What Planet is doing, and says it will do — the filing layer

Source: the FY2026 10-K (filed 2026-03-23, period ended 2026-01-31), Item 1 Business and Item 1A Risk Factors, plus the 424B5's own use-of-proceeds language.

The business, in the company's words: image the world every day and make change "visible, accessible, and actionable" — an archive now averaging over 3,000 images for every point on Earth's landmass, which the filing calls a "non-replicable historical archive." The revenue model is "one-to-many": each image captured sells to many customers via fixed-price subscriptions and usage contracts, with most revenue recurring — that's why the growth carries software-like economics on a hardware asset. Three product tiers per the filing: daily broad-area monitoring (the core constellation), high-resolution tasking (SkySat plus the new Pelican satellites — up to 50cm resolution, API-driven, multiple imaging modes including video), and hyperspectral (Tanager, 400+ spectral bands, built with NASA JPL and sponsored by Carbon Mapper to track methane and CO₂ super-emitters).

The strategic evolution the tape hasn't priced either way: the 10-K elevates satellite services — designing, building, launching, and operating customer-owned satellites end to end, plus dedicated tasking capacity on Planet's own fleet. That's a pivot from "sell the data" to also "sell the factory": mission systems engineering, launch procurement, ground stations, operations. It widens the addressable market into sovereign and defense programs that want their own assets, and it's capital-hungry — which is where the $1.5B ATM's own words matter: proceeds are "to fund future growth, including potential future acquisitions, and for general corporate and working capital purposes." Read plainly: a war chest for the buildout and M&A, raised from strength (they were FCF-positive in FY26), not survival capital. The AI positioning is the same architecture bet in miniature — "AI-ready" data sets and AI-enabled analytics layered on the archive, sold directly or through partners.

What the filing warns about: history of losses and uncertain future profitability; limited operating history in evolving markets; intense competition; customer-acquisition cost; international-operations execution; and — the one unique to this business — the ability to "successfully produce, launch, commission, operate and maintain" the satellites themselves. Launch and commissioning risk is a filed, recurring, physical risk no software comp carries.

The earnings record — the acceleration underneath the crash

Series basis: Massive filed quarterly statements (refreshed 2026-08-04); the two quarters Massive's feed lacks (Q3/Q4 FY26, marked †) are derived from SEC XBRL year-to-date frames (fiscal year minus nine-month, nine-month minus six-month). Growth rates and margins script-computed on the series.

* The two headline-ugly net-loss quarters are the same non-cash story: warrant fair-value remeasurement as the stock 8x'd — $161.4M in FY2026 (10-K) and $106.5M in Q1 FY27 (10-Q), both filed, both finished with the warrant redemption. The Q4 FY26 loss was $152.5M, derived from the full-year and nine-month XBRL frames.

What the series says, in three lines. This is a stalled company igniting, not a hot one cooling: four quarters of an essentially flat top line (~$60–62M, sequential moves under +1.2%) through FY2025, then five consecutive quarters of accelerating growth — +9.8% → +20.1% → +32.6% → +40.9% → +42.1% — with the guide pointing at +39–46% for the quarter that just ended. The crash landed on the two best growth quarters in the company's filed history. The net-loss column is two different stories wearing one number: operating losses sat in a modest −$18M to −$40M band throughout as the top line grew 56%, and the −$139M / −$152M headline quarters are the warrant charge, not the business. The cash is real and it never flinched: FY2026 operating cash flow totaled +$134.4M across four positive quarters. Q1 FY27 operating cash flow remained positive at +$15.4M; the last negative quarter was Q4 FY25. The honest counterweight: gross margin came down from FY25's second-half peak (62.0%) to 53.5% as the lower-margin satellite-services mix grows — the acceleration is partly bought with hardware mix, and that trade-off is the thing to watch as services scale.

What the street and the socials are saying

Morning pulls, 2026-08-04. News: the desk's news pass. Social: hosted sweep saved to the capture corpus. Sentiment context, not price truth.

The press has no crash story to tell — which is itself the story. MarketBeat's movement-aggregator explicitly finds "no major headline news"; Google Finance's summary attributes the month to "broader market sell-off and valuation concerns, despite strong underlying growth metrics." The substantive coverage is comparative: Motley Fool's PL-vs-Snowflake piece gives Planet the edge on competitive moat while flagging government-revenue concentration. Its NVDA-vs-PL growth comparison leads with the $900M backlog and the $102–107M guide. The insider-sale coverage (CFO ~$2.4M, a director ~$170K, both July) uniformly notes the 10b5-1 non-discretionary context. Nobody in the pull argues the business broke; everybody notes the price did.

The social read is scar tissue plus victory laps — not fresh euphoria. The sweep surfaced a year-ago space-stock list victory lap (@Speculator_io claiming PL +1,889% from their own 2025 entry basis — attributed, their math, not the desk's; the desk's 1-year figure on settled closes is +236%), a self-described early holder re-buying the dip after riding $2 shares in 2023–24 and consolidating into Rocket Lab (@SayNoToTrading), and independent corroboration of the Scottish Government award — a second account (GUL, handle in the saved sweep) adds that it runs through Planet's German subsidiary, a detail the desk's roster capture (@Kaizen_Investor, 2026-07-31) lacked. Two unconnected accounts on the same seven-figure contract strengthens it, though it stays attributed until a company or government notice confirms. Net: the crowd around this name is positioned longs nursing a drawdown and dip-buyers with history — a healthier sentiment floor than unanimous conviction, and consistent with the short-interest data showing sellers already sold.

Setup

• Entry zone: no entry in the collapse regime; speculative starter only on a reclaimed-and-held 20-day ($23.22 and falling) with a basing tape; confirmed trend-hold add needs the 200-day ($25.88) reclaimed and the September print clean.
• Stop: for any starter, below the 2026-07-29 structural low ($19.47 close / $19.16 intraday) — a new low into the ATM overhang reopens the leg down, and the next mapped shelf is all the way at the $12.76 point-of-control.
• Target: first the 50-day / value-area-high zone (~$30.73–30.83), then the prior value shelf; trail on structure, no fixed number.
• Conviction: LOW on timing, medium-high on the business — unchanged in words from June, but both halves moved: the business is better-verified (loss explained from the filing, warrants finished, backlog and guide primary-sourced) and the tape is objectively worse (both prior desk lines broke; regime downgraded from pullback to collapse).
• Sizing note: the ATM changes what a bounce is worth — the company is a rational seller into strength until $1.5B is raised or the program lapses. Rallies face supply; position accordingly (starter-size only, add on confirmation, never on the first reclaim day).

Bull case

• Five straight quarters of accelerating growth, primary-verified end to end: +9.8% → +20.1% → +32.6% → +40.9% → +42.1% YoY (the full series above), guided +39–46% for Q2, backlog +72% to $906M+ — a stalled company that ignited, with $335.6M in trailing-twelve-month sales and visibility companies this size rarely have. The 100% earnings beat rate and the guide's +8–14% sequential step say the September print has a low bar to clear and a management team that sets bars it clears.
• The GAAP-optics era just ended. The warrant liability that generated nine figures of non-cash loss noise is redeemed and gone. Future income statements will look like the business actually runs: modest operating losses shrinking against 40%+ growth. FY2026 operating cash flow was +$134.4M. Q1 FY27 operating cash flow was +$15.4M.
• Fortress liquidity before the equity program sells a share: $368.1M cash plus $261.4M of short-term investments (≈$630M) at 2026-04-30, with FY26 free cash flow of +$57.7M (18.8% of revenue). The equity program is opportunistic capacity, not a lifeline — this is a company arming for constellation capex from a position of strength.
• The de-rate was the sector's, not the company's. PL fell −30% in a month in lockstep with LUNR/SPCX/RKLB/ASTS while SPY rose — a liquidation event and a halo unwind, with zero adverse company news in the window (July's filings were routine Form 4s, a 144, a passive 13G/A, and an annual-meeting 8-K). When the tide that went out wasn't about you, it can come back without your permission too.
• Sold-out sellers: short interest FELL −21.7% into the July decline (41.4M shares on 6/15 → 32.4M on 7/15) — the drop was longs liquidating, not shorts pressing, and relative volume at 0.6 on the Aug 3 bounce says the supply that wanted out largely got out. Insider "distribution" also reads softer on inspection: the July sales were pre-established Rule 10b5-1 plan executions (CFO ~$2.4M on 7/23, a director ~$170K on 7/13), not discretionary exits.
• Optionality kicker (attributed, not independently verified): UK/Europe expansion — a London office and a seven-figure Scottish Government agriculture-monitoring contract per @Kaizen_Investor captures (7/02, 7/31) — and PL named as a potential data customer of Google's Project Suncatcher space-based AI datacenter concept in press coverage. None of it is in the model; all of it is free upside if real.

Bear case

• The trend is broken by the desk's own written rules — twice. The June dive's $23–24 confluence failed, then the July 18 note's $21.03 invalidation line took six consecutive closing violations. Regime: collapse. Thesis tag: broken. Price sits below every moving average that matters (−7.3% vs the 20-day, −29.9% vs the 50-day, −16.8% vs the 200-day). Trend-hold doctrine does not buy this shape; it waits for the shape to change.
• The company is now a seller above you. The $1.5B ATM means every rally toward the 200-day meets potential issuance — ~19.5% of the market cap authorized for sale "from time to time," with forward-sale mechanics that let them lock prices quietly. This is the structural difference from June: dip-buyers then fought other traders; dip-buyers now also fight the treasury department.
• Dilution is not hypothetical — it's the recent record: weighted diluted shares grew +12.3% in a single quarter (warrant cleanup), and the ATM could add another ~19% at current prices. A +42% revenue grower diluting ~20% compounds per-share value much slower than the top line suggests.
• Still expensive after a −58% drawdown: ~22.9x TTM sales (computed: ~$7.7B market cap on $335.6M TTM revenue through 2026-04-30). The multiple prices years of execution; a single guide-miss in September re-rates it violently, and the +910% two-year tape means enormous embedded gains still looking for exits on every bounce.
• The bounce is one session of confirmation-free squeeze. The desk's own Aug 3 read: the wrecked cohorts bounced hardest, "squeeze-shaped, not trend-shaped," and rotation flows still read money moving OUT of space (relative strength −22.1 on the desk's rotation map). A +5% day inside a collapse regime is how bear markets pay for attention.
• Government revenue concentration (flagged in press coverage and consistent with the defense/intel-heavy customer base): budget cycles and a handful of large contracts drive the model; a delayed award is a guide-miss with a good excuse.

Catalysts

• Q2 FY27 print, expected early September (quarter ended 2026-07-31; the last two prints landed ~5 weeks after quarter-end; no scheduled date in the desk calendar yet — the desk's own clock applies). The revenue bar is inside or above $102–107M, with backlog still compounding. Operating cash flow must remain positive alongside the first clean post-warrant income statement. A beat-and-raise into a based chart is the re-entry signal; a miss validates the collapse tag.
• The 20-day reclaim (~$23.22, falling) — the mechanical trigger that converts watch into starter. It must HOLD and base; the first touch is not the signal.
• SpaceX (SPCX) reports today, 2026-08-04 — the sector's gravity well printing its first public quarter. A strong print re-rates the whole cohort's sentiment; a weak one extends the halo unwind. Either way it moves PL this week without PL doing anything.
• ATM absorption: any 10b5-1/424B5 progress disclosure showing how much of the $1.5B has been sold. The overhang shrinking is bullish mechanically; heavy issuance into weakness is the bear tell.
• Contract flow: defense/intel and government awards (the structural demand driver) — a large multi-year award puts a fundamental floor under the base-building.

Risks

• Analytical risk — the bounce invites narrative drift. Two green sessions after a −58% drawdown make every bull point feel truer. The regime tag is the discipline: nothing here is an entry until the tape confirms, no matter how clean the September print looks in advance.
• Data risk — the guide and backlog are management's numbers (the guide is the company's own outlook; backlog is a non-GAAP measure excluding unexercised options). Both are primary-sourced but not audited-actuals; September converts them to facts or misses.
• Structural risk — ATM mechanics are opaque in real time. Forward sales mean issuance can be economically locked before it's disclosed; the desk will typically learn how much supply hit the tape a quarter late.
• Sector risk — the space cohort trades as one instrument in stress (July proved it). PL's diversification benefit inside the space/drone-defense books is lower than its business distinctiveness implies.

Financials

Leading-indicator read (2026-08-04): composite Neutral (0.4) — revenue growth bullish (QoQ rate accelerating +1.59pp), earnings strong (100% beat rate), insider trading bearish on the raw count (1 buy / 19 sells, net −$50.7M — though the named July sales were 10b5-1 plan executions).

Tape (2026-08-03 settled close): $21.53, RSI 37.7, SMA20 $23.22 / SMA50 $30.73 / SMA200 $25.88, 1d +5.13%, 7d +2.52%, 30d −30.23%, 3m −44.14%, 1y +236.4%, −58.4% off the $51.76 52-week high (early May), VWAP $23.69, AVWAP-from-low $24.05, POC $12.76, value area $6.10–30.83, relative volume 0.6, golden cross still technically active, regime collapse, thesis state broken.

Cross-references

• Prior desk work: 2026-06-29-pl-deep-dive (the June WATCH at $27.07 — its $23–24 stop line and "verify the catalyst" follow-up both resolved here) and 2026-07-18-pl-focus-name-candidate-answer (the focus-name read at $22.47 — its $21.03 invalidation TRIPPED with six closes below, 2026-07-24 → 07-31; its second entry condition, "the loss spike explained," is now SATISFIED via the 10-Q XBRL; its first, a base reclaiming the 20-day, remains open).
• Watchlists: focus (the user's top tier), space, drone-defense (July 31 scan holds PL in the nine-name pure-play collapse pocket, thesis rated broken), spacex-s1-supply-chain.
• Sector context: the SpaceX-IPO liquidity-event perspective (halo air-out fired June 25; the cohort's July leg was the market-wide liquidation documented in the 2026-08-03 market brief).
• Contrast: unlike the collapse-pocket peers (AVAV, KTOS, RCAT, DPRO et al.), PL is cash-generative with a $906M backlog — the highest-quality name wearing the sector's worst regime tag.

Sources

• SEC EDGAR — Planet Labs PBC, CIK 0001836833 (primary, financial truth):
  • FY2026 10-K, filed 2026-03-23 — business narrative, risk factors, annual financials.
  • Q1 FY27 10-Q, filed 2026-06-05 — the warrant fair-value adjustment and quarter financials.
  • Q1 FY27 earnings release (8-K EX-99.1, filed 2026-06-04) — the $102–107M guide, $906M backlog, warrant-redemption announcement.
  • 424B5 prospectus supplement, filed 2026-06-05 — the up-to-$1.5B ATM equity program (on-disk copy 424B5-2026-06-05.html).
  • 8-K filed 2026-07-10 (annual-meeting votes, routine); XBRL facts refreshed 2026-08-04.
• Tape (price truth, 2026-08-03 settled close): space + focus.json; OHLC closes for the invalidation check 30d; same-morning intraday snapshot (~$22.75) from the deep-dive pack, labeled snapshot, not close.
• Financial statements: * (income/balance/cash-flow, short-interest, sec-filings), fetched 2026-07-18/2026-08-04.
• Filing narrative (business + risks): FY2026 10-K filed 2026-03-23, Item 1/1A — cached at 10k-business and 10k-risk_factors.json; archive/product-tier/satellite-services and use-of-proceeds quotes verbatim from the 10-K and the 424B5.
• Earnings-record series: Massive quarterly statements refreshed 2026-08-04 (income-statements, cash-flow-statements.json); Q3/Q4 FY26 derived from SEC XBRL YTD frames in company-facts (FY minus 9-month, 9-month minus 6-month); FY26 warrant adjustment $161.4M from the same XBRL (10-K basis). All rates script-computed.
• News (event truth, pulls cached 2026-08-04): Motley Fool on the CFO's $2.4M 10b5-1 sale, 2026-08-02 (names the "planned $1.5 billion equity offering" — verified against the 424B5 above); Planet Labs vs. Snowflake, 2026-07-27 (moat edge, government-concentration flag); Nvidia vs. Planet Labs revenue trends, 2026-07-25 (guide + backlog framing); director's 10b5-1 sale, 2026-07-23; local cache news-latest + news/latest.json.
• Sector/market context: 2026-08-03-market-brief (the July liquidation narrative + the Aug 3 turn session + SPCX reporting 2026-08-04); 2026-07-31-drone-defense (collapse-pocket read); space-cohort 30d table computed from summaries/space.json.
• Social (attributed, not independently verified; sweep saved to 2026-08-04-social-search-pl-planet-labs): @Kaizen_Investor on the Scottish Government contract, 2026-07-31 and the London office, 2026-07-02 (roster captures in tweets); sweep items @Speculator_io's space-list victory lap, @SayNoToTrading's dip re-entry, and the German-subsidiary corroboration.
• Leading indicators: the desk's own tooling (2026-08-04).