Planet Labs
Planet Labs
One-line summary: Public Earth-imaging company (CEO/co-founder will-marshall) running the largest daily-refresh satellite fleet; re-framing itself from a satellite-imagery vendor into the data layer for "large earth models" — and a worked example of venture-scale value captured in the public market (~$5 → ~$50 over 12 months).
What it is
Operates ~200 Earth-imaging satellites that photograph the entire Earth every day, with a full time-series archive ("every day going back"). will-marshall frames it as "the Google satellite layer on Google Maps that you can look at, except it's today's date rather than three years old." Customers span agriculture, energy, civil government (flooding/fires/permitting), and defense/intelligence. Went public in 2021 via SPAC at a ~$2B valuation. From will-marshall in 2026-06-06-podcast-all-in-podcast-the-ipo-comeback-why-tech-giants-are-finally.
Why it matters to stock-market
Planet is tracked here on three distinct legs, each tradeable as PL:
- An AI-data moat play. Marshall's thesis is that LLMs are "blind" to the real world and "AI is only as good as the data it's trained upon" — so the unique daily-global Earth archive becomes a scarce training/inference input for "large earth models." This is the canonical chain: ai-real-world-data-gap-to-planet-moat. Marshall sizes the Earth-observation market at "$75 to $100 billion."
- A named entrant in the orbital-data-center race. Planet is partnering with Google to launch TPUs into space and has already launched some Nvidia GPUs into space on an early test — making it a non-SpaceX player in terrestrial-power-flat-to-orbital-dc-arbitrage.
- The worked example of public-market value capture. Planet's ~10x ($5 → $50) happened in the public market over ~12 months, with "90% of the value created" in years three and four post-listing — the central evidence for ipo-comeback-public-market-value-capture. It is also a scarce listed comp that ipo-comp-anchor-repricing expects SpaceX's S-1 to re-anchor.
Key facts
- Fleet: ~200 satellites, images the entire Earth daily; full historical time-series archive. From will-marshall in 2026-06-06-podcast-all-in-podcast-the-ipo-comeback-why-tech-giants-are-finally.
- Revenue mix: security/defense is "about 60% of our revenue today" — a larger fraction than originally guessed, driven by the geopolitical demand for "advance warning" intelligence. From will-marshall.
- Stock move: ~$5 → ~$50 per share over ~12 months (~10x); "90% of the value was created" in years three–four after the 2021 listing. From david-friedberg and will-marshall in 2026-06-06-podcast-all-in-podcast-the-ipo-comeback-why-tech-giants-are-finally.
- Cap table: early VCs were Draper Fisher Jurvetson, Capricorn, Founders Fund (Peter Thiel), and DST (Yuri Milner). Google is the largest single investor and "hasn't sold a share"; Capricorn held until very recently. From will-marshall.
- Orbital compute: did a study with Google "about eight or nine years ago" finding that at a launch cost of ~$200–300/kg, space data centers become "simply cheaper" than ground; partnering with Google to launch TPUs and has already launched Nvidia GPUs into space. From will-marshall. See terrestrial-power-flat-to-orbital-dc-arbitrage.
- Enabling forces: rocket launch costs down "4 or 5x over the last 10 years" plus satellite miniaturization (a capability that "used to cost a billion dollars and weighed 20 tons now costs a few … hundreds of kilograms") — Marshall's "mainframe-to-desktop revolution for space." From will-marshall.
First-party update from 2026-06-26-podcast-moonshots-the-10b-satellite-empire-putting-ai-in-orbit-why (2026-06-26)
CEO will-marshall, on the record:
- Valuation and stock move: "Today Planet's a $10 billion company" and "a 450% increase in the price over the last year". Ticker confirmed on-air ("Ticker is pl").
- Archive depth (the moat claim): "Planet has 3,000 images for every point on the landmass of the Earth over the last 10 years" — ~150 petabytes; the non-replicability argument is temporal, not technical: "until someone invents a time machine".
- Revenue mix: "about 60% defense and intelligence, about 25% civil government and about 15% commercial."
- Selected to fly Google's first Project Suncatcher orbital-TPU demo satellites — which places Planet in orbital-inference-efficiency-to-tpu-advantage as the integrator rather than the chip owner.
- Latent demand claim: "I think we've got about 100x to go in the next couple of years"
Contested in the same source. alexander-wissner-gross challenges whether an orbital archive is the scarce input at all, given frontier multimodal training data. See the tension recorded on ai-real-world-data-gap-to-planet-moat — Step 4 remains open.
Strengths (from a thesis-input perspective)
- Genuinely scarce asset: largest daily-refresh global archive, hard to replicate, with switching costs for dependent governments ("countries that are fully dependent on us").
- Optionality on two secular waves at once — Earth-observation AI today, orbital compute later.
- Defense/intel revenue gives a durable, geopolitically-driven demand floor.
Weaknesses (from a thesis-input perspective)
- The "large earth models" TAM ($75–100B) is Marshall's own estimate from a self-interested CEO; unproven.
- Orbital-compute optionality is long-dated and faces the in-space clustering problem andrew-feldman flags (see terrestrial-power-flat-to-orbital-dc-arbitrage contradictions).
- Concentrated, geopolitically-sensitive customer base (defense/government ~60%).
Open questions
- spacex-public-mark-to-space-peer-rerate — Planet is one of the listed comps SpaceX's IPO will re-anchor.
Sources
Related
- will-marshall
- spacex
- ai-real-world-data-gap-to-planet-moat
- terrestrial-power-flat-to-orbital-dc-arbitrage
- ipo-comeback-public-market-value-capture
- ipo-comp-anchor-repricing
- orbital-inference-efficiency-to-tpu-advantage
- rapid-reusability-to-orbital-compute-capex-arbitrage
- low-orbit-self-cleaning-and-gpu-depreciation
- alexander-wissner-gross