the margin split between data subscriptions and satellite services is inferred rather than observed. |
Method. Unlevered discounted cash flow, methodology version 1.2, in the float form described in Section 02. Ten explicit years from FY2028 to FY2037, discounted mid-period at 10.5% in the base case, with a normalised terminal year taking maintenance capex at 8% of revenue, full 21% cash tax and the float growing only at the terminal rate. Free cash flow to the firm is adjusted EBITDA less stock-based compensation less capital expenditure less cash tax, plus the change in deferred revenue. Stock-based compensation is charged in full and never added back, which is the largest single difference between this model and the company’s own adjusted EBITDA. Cash tax is zero for four years against the $1.6bn accumulated deficit, then 21% on EBITDA less compensation less capital expenditure, with capital expenditure standing in for depreciation. The convertible notes are treated as equity throughout: the share count is 434.5m fully diluted and the full $865m cash balance is added back, which is the same basis on which the market strikes the enterprise value. The reference price used throughout is the $18.12 close of 4 September 2026, the last completed session before publication, and every multiple in this report scales directly with it.