Intelligence notes

The Insider Exit — Karp, Thiel, and Palantir's Liquidation Signal

15 min read

The Insider Exit — Karp, Thiel, and Palantir’s Liquidation Signal

Strategic Intelligence Assessment | intelligencenotes.com

Bottom Line Up Front

Fact. In the first calendar quarter of 2026, Alex Karp (CEO) and Peter Thiel (Chairman) of Palantir Technologies executed documented Form 4 sales totaling approximately $355.7M of personal equity (Karp ~$66M; Thiel $289.7M) through Rule 10b5-1 trading plans adopted in November 2025. Karp’s plan authorizes liquidation of up to 48.9 million shares ($1.23B at adoption-window pricing) over the plan’s full life — a figure that has been misreported as Q1 2026 execution in aggregate coverage. The divestment series executed concurrently with the most expansive government contracting period in Palantir’s history: Q1 2026 revenue of $1.633B (+85% YoY), U.S. government revenue of $687M (+84% YoY), and a Rule of 40 score of 145% — a tier Palantir shares with only NVIDIA, Micron, and SK hynix. Karp described Palantir from the Davos 2026 stage as “indispensable to Western hard power.”

Assessment (Medium). The divestment pattern is legally insulated by Rule 10b5-1 mechanics but analytically divergent from public posture. Three competing hypotheses are evaluated below. The most benign (tax/liquidity management) is consistent with surface evidence; the most significant (internal risk awareness preceding the Anthropic Rupture, Minab strike, and Goldman-Wyden-Velázquez-Garamendi congressional letter) is the least provable but the most analytically interesting given the November 2025 plan-adoption timing.

This assessment is not a security recommendation. It is a political-economy reading of what the divestment signal reveals about how the architects of the Western kill chain price their own enterprise.

1. The Mechanics

Rule 10b5-1, promulgated by the SEC in 2000 and tightened in December 2022, permits corporate insiders to establish pre-programmed trading schedules at a point when they are not in possession of material non-public information. Once adopted, the plan executes automatically — trades occur on pre-set dates, at pre-set prices, in pre-set volumes — and the affirmative defense against insider-trading liability attaches to transactions executed under the plan regardless of subsequent corporate news. The 2022 amendments added cooling-off periods (90 days for officers and directors), good-faith certifications, and mandatory disclosure on Form 4.

The Palantir insider series:

  • November 2025: Karp and Thiel separately adopt 10b5-1 plans. Karp’s plan covers up to 48.9 million shares (~$1.23B equivalent at adoption-window pricing).
  • 2026-02-20: Karp Form 4 — 493,025 shares sold (~$66M), partially coded as RSU tax-withholding cover. First documented tranche.
  • 2026-03-02: Thiel Form 4 — 2,000,000 shares sold (~$280M) via STS Holdings II LLC at $140.67-$147.13 per share.
  • Q1 2026 aggregate (confirmed, EDGAR-verified): Karp $65,955,988 (493,025 shares, Form 4 filed 2026-02-20, tax-withholding RSU cover); Thiel $289,707,507 (2,000,000 shares via STS Holdings II LLC, Form 4 filed 2026-03-02); combined $355,663,495. Fact (High) — Form 4 primary records via SEC EDGAR. Note: (a) Karp’s plan ceiling of ~$1.23B covers up to 48.9M shares over the plan’s full authorized life, not Q1 execution. (b) Three additional Thiel-affiliated vehicles — Rivendell 7 LLC (34,260,451 shares), PLTR Holdings LLC (20,823,993 shares), Rivendell 25 LLC (53,487 shares) — appear in the March 2 Form 4 solely as mandatory beneficial ownership disclosures (Footnotes 11-13); none transacted shares in Q1 2026. The combined $355.7M figure is complete and final for Q1 2026. (c) The origin of the widely-reported “$1.5B Q1 2026” media figure is the arithmetic conflation of Karp’s plan ceiling ($1.23B) with Thiel’s Q1 actual ($289.7M): 1,230M + 290M ≈ $1.52B. This is not a Q1 execution figure.

The plans are legally clean. The timing of plan adoption — November 2025 — is the analytically relevant variable. By the time Karp’s plan was adopted, the Q1 2026 earnings cycle had not yet printed; the Anthropic FASCSA designation was four months in the future; the Minab strike was four months in the future; the Goldman-Wyden-Velázquez-Garamendi multi-member oversight letter was five months in the future. Plan execution after these events is insulated. Plan adoption in advance of them is not the subject of insider-trading law, but it is the subject of analytical inquiry into what the principals knew, or were modeling, when they wrote the schedule.

2. The Valuation Context

At the time the 10b5-1 plans were adopted and through the Q1 2026 execution window, PLTR traded at multiples that are extreme even within the Magnificent-Seven AI cohort:

  • Trailing P/E: ~247-255x
  • Price-to-Sales: ~67x
  • Rule of 40 (Q1 2026): 145%

A trailing P/E of ~247x implies the market is pricing approximately 250 years of current earnings into the equity. A P/S of ~67x implies the market expects revenue to multiply by an order of magnitude before standard software-sector multiples (8-15x P/S) would re-anchor the valuation. These are not merely high multiples; they are multiples that require sustained hyper-growth across a 10-year horizon for valuation mean-reversion to be averted.

The Rule of 40 score (revenue growth plus free cash flow margin) of 145% is operationally extraordinary. It is achieved by combining 85% YoY revenue growth with FCF margins that the firm’s GAAP-to-non-GAAP bridge makes structurally legible. Stock-based compensation (SBC) is the bridge’s largest reconciling item. Palantir’s non-GAAP profitability excludes very large SBC charges — and the shares sold under the 10b5-1 plans are predominantly RSU grants previously issued as compensation. The GAAP/non-GAAP gap is not a disclosure error; it is the structural feature that produces both the reported profitability and the share supply being liquidated.

[Michael Burry’s public short position on PLTR (referenced in the parent dossier §5.2 Accounts Receivable Crisis section) is anchored on the inverse of this same arithmetic: at 67x P/S, any deceleration in the growth trajectory compresses the multiple non-linearly. Burry’s thesis is the bear-case mirror of the bull-case Rule of 40 print.]

3. Three Hypotheses — A Competing Analysis

Applying the Analysis of Competing Hypotheses (ACH) framework to the divestment pattern:

Hypothesis 1 — Tax and Liquidity Management

Claim: The Q1 2026 sales are routine RSU-vesting tax optimization and concentration management. Executives holding large illiquid blocks are advised to diversify; 10b5-1 plans are the standard legal vehicle.

Supporting evidence: The 2026-02-20 Karp Form 4 explicitly codes part of the transaction as RSU tax-withholding cover. The 90-day cooling-off period was observed (plans adopted November 2025; first executions February-March 2026). The transaction volumes are within range of prior Palantir insider activity. Thiel’s vehicle (STS Holdings II LLC) is a long-established estate-planning structure.

Disconfirming evidence: Even the documented Q1 2026 aggregate (~$355.7M) is materially larger than prior comparable quarters at Palantir. If Thiel-affiliated entity filings (Rivendell 7 LLC, PLTR Holdings LLC, Rivendell 25 LLC) add to the STS Holdings II LLC figure, the true Q1 total may approach or exceed the widely-cited ~$1.5B. Concentration management at that scale would be extraordinary — founders typically taper across multi-year horizons rather than execute >$300M in 90 days from a single named vehicle, let alone across multiple.

Confidence: Medium-High that this hypothesis explains a substantial fraction of the total but does not fully account for the aggregate scale or the November 2025 plan-adoption timing.

Hypothesis 2 — Valuation Skepticism

Claim: Karp and Thiel internally assess PLTR as overvalued at 247x P/E and 67x P/S. The plans rationally reduce concentration before any valuation mean-reversion. This is not illegal; it is a bearish signal from the most-informed sellers.

Supporting evidence: The multiples are objectively extreme. Founders are not under any disclosure obligation to volunteer that they consider their firm overvalued; they are entitled to act on that assessment under Rule 10b5-1. The plan adoption (November 2025) follows the autumn 2025 peak in AI-cohort valuations. Thiel has a documented history of selling near sector peaks (Facebook IPO; the 2020-2021 SPAC cycle).

Disconfirming evidence: Karp’s public posture (Davos 2026; The Technological Republic) consistently asserts that Palantir is structurally underpriced relative to its strategic function. Either the public posture or the trading behavior is correct; both cannot be.

Confidence: Medium. Internally consistent with the financial-forensics evidence; partially in tension with the public posture but not provably so (public statements are not under oath).

Hypothesis 3 — Internal Risk Awareness

Claim: The founders possess higher internal visibility into Palantir’s structural risk surface than the public market has priced. The November 2025 plan adoption predates the Anthropic Rupture (March 2026), the Minab School strike (March 2026), the Feinberg Memorandum’s PoR transition (March 2026), the Goldman-Wyden-Velázquez-Garamendi multi-member oversight letter (April 2026), and the UK NHS FDP break-clause signal (April 2026). Plans adopted before these events reflect, at minimum, awareness of the risk taxonomy from which they emerged.

Supporting evidence: The dossier’s red-team analysis identifies five structural vulnerabilities the market has not fully discounted: (a) the Ontology single-point-of-failure architecture, (b) the LLM supply-chain fragility demonstrated by the Anthropic Rupture, (c) congressional-oversight escalation, (d) FASCSA-type designation exposure across the broader AI vendor stack, and (e) Minab-style litigation precedent. Founders necessarily possess earlier visibility into items (a) and (b) than external analysts. Item (c) was structurally foreseeable from the June 2025 Wyden/AOC letter. The November 2025 plan adoption follows that letter by ~5 months.

Disconfirming evidence: Rule 10b5-1 plans by design insulate the principal from any inference that subsequent execution is informed by non-public material information. Pre-planning is precisely the affirmative defense. The 90-day cooling-off period was observed. No publicly available evidence directly establishes that the founders modeled any specific downstream risk event when adopting the plans.

Confidence: Assessment / Low. This is the most analytically interesting hypothesis but also the least provable. It cannot be falsified or confirmed from public records alone; doing so would require either internal communications discovery (unavailable absent litigation) or post-hoc admission. The framework, however, is not whether this hypothesis can be proven — it is whether the divergence between public posture and divestment behavior is large enough to warrant the hypothesis being held open. It is.

4. What the Market Is Pricing In

Against the divestment series sits a tailwind structure of unusual magnitude:

  • $2.3B FY2027 Pentagon Maven budget request (confirmed on the 2026-05-04 earnings call).
  • $185B Golden Dome homeland missile defense initiative — Palantir is a foundational software architect alongside Anduril, Aalyria, Scale AI, and Swoop.
  • $10B Army Enterprise Service Agreement through 2035 (awarded July 2025).
  • Maven Smart System program-of-record transition directed by the Feinberg Memorandum (2026-03-09), with a 2026-09-30 milestone — removing MSS from prototype/OTA framing and locking it into a permanent budget line.
  • NCIA MSS NATO procurement (March 2025) extending the epistemological monopoly to 32 alliance members.

The insider sell is a divergent signal against this tailwind. The market is pricing 10+ years of compounding hyper-growth into the equity at 67x P/S. The founders, simultaneously, executed documented Q1 2026 sales of ~$355.7M against a Karp plan ceiling of $1.23B over the plan’s full execution horizon — with Thiel-affiliated entity totals unresolved. Either the market is correctly pricing future cash flows that the founders also believe will materialize (in which case the sell is pure liquidity management), or the founders’ liquidity preference is itself information about the discount they apply to those future cash flows.

5. Historical Analogues — A Framework, Not an Allegation

The literature on founder-divestment-at-peak-multiple is shallow but not empty. Three reference cases are framework-relevant; none is alleged here as direct analogue:

  • Amazon, 2000: Jeff Bezos sold large blocks across 1999-2000 ahead of the dot-com collapse. Amazon ultimately validated its long-horizon thesis; the divestment signal in the moment was nevertheless predictive of multiple compression.
  • Enron leadership, 2000-2001: Senior executives liquidated significant equity in the 12 months before the firm’s collapse. The legal frame for those sales is materially different from Rule 10b5-1 because the disclosures were materially fraudulent; the case is cited here only as a reminder that sufficient divergence between public posture and insider behavior is a documented red flag in retrospect — never in real time.
  • Founder-led IPO-decade software firms more broadly: Multiple cases (CRM, NOW, SNOW) of founders executing structured liquidity at extreme multiples that did, in fact, mean-revert without firm-level distress. The divergence-signal hypothesis is therefore not a single-state predictor; it is a probability shift, not a binary call.

The framework value of the analogues is methodological. None licenses a claim that Palantir is Enron, Amazon, or Snowflake. The framework licenses only the proposition that founder divestment at extreme multiples is a non-zero signal worth surfacing — particularly when paired with a public posture that asserts the firm is structurally underpriced.

6. Strategic Implications

For investors

If the analyst is holding Palantir at current multiples, the insider divestment is material context that the firm’s IR communications do not foreground. The Form 4 series is public; the documented Q1 2026 aggregate (~$355.7M across named vehicles, with Thiel-affiliated entities unresolved) is rarely surfaced with precision in sell-side coverage — and the conflation of Karp’s $1.23B plan ceiling with Q1 execution has produced misleadingly large aggregate figures in financial media. The Burry short thesis (parent dossier §5.2) and the Karp/Thiel sells are operationally consistent — both interpret a 67x P/S print as a window for asymmetric position management. The asymmetric question for the holder is whether their information set on Palantir’s structural risk surface (Ontology SPOF; Anthropic Rupture; FASCSA exposure; congressional oversight tempo) is closer to the founders’ or to the sell-side consensus.

For policymakers

The architects of the Western kill chain — the principals responsible for the Maven Smart System program-of-record, the Golden Dome software architecture, the ELITE/ImmigrationOS deployment, the NHS Federated Data Platform, and the Persian Gulf battle-management substrate — executed documented Q1 2026 sales of ~$355.7M (with Thiel-affiliated entity totals pending EDGAR resolution) against a Karp plan ceiling of $1.23B over the plan’s full authorized life. This is not illegal. It is, however, a fact that should inform legislative oversight calibration. A firm whose principals are simultaneously declaring it indispensable and extracting nine-figure liquidity at peak multiples is a firm whose internal cost-of-capital signal diverges from its external strategic posture. Oversight instruments calibrated against the external posture alone will be miscalibrated.

For accountability structures

Insider divestment at this scale is a legal mechanism. Rule 10b5-1 is functioning exactly as designed: the plans were pre-adopted, the cooling-off period was observed, the Form 4 disclosures were timely. The political-economy implications are a separate matter from the legal compliance. The locus of operational decision authority over lethal targeting, civilian-data aggregation, and federal-workforce automation has migrated to a publicly traded firm whose CEO and chairman are managing their personal exposure to that enterprise via pre-programmed liquidation. The accountability question is not whether this is permitted (it is). The accountability question is whether the structural alignment between the firm’s strategic claims about its indispensability and its principals’ financial behavior is the alignment that policymakers, allies, and citizens have been led to assume.

7. Confidence Summary

  • Fact (High): Q1 2026 divestment confirmed via EDGAR — Karp $65,955,988 (493,025 shares, Feb 20, 2026) + Thiel $289,707,507 (2,000,000 shares via STS Holdings II LLC, Mar 2, 2026) = $355,663,495 combined. Complete and final. Thiel’s three passive LLC vehicles (Rivendell 7 LLC, PLTR Holdings LLC, Rivendell 25 LLC) transacted zero shares in Q1 2026 — they appear only as mandatory beneficial ownership footnotes in the March 2 Form 4. The “$1.5B Q1” media figure originates from arithmetic conflation of Karp’s plan ceiling ($1.23B life of plan) with Thiel’s Q1 actual ($289.7M): 1,230+290 ≈ $1,520M. No factual basis in Form 4 record.
  • Gap (Open): Karp May 20, 2026 RSU vesting tranche — the next confirmable execution event under the Nov 21, 2025 10b5-1 plan. Expected sell-to-cover at RSU vesting; filing deadline approximately May 22-23, 2026. At prevailing PLTR price, this tranche is estimated at ~$40-50M. Will update Hypothesis 1 vs. Hypothesis 3 weighting if the cadence continues at scale. Sweep on or after May 22, 2026 will confirm.
  • Gap (Open): Thiel’s larger passive vehicles (Rivendell 7 LLC at 34.3M shares; PLTR Holdings LLC at 20.8M shares) remain intact. If subsequent tranches under the Nov 14, 2025 plan execute from these vehicles, the figures would be materially larger than the STS Holdings II LLC Q1 tranche.
  • Fact: Concurrent Q1 2026 revenue print of $1.633B (+85% YoY) and Rule of 40 = 145%. High (Palantir IR release; earnings call transcript 2026-05-04).
  • Fact: Valuation multiples (P/E ~247-255x; P/S ~67x) at time of execution. High (market data).
  • Assessment: Hypothesis 1 (tax/liquidity) explains a substantial fraction but not the full aggregate. Medium-High.
  • Assessment: Hypothesis 2 (valuation skepticism) is internally consistent with the multiples and Thiel’s documented pattern. Medium.
  • Assessment: Hypothesis 3 (internal risk awareness) is the most analytically interesting and the least provable from public records alone. Low confidence on the specific causal claim; Medium confidence that the divergence between public posture and divestment behavior is large enough that the hypothesis must be held open.
  • Gap: Karp’s May 2026 post-earnings-blackout Form 4 disclosures (whether the ~$1B residual liquidation track has continued executing) will materially update Hypothesis 1 vs. Hypothesis 3 weights. Resolution pending EDGAR filings.

Key Connections

Assessment confidence: High on financial and contractual claims (SEC filings, primary IR releases, earnings call transcript). Medium on the ACH weighting across the three hypotheses. Low on the specific causal claim of Hypothesis 3. See full dossier for sourcing.

Graph View