EXPERIMENTALExperimental model. The model is driven by the order book and rests on a handful of judgements, above all the future book-to-bill, how much of each extra dollar of revenue becomes profit, and the multiple the market pays in five years. Change any of them and the numbers move a lot. Read them as scenarios, not forecasts. Not investment advice.
Planet Labs: the order book, priced
Planet photographs the whole Earth every day with a fleet of small satellites and sells the pictures, and what AI reads from them, mostly to governments and armies. The stock rose fivefold and then lost three quarters of it in 2026. We asked what today's price pays for, built from the order book up, the same way as our Theon model.
The price since 2021, and five years of simulated paths
Price (history)Median pathMiddle half of paths1 in 20 / 19 in 20
Source: NYSE close (Yahoo Finance), USD; model: planet-eval, pathsnxtdev.sk NXTlab
What Planet is
A San Francisco company founded in 2010 by three former NASA scientists, public since a 2021 SPAC merger. It flies the largest fleet of Earth-imaging satellites: the Dove fleet scans all land on Earth every day, and the newer Pelicans (high resolution) and Tanager (hyperspectral) are being launched now. Revenue is almost all subscriptions: 98% of contract value recurs.
The order book is lumpy. In fiscal 2026 (to January) the backlog jumped 79% to , mostly on a EUR 240m German deal for dedicated Pelican satellites. In the six months since, orders were only against of revenue, and the backlog fell to .
It has only just started making money. Adjusted EBITDA went from −10.6m to +15.5m to a guided +3m to +10m this year; with capex at a quarter of revenue for the new satellites, free cash flow is negative this year.
Cash is not the problem. in cash against of 0.5% convertible notes due 2030. They convert at USD 11.95, below today's price; a capped call absorbs the dilution up to USD 18.04.
Shares keep coming. Stock pay is about 15% of revenue, and a USD 1.5bn at-the-market programme from June 2026 still has about USD 1.38bn unused.
It moves like a lottery ticket: about a year. From USD in 2026 it fell to the low teens.
The model in seven lines
Every year on every path, in USD:
orders = book-to-bill × revenue
backlog = last backlog + orders − revenue
revenue = conversion × backlog at the start of the year
EBITDA = this year's EBITDA + incremental margin × extra revenue
cash = cash + EBITDA − capex + interest
EV/sales = K × (growth + FCF margin) in five years
price = (EV + cash − convertible) / shares, then into EUR
Organic growth in a steady state is (book-to-bill − 1) × conversion: 1.40 with 55% of the backlog delivered each year is about 22% a year, in line with management's "20%+" target. The exit multiple pays for growth plus cash margin together (the "rule of 40"), so a path that grows fast but never makes cash is valued like one that doesn't grow. Stock pay dilutes at each year's price. Returns are in EUR, with the euro moving against the dollar on its own path.
Input
Value
Where it comes from
Revenue: reported, then simulated
ReportedModel median, with the 5th–95th percentile
Source: Planet reports, fiscal years ending January (FY2027 = guidance midpoint)nxtdev.sk NXTlab
Where the business ends up
Median of the paths, with the 5th to 95th percentile in brackets. USD millions; fiscal years end in January.
Fiscal year
Revenue
Backlog
Adj. EBITDA
Free cash flow
Cash
Shares (m)
EV / sales
What it returns
Five-year return a year in EUR, every path
Return in EUR before tax; Planet pays no dividendnxtdev.sk NXTlab
Held for
Expected value a year
Median a year
Worst 1 in 20 (total)
Best 1 in 20 (total)
What the answer depends on
Each row changes one judgement and re-runs the paths. Five-year return in EUR, before tax.
Change
Expected a year
Median a year
Chance of a loss
Beats
Median max drawdown
The same paths at a different price
Entry price (USD)
Expected a year
Median a year
Chance of a loss
Beats
Strong and weak sides of this model
What it gets right
Revenue comes from orders. The next year is mostly the backlog already signed, so the near years are close to measured, and the rest is one number you can check every quarter.
Stock pay is a cost. It is paid in shares at each year's price, not added back the way adjusted EBITDA does.
The convertible is handled by its terms: debt below the USD 18.04 cap, shared upside above it.
EUR returns, with EUR/USD as its own random path.
What it cannot see
Working capital is left out. Governments often pay up front: in fiscal 2026 that turned a cash burn of about USD 66m (EBITDA minus capex) into USD 53m of free cash flow. The model's cash is therefore on the low side.
One mega-deal changes everything. A single EUR 240m contract made fiscal 2026. The yearly book-to-bill noise covers lumpiness, but the next big deal (or none) is the real risk either way.
Taxes are ignored (Planet has large tax losses to use first), and so is any acquisition.
The exit multiple is a judgement. K = 0.16 is roughly where ordinary software trades; today's price implies about 0.56.
What to watch
Q3 fiscal 2027 results (early December). Book-to-bill for the quarter. The model needs about 1.4 on average; the last six months were 0.6. One or two new satellite-services deals the size of Germany's would flip the answer.
The fiscal 2028 guidance (March 2027). Analysts expect revenue of about USD 570m (+30%). The model's median is lower; if the guide lands near 570m, the order book is refilling.
Use of the USD 1.4bn ATM. Selling shares at USD 13 instead of USD 32 costs existing holders far more per dollar raised.
Method
Sources: Planet Q4 FY2025 and FY2026 results, Q2 FY2027 results and 10-Q (Sep 2026), convertible notes releases (Sep 2025), ATM prospectus (Jun 2026), Investor Day targets (Oct 2025); NYSE/Yahoo Finance prices, shares and analyst estimates; EUR/USD. Tax: Slovak rules for a listed share held over a year (time test). Research, not investment advice.