Planet Labs

IThe business

Planet takes pictures of the whole land surface of the Earth every day and sells access to them. It makes its own satellites in San Francisco, buys rides to orbit from other companies, and runs them from the ground. It has one reportable segment.³ In the year to 31 January 2026, which Planet calls fiscal 2026, revenue was $307.7M, up 26%. Adjusted EBITDA, the company’s own measure, was $15.5M, its first profit on that measure. The net loss was $246.9M, of which $161.4M was a non-cash charge for warrants that gained value as the share price rose.³ I use “FY26” for that year and “FY27” for the year ending 31 January 2027 throughout.

There are two ways Planet earns money. The older one is a licence. A customer pays for access to Planet’s imagery and tools, mostly for a year or more, billed in advance every quarter or year. The same image can be sold to any number of customers, which is the company’s “one-to-many” point.³ The company says 98% of the value of its subscription contracts is recurring.³ The newer one is satellite services, which began to matter in January 2025. A customer, in practice a government or a company serving one, pays Planet to design, build, launch and operate satellites that the customer owns, or pays for dedicated use of satellites Planet flies. The money comes in on milestones, and Planet recognizes it as revenue as the work is done or when a satellite is accepted.⁴ On the owned-satellite deals Planet gets licensing rights to some of the images the customer’s satellites take, and counts those rights as part payment.⁴

The customers are mostly governments. In FY26 defence and intelligence customers paid $180.2M, 59% of revenue, civil government $71.9M and commercial customers $55.6M.³ In the quarter to 31 July 2026 defence and intelligence was $81.0M of $116.1M, or 70%, and up 93% on a year earlier.⁴ By region, North America was $132.0M in FY26, of which the United States was $123.9M, Europe, the Middle East and Africa $103.7M, Asia Pacific and Japan $59.8M and Latin America $12.2M.³ Ukraine alone was $35.9M, 12% of revenue, and Japan $38.0M.³ In the latest quarter Europe, the Middle East and Africa was $55.2M, almost half of the total, against $23.5M a year earlier.⁴

A few customers matter a great deal. Two customers were 13% and 12% of FY26 revenue, and in the quarter to July three customers were 17%, 11% and 10%, 38% together C (calculated).³⁴ One customer owed 33% of receivables in January and two owed 25% and 11% in July.³⁴ The filings do not name any of them. I will not guess.

What it flies

Planet has launched more than 600 satellites and says it operates over one hundred.¹²³ There are four kinds. SuperDoves, small satellites that scan the whole landmass each day at a resolution of about 3.5 metres, are the base of the subscription business. Pelicans take sharper pictures, to 50 centimetres after processing, and can be pointed at a place several times a day. Ten were in orbit after the July launch of Pelican-11, a test of a second generation the company says can reach the 30-centimetre class.³⁵⁸ Tanager sees about 400 colour bands at 30 metres and was built with NASA’s Jet Propulsion Laboratory for a charity-funded project that tracks methane. Tanager-1 launched in August 2024 and Tanager-2 has been shipped for launch.³⁵ Owl is a planned constellation of daily one-metre images.¹² The archive goes back to 2009, and daily scans to 2017, and averages more than 3,000 images for every point on land.³

Planet buys its rides to orbit. The two latest shipments to Vandenberg Space Force Base were for SpaceX’s Transporter missions, and Planet has also signed with Isar Aerospace in Germany to launch a Pelican.⁸⁵ The 10-K lists nine launch suppliers it could use, including Rocket Lab and SpaceX.³ I look at both on this site. SpaceX does most of the world’s lifting to orbit, and so is both a supplier and, in Planet’s own risk list, a possible competitor. Rocket Lab is also on the list, and I use it as the closest comparison in Valuation.¹³¹⁴

How the money works

The subscription side is cheap to serve once the satellites exist. Revenue from the same pictures can be sold to more customers, and customers pay ahead. The cost is the fleet. Capital spending was 26% of revenue in FY26 and is guided to about 25% in FY27.³⁵ Planet depreciates its satellites over 2.5 to 9 years, so the fleet has to be replaced, and it spent $81.5M on property, equipment and capitalized software in FY26 against $41.8M of depreciation and amortization C (calculated).³ The services side works differently. Planet builds hardware for another party, so the revenue is large, the margin is lower, and the timing is lumpy. In the latest quarter $14.1M of revenue was recognized at a point in time, mostly when satellite hardware passed to the customer, against $0.8M a year earlier.⁴ Gross margin was 57% in the quarter on the GAAP measure and 59% on the company’s non-GAAP measure, against 58% and 61% a year earlier, and the company expects 55% to 57% for the full year on the non-GAAP basis, against 59% in FY26.⁵³

The contracted pile, and what it hides

Planet reports two measures of work already sold. Remaining performance obligations (RPO) are the contracted revenue not yet recognized, counted only where the customer cannot walk away without a penalty. Backlog adds the cancelable part. At 31 July 2026 RPO was $753.1M and backlog $814.9M. Of the RPO, about 46% is expected to become revenue in twelve months and 68% in twenty-four.⁴⁵ That is $346M and $512M C (calculated). Backlog, which adds the cancelable part, puts 50% in twelve months and 70% in twenty-four, or about $407M and $570M C (calculated).⁵ The last twelve months of revenue were $378.3M C (calculated).⁴ Neither measure counts a subscription renewal that has not been signed yet, or usage-based work, so the second year looks thinner here than it will be. The first year is covered. For the second, the pile is a floor and not a forecast.

The direction matters more than the level. RPO doubled in FY26, from $412.8M to $852.4M, when Germany, Sweden and other work was signed, passing $672M in October 2025 on the way, and has fallen in each quarter since, to $816.0M in April and $753.1M in July.³²⁹⁸⁵ Revenue plus the change in RPO is a rough measure of what was signed in the period. It was $525M in FY25 and $747M in FY26. In the six months to July 2026 it was $111M, against $210M of revenue C (calculated).¹¹³⁴ It is rough because RPO leaves out usage-based and cancelable work, and revenue includes some of that, so I treat it as a direction and not a figure. But the direction is plain. In the first half of FY27 Planet delivered about twice what it signed. Large orders come in lumps, and the quarter to October may bring the next one. The results will tell.

The latest quarter

Planet reported the quarter to 31 July on 3 September, after the close. Revenue was a record $116.1M, up 58%. Adjusted EBITDA was $13.9M against $6.4M a year earlier, and the net loss was $9.4M against $22.6M.⁵ The company said it was the fourth quarter in a row at or above the “Rule of 40”, which is revenue growth plus adjusted EBITDA margin; growth of 58% and a margin of 12% make 70.⁵ Free cash flow for the six months was $21.3M. It also sold 3.8M shares through an at-the-market program for $122.4M before costs, about $32.36 a share, and received $107.8M from the exercise of warrants.⁵⁴ Cash and short-term investments were $865.4M.⁵ The stock closed at $18.35 that day and $18.12 the next.¹

The outlook was lower for the next quarter and higher for the year. Revenue for the quarter to October is guided to $101M to $105M, adjusted EBITDA to a loss of $1M to $6M and capital spending to $30M to $37M. For FY27 the company guides revenue of $430M to $441M, adjusted EBITDA of $3M to $10M and capital spending of $100M to $115M.⁵ In June the full-year capital guide had been $80M to $95M.⁸ If the quarter lands at its middle, the fiscal year guide leaves about $122M for the quarter to January C (calculated). The next report covers the quarter to 31 October. Last year the third-quarter report came on 10 December, and I found no date announced for this one by 9 November, so I treat December as expected and not confirmed.⁵²⁹

Guided against delivered

Planet has beaten its own quarterly guide twice. In the quarter to April, revenue was $94.2M against a guide of $87M to $91M, non-GAAP gross margin 56% against 49% to 51%, and adjusted EBITDA a loss of $1.0M against a loss of $3M to $6M. In the quarter to July, revenue was $116.1M against $102M to $107M, margin 59% against 52% to 55%, and adjusted EBITDA $13.9M against $0 to $5M. Capital spending was inside its range in April, at $18.0M, and above it in July, at $29.1M against $21M to $27M C (calculated).⁶⁸⁵ The full-year revenue guide has been raised at each report.

There is a second reading. After a quarter that ended about $11.6M above the middle of its range, the middle of the full-year revenue guide moved up only $2.5M, from $433M in June to $435.5M in September. That leaves the second half about $9M lower than the June guide implied C (calculated). The quarter to October is guided 11% below the quarter to July. Revenue recognized at a point in time, mostly satellite hardware, was $14.1M in the quarter to July against $1.8M in the quarter to April, a step of about the size of the beat C (calculated).⁴⁵ I read the beat as partly a delivery that landed early, and I cannot tell how much. The margin guide moved up by five points at both ends, which is real, and so is the capital spending guide, which rose by $20M at both ends.

What was won since January

The three large contracts are in the table above. Since then the company has named these awards. Most are small or last a year.

Against a fall of $99M in remaining performance obligations in six months, these are small. The company called its August wins with the NGA and the German government “landmark”.⁵ The sized ones add to $37.4M plus up to €25M over five years, and one more is “eight figures”, which could be anywhere from $10M to $99M. I find no new contract of the scale of JSAT, Germany or Sweden since December 2025. That is the gap the first row of the table in Section VII tracks.

Where Planet is from

Three NASA scientists founded it in 2010.⁵ It listed on 8 December 2021 by merging with a blank-check company, dMY Technology Group IV, which brought $590.4M of gross proceeds including a $252.0M private placement.⁷¹ It is a Delaware public benefit corporation, so the board must weigh a stated public purpose, to illuminate environmental and social change, alongside profit. It has headquarters in San Francisco, about 1,000 employees in 29 countries, and offices in Berlin, where it plans to start making satellites this year, and London.³⁵

IIThe moat

Planet has things that are hard to copy and nothing that stops a determined rival. I think it helps to keep the two apart.

What is hard to copy

The first is the archive. Planet has scanned the whole land surface every day since 2017 and holds imagery back to 2009, averaging more than 3,000 images of every point on land.³ You cannot go back and take yesterday’s picture. For anyone who wants to train a model to spot change, or test one against what happened, that history has value that grows with age.

The second is the fleet and the line that makes it. Planet has launched more than 600 satellites and builds them in one factory in San Francisco, with a second planned in Berlin.¹²³⁵ A rival has to match the cost, the cadence and the software that runs them. The FY25 10-K warned that a rival could build a larger fleet, partly because Planet might keep its own at about its then size for the near term. The FY26 10-K no longer says so, and the fleet is growing again.¹¹³

The third is permission. Planet needs licences from the National Oceanic and Atmospheric Administration to operate its satellites, and a facility clearance for classified work in its US federal subsidiary.³ These slow a newcomer, though not an established rival.

The fourth is the sovereign contract itself. When a government owns satellites that Planet built and flies for it, the data pipes, the ground stations and the staff are tied together, and replacing that is a project. That is my inference from how these contracts are described, not a finding in the filings E (estimate). It would be proved by a renewal, and none has come due.

What is not protected

No law reserves any part of this market for Planet. Nothing stops a government buying from a prime with a longer record, and nothing makes a commercial customer stay. The signs are mixed. Net dollar retention, the share of last year’s contract value that is still there plus any growth, was 131% in FY23, 101% in FY24 and 116% in FY26.¹⁶¹⁵³ The number of customers fell from 1,018 to 897 in two years, which the company puts down to focusing on bigger customers, and it will no longer report it.³ Civil government and commercial revenue was $126.1M in FY24, $128.1M in FY25 and $127.5M in FY26, flat for three years C (calculated).³ All of the growth has been defence and intelligence. The latest 10-Q credits large expansions by existing customers and, in part, deliveries of satellite hardware.³⁴

The risk that worries me most is SpaceX, and Planet’s own filing says the same. SpaceX launches its own constellations, and the 10-K says its vertical integration of launch and satellites could let it offer better or cheaper services if it chose to compete directly.³ The filing frames it as a possibility and I have not tried to judge how likely it is. On my SpaceX page I count 165 Falcon launches in 2025 against 18 for Rocket Lab.¹³ If a company with that position decided to offer pictures, Planet would be competing against its own launch supplier.

IIIThe competitive advantage

The last section says where Planet is protected. This section asks what it does better than the people it competes with, and whether the numbers show it.

Speed, and who pays for the satellites

Planet’s claim is speed. It calls the method “agile aerospace”: small satellites, built in volume with an in-house design, launched often, with software that runs them with few people.³ The best evidence is Sweden. The contract was signed in December 2025 and a Swedish satellite was in orbit in May 2026, just over four months later.⁸ The Swedish Armed Forces now own a reconnaissance satellite Planet built, and Planet is licensed to offer some of what it sees to others.³ That is the second part of the advantage. In a normal fleet Planet pays for every satellite and sells the pictures. Here the customer pays for the satellite, Planet earns a margin building it, and Planet gets image rights as part payment, which it records at fair value.⁴ If it works the way the company says, Planet’s fleet grows faster than its own spending. The German deal may point the same way. Its release says the capacity includes satellites “being built for our previously announced satellite services agreement”, and the only such agreement announced by then was JSAT’s, so Germany may be paying for use of satellites that JSAT paid for. That is my reading and the release does not say it E (estimate).²⁸¹²

I can check that against the cash. It does not show up yet. Capital spending was $81.5M in FY26, 26% of revenue, up from 20% the year before, and the company expects about $108M in FY27, and says the rise is for its own next-generation Pelicans and medium-resolution satellites.³⁵ The customer-funded satellites sit in cost of revenue, which is why gross margin is lower. In a business where the customer pays for the hardware, spending on the company’s own fleet would be expected to fall. Planet is still building its own.

One-to-many, and what it does to margins

The subscription business is where the one-to-many model pays. A picture taken once is sold again and again. The signs are there: non-GAAP operating income was $4.1M in the latest quarter against a loss of $3.3M a year earlier, because non-GAAP operating costs rose 33% while revenue rose 58% C (calculated).⁵ R&D was $106.7M in FY26, 35% of revenue.³ Growth in costs lagging growth in revenue is what a software-like business looks like.

The gross margin tells a different story. It was 65% on the company’s non-GAAP measure in the quarter to January 2025 and 56% to 60% in the last four quarters. That fits a mix that is moving toward building satellites for others.⁶⁸⁵²⁹ The company guides to 55% to 57% for FY27, which is five points above the 50% to 52% it guided in March, so this year the margin has come in above plan even as the mix moved toward hardware.⁶⁵ Satellite hardware is low-margin work compared with licences. The more of the next five years comes from it, the lower the margin on each dollar.

Where Planet is ahead, and where it is not

• Ahead on cadence. Ten Pelicans are in orbit, and a Tanager and 18 SuperDoves were shipped to launch in early September. Planet launched 40 satellites in FY26.⁶⁵
• Ahead on data. Daily global coverage with an archive to match. Nobody else sells that, as far as I can tell from the 10-K’s list of rivals.³
• Behind on resolution at the high end. The 10-K says incumbents run very high-resolution satellites at very high cost, and Planet’s Gen-2 Pelican, aimed at 30-centimetre class, has only flown as a test.³⁸
• Dependent on launch. Planet buys launches. The recent launches named in its releases were on SpaceX rockets. Its 10-K lists nine possible suppliers, and it has signed with Isar Aerospace for a Pelican.³⁸⁵ A failure or delay does not just cost a satellite; it can delay revenue booked on delivery.⁴
• Dependent on a few buyers. Three customers were 38% of the last quarter’s revenue.⁴

Putting it together: the advantage is real in speed and in the archive. The pricing power is unproven. I would want to see the gross margin stay near 58% as satellite revenue grows and the civil and commercial business start to grow again before I treated the advantage as durable. Neither has happened yet.

IVManagement

Who runs it

Two co-founders run Planet. Will Marshall has been chief executive since 2011 and is chairperson. He holds a doctorate in physics from Oxford and worked at NASA before founding the company in 2010.¹⁰ Robbie Schingler, a co-founder, is chief strategy officer and a director; he has held the chief operating and chief financial roles earlier in the company’s life.¹⁰ Ashley Johnson has been chief financial officer since February 2020 and president since March 2024; before that she was chief financial and operating officer at Wealthfront.¹⁰ The board has nine directors, seven of them independent. They include the chief executive of Ciena, a former chief legal officer of Twitter and a retired general who was the first chief of space operations of the US Space Force. The two founders are the only directors who are not independent.¹⁰

The founders also control the company. Each founder holds 11,746,898 Class B shares with 20 votes each, and together they have about 58% of the votes: their 23.5M Class B shares carry 470M votes against 340M for all the Class A, on the August count C (calculated).¹⁰⁴ The 10-K still says about 62%, which looks out of date.³ In July 2026 the three directors up for election won, though Vijaya Gadde had 72.1M votes withheld and General John Raymond 41.7M.¹⁷ The Class B shares convert to ordinary shares on most transfers, on a founder’s death or incapacity, on the tenth anniversary of the December 2021 merger, or six months after a founder stops working for the company.³ So control lasts about five more years unless a founder leaves or sells first.

How they are paid, and what that pays them to do

Mr Marshall’s reported pay for FY26 was $4.63M: $450,000 of salary and $4.18M of stock awards. Ms Johnson’s was $2.54M and Mr Schingler’s $1.69M.¹⁰ Most of it is stock. Each year they receive restricted stock units that vest every quarter over four years, with no performance test: in FY26, 946,970 units for Mr Marshall, 511,153 for Ms Johnson and 315,986 for Mr Schingler.¹⁰ They also own options, from the days before the merger: 4.49M for Mr Marshall, 2.03M for Ms Johnson and 1.32M for Mr Schingler were exercisable in May.¹⁰

The annual bonus is the part that sets what they chase. It pays on two measures: GAAP revenue for 60% and adjusted EBITDA for 40%, and the executives can take it in shares. For FY26 the targets were $280.0M of revenue and an adjusted EBITDA loss of $7.0M. Actual revenue was $307.7M and adjusted EBITDA was a profit of $15.5M, so the committee paid 125% of target, the maximum.¹⁰ Three things follow. First, revenue counts whether it comes from a licence or a satellite built at a thin margin, so the plan pays the same for either. Second, adjusted EBITDA leaves out stock pay of $55.0M and leaves out capital spending of $81.5M.³ Third, a target below zero was beaten by a profit. The plan is a fair test of whether the company met its own budget. It does not test whether an owner made money. That is my reading of the definitions, not a statement by the board E (estimate).

How much of their own money is in

The proxy shows large stakes. Mr Marshall owned 17.3M shares including options and Class B, worth $248.8M at $14.37, and Mr Schingler 13.5M, worth $194.2M.¹⁰ Those are ownership figures from May. Since then the Forms 4 show open-market sales. Between 2 April and 17 September Mr Marshall, Ms Johnson and Mr Schingler sold 981,207 shares for $28.4M, an average of $28.93 a share, with prices from $16.96 to $35.22 C (calculated).⁹ That is about 3% of what the proxy showed them owning. Two other directors sold $3.3M of stock between April and October, and a third $0.3M. Altogether six insiders sold 1.13M shares for $32.0M C (calculated). In the 21 months I read, the only purchase was 20 shares by a director in October 2025.⁹ Sales on a plan prove little, and three directors did file trading plans in January 2026, for 71,000, 117,107 and 32,468 shares, which Planet disclosed only in an amendment in June.¹⁸ But these were sales by the people who know the business best, at an average price twice today’s.

One more holder matters for the share count. Alphabet Inc.’s 13F, filed on 6 November 2026 for the quarter ended 30 September, lists 35,248,893 Planet Class A shares, worth $577.7M at the quarter-end price of $16.39. Its June report listed the same number. The proxy lists Google LLC as the only holder of more than 5% of the Class A stock apart from the founders, at 10.59% in May.¹⁹²⁰¹⁰ I do not know why any holder owns what it owns, and I do not infer it.

Planet also has contracts with Google LLC, which the proxy lists. It purchases cloud hosting under a contract that commits it to $193.0M through January 2028, of which $48.0M remained at 31 July, and spent $16.1M on it in the six months to July.¹⁰⁴ In October 2025 it agreed to do research and development work on a payload carried on prototype satellites, with funding paid on milestones; $9.5M had been received by July.⁴ A 3% fee is payable on any imagery sold through its marketplace listing.¹⁰

What they did with capital over a full cycle

The company raised $590.4M gross when it listed, including $252.0M from a private placement, and $460.0M in September 2025 from convertible notes.⁷³ In the six months to July it received $107.8M from warrant exercises and, in the second quarter, $122.4M gross from selling stock at the market.⁴ Against that, from FY20 to FY26 it lost $0.88B at the operating line, burned $340M of free cash flow C (calculated) and paid $297M of its staff in stock C (calculated).²¹ Its shares outstanding went from 267.9M in April 2022 to 363.8M in August 2026, up 36% C (calculated).⁷⁴ The merger agreement gave former holders up to 27M more shares if the price reached $15, $17, $19 and $21; the price did, and about 22.9M were issued in January and February 2026 C (calculated).²² In FY26 Planet also paid $72.7M in cash to cover tax on employees’ vesting shares, withholding the shares instead of issuing them, which is more than its free cash flow that year.³ That payment works like a buyback of shares it had just created.

I can’t tell skill from luck in that record. The company listed by merging with a blank-check company. It got through FY25 without raising money, cut its operating loss by 46% from the FY23 level, and in FY26 turned cash flow positive on prepayments. It sold stock at about $32 in the summer, a price it has not seen since, and issued notes in September 2025 that convert at $11.95 when the stock was near $9. A company that sells shares when the price is high is acting in its owners’ interest, as long as it does not spend the proceeds poorly. The test is what it does with the cash. In the six months to July it spent $47.1M on capital and ran an operating loss of $48.4M. Its filings say it may need to raise more.⁴¹

The original 10-K left out the three directors’ trading plans and an exhibit setting out the clawback policy, both added on 5 June.¹⁸ The court matters are small beside the numbers above, and are in Risks.

VFinancials

This is seven years, not ten. Planet was private until December 2021, and I opened no filing that shows earlier statements. The first seven years are mostly a story of losses; the last is the first with positive cash from operations, and that cash came from customers paying early.

Revenue grew 21% a year over six years, from $95.7M to $307.7M. The gross margin on the GAAP measure went from negative to 56%, helped as old satellites became fully depreciated and depreciation fell from $77.6M to $41.8M. The operating loss narrowed from $175.7M in FY23 to $95.1M in FY26 while operating expenses stayed near $268M, so the improvement came from gross profit. Adjusted EBITDA turned positive in FY26.²¹³

The quarters show where the growth came from. Revenue was flat at about $61M a quarter through FY25, then rose to $66M, $73M, $81M and $87M in FY26, when defence and intelligence revenue rose by $64.0M over the year, and to $94M and $116M in FY27.³ The company guides to a fall to about $103M in the quarter to October, then a recovery to about $122M in the quarter to January C (calculated).²¹⁵ The 98% recurring figure covers subscription contracts and leaves out satellite services. Satellite hardware moves in steps.

What an owner could really take out

Reported free cash flow for the twelve months to July 2026 was $19.9M. That is the company’s own measure and it is positive. I take three things away from it. Stock pay was $62.5M, a real cost paid in shares. Customers prepaid $122.4M more than they used, mainly for satellites not yet built; that cash is theirs until Planet does the work, so I don’t count it as income. And I put back the part of capital spending that goes beyond standing still. I think about $60M a year keeps the fleet and ground stations at their present size, since depreciation and amortization was $41.8M in FY26, of which satellite depreciation was $28.4M, and the next Pelicans cost more to build than the first ones; actual spending was $97.8M, so about $37.8M was for growth E (estimate).³ The answer is owner cash of −$127.3M.

That is the gap between how the business looks and what it is: adjusted EBITDA of $20.8M, free cash flow of $19.9M, and what an owner could take out, −$127.3M C (calculated). I don’t think the last is permanent. Customers who prepay while Planet builds fund the work, which is a real source of cash for a growing business, but it can reverse in a year. In the first half of FY27, of $68.4M of operating cash flow, $47.1M was extra prepayments and $23.9M was receivables collected; without both it would have been about −$2.6M C (calculated).⁴

Quality of earnings

• The warrants. The FY26 net loss of $246.9M includes $161.4M for warrants that rose in value with the share price, and the first half of FY27 $106.5M more. Planet redeemed the public warrants on 27 April 2026; holders exercised 9.37M of them for $107.8M, and none remain, so the charge will not come back.³⁴²³
• Stock pay. Stock-based compensation was $55.0M in FY26, 18% of revenue, and $33.5M in the first half of FY27. Adjusted EBITDA adds it back. Without that one add-back FY26 would show −$39.5M C (calculated).³⁴ Unrecognized cost on stock units already granted is $149.5M, to be expensed over about 2.2 years.⁴
• Prepayments. Deferred revenue rose $151.1M in FY26, which is more than operating cash flow of $134.4M. Without it, operating cash flow would have been −$16.7M C (calculated).³
• Smaller items. FY26 adjusted EBITDA also leaves out $11.2M of litigation costs and $2.5M of employer payroll tax on earn-out shares.³ Interest income was $11.6M in six months of FY27; I count the cash itself in the value and not the interest as well.⁴
• Tax. Federal losses carried forward are $967.3M, of which $708.1M do not expire, with $603.6M more at state level and $41.9M of federal credits. They will keep cash tax near zero in the United States for years if profits arrive, and I count that in the value.³
• Leases. Operating lease liabilities rose from $15.6M to $50.4M in six months, after $38.1M of new leases. Rent is inside operating costs, so I add no debt for it.³⁴

Capital spending is the other half of the story. For five years, FY20 to FY24, Planet spent less than its depreciation, which flatters free cash flow because the fleet was getting older. In FY25 and FY26 it spent more, and the FY27 guide rose from $80M to $95M in June to $100M to $115M in September, as the company builds its own Pelicans and medium-resolution satellites.²¹³⁸⁵ Accumulated depreciation is 61% of the cost of everything it owns, $238.0M of $388.6M C (calculated), so a lot of the fleet is old.³

The share count is the third thing. The 363.8M shares outstanding on 27 August were 36% more than in April 2022 and 5% more than in March.⁷³⁴ On top of that are 13.5M options at an average $5.80, 1.07M warrants at $9.38, 23.9M restricted stock units that vest as people stay, and 38.5M shares that the notes would produce if converted.⁴ The program for selling up to $1.5B of stock at the market, signed in June, had $1,377.6M of capacity left on 31 July. At today’s price that is about 96M shares, 26% of those now outstanding C (calculated). The company is not obliged to use any of it. It sold 3.8M in the quarter to July, at about $32.36 gross.⁴²⁴ I come back to this in Valuation, because the price at which the company sells matters as much as the amount.

The balance sheet

On 31 July 2026 Planet had $415.1M of cash and $450.3M of short-term investments, $865.4M together, plus $8.5M of restricted cash I leave out. It owed $460.0M on 0.50% convertible notes due 15 October 2030, carried at $448.3M, and had no other borrowing.⁵⁴ That is net cash of $405M C (calculated). Deferred revenue was $298.6M, equity $564.4M and the accumulated deficit $1.6B.⁵ The notes convert at $11.95 into 38.5M shares, and Planet bought capped calls for $39.6M that offset dilution up to $18.04. Holders could convert in the quarter to October because the stock had been above $15.54 for enough days. By my count only 16 of the 30 trading days to 30 October closed at or above $15.54, short of the 20 needed, so that right looks to have lapsed for the quarter to January unless another condition is met C (calculated). The October 10-Q will say.⁴¹ At $14.37, between the conversion price and the cap, the capped call covers the in-the-money value, so I count the notes as $460.0M of debt and no extra shares. Above $18.04 they would add shares at a cost the call does not cover, and I model that.

VIValuation

First, how this goes badly wrong

Before the numbers I asked what would most likely hurt an owner at this price, and worked back. Three things stand out.

New orders do not replace deliveries. In the six months to July 2026 Planet recognized $210M of revenue and signed, by my rough measure, $111M. The contracted pile covers $346M of the next twelve months of revenue and $166M of the twelve after that; counting cancelable work it is $407M and $163M.⁴⁵ If the next large contracts do not come, revenue falls back toward the $380M to $400M that the subscription business and what is left to build can support, while the staff and spending added for growth stay. That is my bear case. It cuts the value from $2.52 to $0.43 E (estimate).

The cost of standing still and stock pay is bigger than the profit. Stock pay is about $68M a year and I think keeping the fleet at its present size takes about $60M. The business needs adjusted EBITDA of $130M or more before it has paid for either, and the FY27 guide is $3M to $10M.⁵ Each five points of adjusted EBITDA margin, every year, moves my base value by about $1.10 a share; five points less takes it to $1.40 and five more to $3.60 E (estimate).

A few customers, and a war. Three customers were 38% of the last quarter’s revenue and Ukraine was 12% of FY26 revenue.⁴³ At the Q2 rate of $464M a year, a customer with 17% of revenue is about $79M a year C (calculated). Government budgets and the course of a war are not things I can forecast.

None of these needs anything exotic. They set the bear case, and they are why I begin with the cash an owner can take out and not with adjusted EBITDA.

What an owner could really take out

I value cash after stock pay and after the capital needed just to stand still. For Planet that number is negative, and it is a long way from the reported figures. Over the last twelve months free cash flow was $19.9M and adjusted EBITDA $20.8M. Owner cash was −$127.3M E (estimate). The gap is stock pay of $62.5M, customer prepayments of $122.4M, and an add-back of $37.8M for growth spending.

I don’t think −$127M is where it stays. If Planet keeps winning contracts, the prepayments keep coming and the model below counts them in the years they happen. In my base case the balance falls from $299M to $270M by January and then drifts up with revenue, so I assume no repeat of the last year’s $122M. Owner cash turns positive in FY32, and the cash burned before then, about $460M, is 53% of today’s cash balance E (estimate).

What the price implies

Put the other way round: what does $14.37 need? On today’s diluted share count the market value is $5.35B. Planet has net cash of $405M after counting the notes as debt, so the business is priced at $4.94B, 13.1 times the last twelve months of revenue. At a 12% discount rate, my base margins and spending, the price needs $7.3B of revenue in FY41, growing 22% a year for fourteen years from the FY27 outlook. My base has $1.6B and my bull $3.5B. The other way to read it is that my bull case is worth the price at a discount rate of 9.7%. The price therefore asks for the bull case and for an investor content with a return below 10%.

What I can’t judge

• How much of the revenue is subscription and how much is satellite building. Planet does not split it. The 98% recurring figure excludes satellite services, and the customer-type table mixes both.
• What a satellite contract earns. The gross margin tells me the mix is lower margin than licences. It does not tell me by how much, and the company gives no contract margins.
• What is in the pipeline. The company says the pipeline expanded. The filings give me no number; the awards it names are in Section I.
• Who the large customers are, and when they renew. The filings name none.
• What the next Pelicans and Owl cost, and how long they last. My $60M a year to stand still is a judgement between $45M and $85M.
• Whether the notes convert, and what the board does with $1.38B of capacity to issue stock.
• Politics. Defence budgets, export rules and the war in Ukraine.

Because I can’t know these, I show a range and where the price sits in it, and I don’t treat $2.52 as more exact than it is. A fair reading of my work is that the base is somewhere from $1.40 to $5.40 a share. Using the top of the FY27 revenue guide, $441M, in place of the $435.5M middle moves the base from $2.52 to $2.57 E (estimate).

The discount rate

The ten-year Treasury yield was 4.88% on 9 November, and Damodaran’s implied equity risk premium is 4.20%.²⁵²⁶ Planet’s weekly beta against the S&P 500 ETF is 2.01 over five years, 2.62 over three and 2.50 over two C (calculated).¹²⁷ On the five-year beta the cost of equity is 13.3%; on the others 15.4% to 15.9%. I use 12.0%, a little below the formula, because a beta measured through a rise from $3 to $51 and a 72% fall overstates how risky a company becomes once most of its revenue is from governments. It is above the 10% on most of my pages because satellites fail and contracts arrive in lumps. I use 13% for Rocket Lab and 10.5% for SpaceX.¹⁴¹³ At 10% my base is $3.56 and my bull $13.46; at 14% they are $1.89 and $7.21.

Three cases

Bear

New orders stop replacing deliveries. Revenue slips 5% in FY28 and 3% in FY29 to $396M as the satellite contracts are worked off, then grows 2% to 4% a year. Civil and commercial revenue stays flat, as it has for three years. Management cuts stock pay from $60M toward $35M and capital spending to 11% of revenue by FY30, and the adjusted EBITDA margin climbs to 14% in FY30 and 22% from FY36 (it was 5% in FY26). Owner cash stays negative through FY32, a burn of about $510M, and $474M of it comes before January 2030. The $460M of notes fall due in October 2030, so the company would have to refinance, sell shares or cut harder, and I model none of that. Value $0.43: the $405M of net cash, less about $430M of burn in today’s money, plus about $180M for the business that is left E (estimate).

Base

Planet signs about as much as it delivers. Revenue grows 20% in FY28, 15% in FY30, 10% in FY33 and 7% in FY36, reaching $1,238M in FY36 and $1,595M in FY41. The margin climbs to 18% in FY30 and 30% in FY36, capital spending falls from 22% to 10% of revenue, and stock pay grows 3% a year. Value $2.52.

Bull

Sovereign contracts arrive faster than they are delivered. Revenue grows 35%, 30% and 25% in FY28 to FY30, reaching $967M, then 17% in FY33 and 11% in FY36, ending FY41 at $3,510M. The margin reaches 22% in FY30, 36% in FY36 and 38% in FY41, and capital spending falls to 9%. Value $9.54.

Other uses of the same dollar

The cleanest alternative is a Treasury note. Ten-year paper yielded 4.88% on 9 November with no risk of the kind above.²⁵ At $14.37 my base cash flows return 5.2% a year, 0.3 points more than a Treasury, for a business that has to keep winning contracts, and my bull returns 9.7% against 12% that I ask E (estimate). The best comparison on this site is Rocket Lab, which also builds spacecraft for governments, also pays staff in stock and also trades far from my values. At its price my Rocket Lab base returns −2.9% a year and its bull 5.1%, against 13% that I ask there.¹⁴ SpaceX is the other space company I have valued: its base returns 6.4% and its bull 12.6% against 10.5%.¹³ By these numbers the shortfall from my bar is 4.1 points for SpaceX, 6.8 for Planet and 15.9 for Rocket Lab. That ranks them. It does not make the middle one cheap.

Planet is also closer to Rocket Lab in size than to SpaceX. Its enterprise value is about 13 times past-year revenue, against about 44 times for Rocket Lab on its page.¹⁴ Planet is the cheaper stock per dollar of revenue, and that is a fact about the price. It does not say the revenue is worth more.

What the market paid before

The stock was $12.81 a year ago, and is up 12% since. In between it reached $51.40 on 28 May 2026 and fell 72% to now.¹ It is up 4.9 times from the $2.96 close of 21 April 2025. The market has paid very different prices for a dollar of Planet’s revenue. At the end of January 2024 it paid 1.6 times, at the end of January 2026 it paid 27.5 times, and now it pays 12.8 times, on basic shares C (calculated). Revenue is up 71% on the January 2024 year and the multiple is up about eight times, so the price has moved more on what people expect than on what Planet has earned.

The financing and the value per share

The at-the-market program changes the arithmetic, and it can help or hurt. If Planet sold the remaining $1,378M at $14.37 it would issue about 96M shares and take in about $1.35B after costs of 2%. On my base case, where a share is worth $2.52, receiving $14.37 for a share is a gain for the existing owners: the base value per share would rise to $4.93 and the bull to $10.46 E (estimate). That is arithmetic about value per share. It says nothing about where the price goes. It also cuts both ways. A share sold at $5 would add value if my base is right and take it away if my bull is right. The board sold 3.8M shares at about $32 in the quarter to July. Class A shares rose by only 190,361 between 31 July and 27 August, so any sale in August was small, and I have no figure after that. Whether it uses the rest will tell me more about its view of value than its press releases.

The margin of safety

I set it from the bear case: I want a price at which my $0.43 bear value is at least three quarters of what I pay, which is $0.57. Net cash is $1.09 a share C (calculated), so at that price someone would pay less than the cash after the notes and get the business for nothing. The price has never been that low. The lowest close since the first day of trading was $1.69 on 30 April 2024.¹ I am not saying the price will fall to $0.57. I am saying where the margin of safety begins.

Two different questions

The first is how good the business is. By the signs it is improving: revenue grew 26% in FY26 and 51% in the six months to July, retention was 116%, and it moved from losing cash to a small free cash flow profit in a year.³⁴ Against that, it has lost $0.88B at the operating line since FY20 and has not yet earned a cash return on what has been put into it. The second question is how good the price is. At $14.37 the price is 470% above my base and 51% above my bull, and the market is paying for several things going right at once. A good business can be a poor purchase when the price needs the best outcome. This looks like that case, and the market may be seeing something I can’t: that sovereign demand for owned satellites is a new market and Planet is the one company with the factory, the launches and the archive. I have tried to show what that would have to be worth.

VIIThe case for the price

What most people miss

Three facts cut in Planet’s favour and get little attention. First, on the owned-satellite deals the customer pays for the satellites and Planet is licensed to offer some of their imagery to others, which it counts as part payment.³⁴ Second, speed: Sweden signed in December 2025 and had a satellite in orbit in May 2026.³⁸ Third, the largest region is no longer North America. Europe, the Middle East and Africa was $55.2M of $116.1M in the quarter to July, against $39.0M for North America.⁴

The bull case

I found no named analyst or investor with a view on Planet, published after the September results, that I could check against a dated document, so I give the case as the company makes it. On 3 September Mr Marshall said the strategy pairs AI-enabled analytics with sovereign satellite services, and that the pipeline had grown. Ms Johnson said non-GAAP gross margin beat expectations.⁵

In plain words: governments in Europe and Asia want satellites of their own, in months, and Planet is one of very few companies that can build, launch and run a constellation while keeping the right to license the pictures. The company adds that artificial intelligence makes the archive searchable in plain language.⁵ The record supports some of this. Planet beat the top of its quarterly revenue range twice this year, raised its full-year revenue guide at each report, and raised its gross margin guide by five points.⁶⁸⁵ Revenue from Europe, the Middle East and Africa went from $23.5M to $55.2M a quarter in a year, and the company signed three large satellite services contracts in the twelve months to January 2026.⁴¹² The $865M of cash is more than the roughly $460M my base burns before owner cash turns positive.

Against reality

If Planet signs sovereign satellites as fast as it delivers them, keeps a 58% gross margin, and grows costs at half the rate of revenue, the price is possible, for an owner content with under 10% a year. That is my bull case. It is worth $9.54 at 12% and $13.46 at 10%, and the price is its value at 9.7%. Today’s numbers are far from it. It needs revenue of $967M in FY30, 2.2 times the FY27 guide; orders at least equal to deliveries, where the first half of FY27 signed about half; an adjusted EBITDA margin of 22%, where FY26 was 5% and the FY27 guide is 1.5%; and owner cash after all capital spending of $40M in FY30, where the last twelve months were −$165.0M on the same basis E (estimate).

It helps to multiply it out. The price needs six things: orders at least equal to revenue for three years, about 35%; a gross margin near 58% as satellite work grows, 65%; costs growing at half the rate of revenue, 65%; capital spending down to 12% of revenue by FY33, 70%; no loss of a large customer, 80%; and funding on fair terms, 85%. Multiplied, that is about 7%, or one chance in 14. These are my guesses, they are not independent, and I do not use them in the values E (estimate).

What a 10x would take. Ten times $14.37 is $144, about $53B of market value on today’s diluted shares, before the new shares such a price would bring. At 25 times owner cash that needs about $2.1B a year. At a 25% owner cash margin, above anything Planet has earned and above my bull case in FY41, that is about $8.6B of revenue, 20 times the FY27 guide. By FY36 that means growth of about 39% a year for nine years, and by FY41 about 24% a year for fourteen C (calculated). Planet’s revenue has grown 21% a year for six years, and its best year was 46%, in FY23.²¹ Tenfold moves in this stock have happened, but on sentiment: it rose 17 times in 13 months, from $2.96 on 21 April 2025 to $51.40 on 28 May 2026, and is down 72% since.¹ On my Rocket Lab page the ten-year exceptions I name are NVIDIA, up 115 times, and Tesla, up 30 times.¹⁴ For a company with $380M of revenue and no cash profit to repeat that is not impossible, but it is rare, I have not measured how rare, and I would not build a case on it.

Where I come down

Three things moved me: a first year of positive adjusted EBITDA and free cash flow, a satellite flown for a government four months after signing, and $405M of net cash. The strong quarter did not move my base: delivering work signed earlier does not show that more is coming, the contracted pile fell, and the second half is guided about $9M below what June implied. What would raise my values is evidence that Planet signs as much as it delivers while the gross margin holds near 58%, with obligations back above $850M or owner cash positive before FY32. Until then the price sits above my bull case.

VIIIRisks, and what would change my view

The three ways this goes badly wrong are at the start of Valuation. These are the rest.

• Fixed-price building. Satellite services are fixed-price contracts, with revenue recognized by the share of cost incurred. If a build costs more than planned, the company records a catch-up adjustment in one quarter. The 10-K names the estimate as critical.³
• Government budgets and cancellation. The cancelable part of contract value rose from $48.0M to $61.7M in six months. US government contracts can be terminated for convenience, and the 10-K lists shutdowns and funding changes as risks.⁵³
• Export and licences. Planet needs approvals from US regulators to operate and to supply some governments, and says it cannot be sure they will continue.³
• Launch and failure in orbit. A launch failure takes the satellites with it, and a satellite that fails early shortens the depreciation life. Revenue booked on delivery waits for the satellite to be accepted in orbit.³⁴
• Control. The founders keep about 58% of the votes until 2031 and can decide most matters put to a vote.⁴³
• Litigation and tax. A Delaware class action settlement in principle is awaiting approval and Planet is covering the former officers’ costs; the IRS is auditing the year to January 2024.⁴
• Insiders. Six insiders sold 1.13M shares for $32.0M between April and October and none bought in size.⁹