Real Profits. Real Growth. Zero Shareholder Recourse. (NASDAQ: PLTR)

Hawkmont Research
Aug 2
15 min read

The Rare AI Company That Is Actually Profitable Trades Like It Can Never Be Wrong, and Its Founders Have Engineered Permanent Control No Matter How Much Stock They Sell

Published August 2, 2026, ahead of Palantir's Q2 2026 earnings report, scheduled for August 3.

Company Snapshot

Metric	Value
Ticker / Exchange	NASDAQ: PLTR
Recent price (Aug 1, 2026)	$122.78
52-week range	$106.37 - $207.52
Approx. market capitalization	~$295 billion
Shares outstanding	~2.40 billion
Trailing P/E	138.6x
Forward P/E	77.3x
EV/EBITDA	142.3x
Cash, equivalents, and short-term treasuries	$8.0 billion (Q1 2026)
Debt/equity	0.02, effectively debt-free
Share structure	Three classes: Class A (1 vote), Class B (10 votes), Class F (voting trust)
Founders (voting trust)	Alexander Karp (CEO), Peter Thiel, Stephen Cohen (CTO)
Founder voting floor	Class F shares automatically adjust to guarantee Founders a minimum of 49.999999% of total voting power
Next earnings report	August 3, 2026, after market close
YTD price change	Approximately -30% to -31%

Executive Summary
Business Overview
Deep Dive: The Governance Ledger, Permanent Control by Design
Deep Dive: Customer Concentration and the Government Revenue Base
Financial Snapshot
Competitive Positioning
Valuation
Bull and Bear Scenarios
Risks and Sensitivities
Conviction and Portfolio Framework
Appendix: Sources and Comparables
Disclosures and Important Notices

1. Executive Summary

Palantir occupies a position almost no other company in the current AI cycle can claim: it is genuinely, substantially profitable, at scale, while growing faster than nearly every large-cap software peer. First-quarter 2026 revenue grew 85% year over year, the fastest pace since the company's 2020 direct listing, with a 60% adjusted operating margin and a Rule of 40 score of 145%, a combined growth-plus-margin metric that management itself notes places Palantir alongside chipmakers like Nvidia rather than typical software companies. This is not a story about a promising but unprofitable AI bet. It is a story about a company that has already proven the model works.

The market has responded by pricing PLTR at levels that assume the model keeps working without interruption for years. At roughly 77 times forward earnings, 38 times forward revenue, and 142 times EV/EBITDA, Palantir trades at a multiple several times higher than comparable high-growth software peers such as Snowflake, and even above hardware-cycle beneficiaries like Nvidia on a sales basis. The stock is down roughly 30% year to date and more than 40% from its 52-week high, which tells its own story: even a company executing at a genuinely elite level can see its multiple compress hard once a market priced for perfection starts pricing in the possibility of a single wobble.

Underneath both the growth and the valuation risk sits a governance structure that is unusual even among founder-controlled technology companies. Palantir has three classes of stock. Class A trades publicly and carries one vote. Class B carries ten votes. Class F, held exclusively in a voting trust by CEO Alex Karp, director Peter Thiel, and CTO Stephen Cohen, is engineered to automatically adjust its voting weight so that the three Founders collectively retain at least 49.999999% of total voting power, regardless of how much stock they sell or how diluted their economic ownership becomes over time. This is not a typical dual-class discount investors can model with a standard governance haircut; it is a mechanism explicitly designed to make founder control permanent and independent of ownership stake.

That mechanism matters more given the insider trading pattern of the past several months. Thiel sold more than two million shares in a single day in March 2026. Karp and Cohen both sold in February. The trailing three-month selling-to-buying ratio among insiders has run as high as 9.3 to 1. Much of this activity is mechanical, tied to pre-planned 10b5-1 programs and RSU tax-withholding rather than a single discretionary decision to exit, and it should be read as such rather than as a uniform signal of insider pessimism. But the combination of heavy, sustained insider selling with a governance structure that guarantees those same insiders permanent voting control regardless of how much they sell is a specific and identifiable feature of this stock that most comparable software companies do not share.

The business itself carries a second first-class risk that is independent of valuation or governance: revenue concentration in politically sensitive US government work. US government revenue represented roughly 42% of total company revenue in the first quarter of 2026 alone, built on relationships with the Department of Defense, the US Army's Vantage program, and Immigration and Customs Enforcement, among others. These contracts are sticky and defensible, which is precisely why they are also a source of recurring political controversy, protest, and headline risk that has followed the company domestically and internationally.

This note treats Palantir's operating excellence and its structural risk as equally real and simultaneously true. The company has done something almost no AI-adjacent business has managed: turn a genuinely hard, defensible enterprise software product into GAAP profitability at scale. It has also built a valuation, a governance structure, and a customer base that leave very little room for anything to go wrong, and very little recourse for public shareholders if it does.

2. Business Overview

Palantir builds software platforms that turn an organization's data into a governed, operational system for decision-making, distinct from what management calls "model wrapper" AI products. The company operates four principal platforms:

Gotham: the original platform, built for intelligence and defense customers, used for counterterrorism and operational military analysis.
Foundry: the commercial-facing data integration and operations platform, the primary engine behind Palantir's US commercial growth.
Apollo: a software delivery and deployment layer that pushes updates and enables deployment across any environment, including classified and air-gapped settings.
Artificial Intelligence Platform (AIP): Palantir's flagship AI product, providing governed access to open-source, self-hosted, and commercial large language models, with an emphasis on auditability, cost attribution, and enterprise-grade authorization that differentiates it from consumer-facing AI tools.

Palantir's core structural bet is that "operational AI," AI wired into an organization's actual workflows and decision rights with governance controls attached, is a fundamentally different and more defensible product category than general-purpose AI model access. Management has repeatedly framed this distinction as the reason Palantir can charge premium prices and retain customers at high net dollar retention rates, and the first-quarter 2026 numbers, particularly a 133% year-over-year jump in US commercial revenue, are the strongest evidence yet that the framing is resonating with buyers beyond the company's original government base.

Revenue splits along two axes: US versus international, and government versus commercial. In Q1 2026, US revenue was $1.282 billion (78.5% of total), growing 104% year over year, split between US commercial ($595 million, up 133%) and US government ($687 million, up 84%). International revenue, the remaining roughly $351 million, has historically grown more slowly and remains a smaller share of the business than either US segment.

3. Deep Dive: The Governance Ledger, Permanent Control by Design

Most dual-class technology companies (Alphabet, Meta) give founders outsized voting power through supervoting share classes that nonetheless dilute over time as founders sell stock or as the company issues new shares. Palantir's structure is different in a specific, disclosed, and important way.

Three share classes, one designed to never lose control. Class A common stock, the shares that trade on Nasdaq, carries one vote each. Class B carries ten votes each and is held primarily by insiders and early investors. Class F is the outlier: it is held exclusively in a voting trust established by Karp, Thiel, and Cohen, and by its own terms it automatically adjusts its voting weight to ensure the three Founders collectively control at least 49.999999% of total voting power, as long as they retain a minimum ownership threshold. In practice, this means the Founders' control is not simply large, it is structurally insulated from the normal dilution mechanics that erode founder control at most other public companies over time.

Insider selling has been persistent and sizable. Peter Thiel, still the largest individual shareholder at roughly 3% of shares outstanding (about 69 million shares, worth more than $10 billion as of mid-2026), sold more than two million shares in a single day in March 2026 at prices between $140.97 and $146.80. Karp and Cohen both sold in February 2026. Over the trailing three months as of mid-May 2026, the insider selling-to-buying ratio ran as high as 9.3 to 1. Some of this is mechanical: Karp's late-May 2026 transactions, for example, were tied to RSU vesting and automatic tax-withholding sales under a pre-established 10b5-1 plan, leaving his net position roughly unchanged. But the pattern, sustained selling by the three people who also hold permanent voting control regardless of their economic stake, is worth tracking as its own data point, separate from any single transaction.

Stock-based compensation is a real and growing dilution factor. Palantir recorded $684 million in stock-based compensation in fiscal 2025. Combined with continued share issuance to fund growth, this is a genuine, ongoing dilution mechanism for Class A holders, even though it does not touch the Founders' guaranteed voting floor.

What this means for public shareholders. Ordinary dual-class discounts assume that founder control, while outsized, is still ultimately a function of economic ownership that erodes with sales and dilution over time, giving public shareholders a long-run path toward more influence. Palantir's Class F mechanism removes that assumption entirely. Public shareholders are not on a long-run path toward more say in the business; the structure is explicitly built so that outcome cannot happen, regardless of how the Founders manage their own holdings.

4.1 Deep Dive: Customer Concentration and the Government Revenue Base

Palantir's single largest revenue segment by dollar amount in Q1 2026 was US government, at $687 million, or 42% of total company revenue for the quarter. Combined with international government revenue, government work has historically represented slightly over half of total company revenue in recent periods, though that share is falling as US commercial growth outpaces it.

Key relationships. The US Army's Vantage program, a roughly $10 billion multi-year agreement, is among the largest and most visible defense relationships. Immigration and Customs Enforcement is another major government customer, and its use of Palantir's software has become a recurring flashpoint for public protest in the US. France's domestic intelligence agency, DGSI, and the US Department of Agriculture (a contract with up to a $300 million ceiling awarded in 2026) round out a customer list that skews heavily toward agencies whose missions are, by design, politically sensitive.

Why this is a first-class risk, not a footnote. Government contracts of this type are genuinely sticky, deeply embedded in mission-critical workflows, requiring extensive security accreditation that competitors struggle to replicate quickly, which is exactly why bulls point to them as a durable moat. That same stickiness means Palantir's revenue base is directly exposed to shifts in government budget priorities, changes in political administration, and the reputational and protest risk that has already followed specific contracts (particularly the ICE relationship) both domestically and internationally. A company this concentrated in politically contested government work does not carry the same risk profile as a similarly sized enterprise software company selling primarily to private-sector customers, and that distinction should be priced, not waved away as background noise.

5. Financial Snapshot

Q1 2026 results (reported May 4, 2026):

Metric	Q1 2026	YoY change
Total revenue	$1.633B	+85%
US revenue	$1.282B	+104%
US commercial revenue	$595M	+133%
US government revenue	$687M	+84%
Adjusted income from operations	$984M (60% margin)	n/a
GAAP net income	$871M	n/a
GAAP EPS	$0.34	n/a
Adjusted EPS	$0.33	vs. $0.28 consensus
Cash from operations	$899M	n/a
Adjusted free cash flow	$925M	n/a
Rule of 40 score	145%	n/a
Remaining performance obligations (RPO)	$4.5B	+134% YoY, +9% sequential
Total remaining deal value	$11.8B	+98% YoY, +6% sequential

Q2 2026 guidance (reporting August 3, 2026): Revenue of $1.797-1.801 billion (consensus was tracking to roughly $1.81 billion, implying ~81% year-over-year growth), adjusted income from operations of $1.063-1.067 billion.

Full-year 2026 guidance (raised at Q1):

Metric	FY2026 Guidance	Change vs. prior guide
Revenue	$7.650-7.662B (+71% YoY)	Raised from $7.182-7.198B, a 10-point raise, the largest in company history
US commercial revenue	>$3.224B (+120%+ YoY)	Raised
Adjusted income from operations	$4.440-4.452B	Raised
Adjusted free cash flow	$4.2-4.4B	Raised

Reading the numbers: Palantir closed 206 deals of at least $1 million in Q1 2026 alone, including

47 deals of at least $10 million, evidence of broadening enterprise adoption rather than dependence on a small number of mega-contracts. The company ended the quarter with $8.0 billion in cash and short-term treasuries and effectively no debt, giving it enormous flexibility to keep investing in commercial expansion without external financing. The one note of caution inside otherwise exceptional results: total remaining deal value growth (6% sequential) has been less consistent quarter to quarter than headline revenue growth, suggesting new bookings momentum, while still strong, is somewhat less smooth than the topline figures suggest.

6. Competitive Positioning

Palantir's competitive set splits into distinct buckets rather than a single head-to-head rival, and the company's public commentary and independent competitive tracking suggest it wins or loses share differently in each:

Enterprise data and AI platforms. Snowflake and Databricks (private, last valued near $43 billion in its 2024 Series J round) increasingly overlap with Palantir's Foundry and AIP products on the data-plus-AI layer, though the relationship is partly coopetitive: a meaningful share of Palantir deployments run on top of Snowflake or Databricks infrastructure rather than replacing it outright. C3.ai competes more directly on industry-specific AI applications but trades at a fraction of Palantir's multiple and has not demonstrated comparable growth or margin performance.

The most underrated competitive threat. Independent competitive analysis flags Microsoft Fabric, which bundles Foundry-equivalent data and AI orchestration functionality directly into existing Azure enterprise agreements many large customers already hold, as a structurally significant threat precisely because it does not require a net-new purchasing decision the way a Palantir deployment does.

Defense and government-adjacent. Anduril (private, last valued near $14 billion in 2024) competes for command-and-control and defense-technology budgets rather than data-platform budgets directly, while Booz Allen, SAIC, and CACI compete on government accreditation and systems-integration relationships that Palantir has had to build over more than a decade.

Structural advantage. Palantir's genuine edge is the combination of deep government security accreditation, built over roughly two decades, with a governed, auditable approach to AI deployment that differentiates it from lighter-weight AI tools in regulated, high-stakes environments. That combination is difficult and slow for competitors to replicate, which supports the bull case for durable pricing power, but it is not the same as being unassailable, particularly as Microsoft and other hyperscalers bundle adjacent capability into agreements customers already have in place.

Join 100,000+ readers who get Hawkmont Research delivered straight to their inbox. Independent, conflict-free equity research, no paywall, no compensation from the companies we cover. Hawkmont Research is free and will stay free.

7. Valuation

At a recent price of $122.78 and roughly 2.40 billion shares outstanding, PLTR carries a market capitalization of approximately $295 billion.

Multiple	PLTR	Selected peer context
Forward P/E	77.3x	Well above typical large-cap software peers
Trailing P/E	138.6x	n/a
PEG ratio	1.50	n/a
EV/EBITDA	142.3x	n/a
EV/FCF	106.8x	n/a
Forward P/S (on FY2026 guidance)	~38.5x	Snowflake ~15-16x; C3.ai a small fraction of that; Nvidia ~17x
Trailing P/S (on FY2025 actual revenue)	~66x	n/a

Palantir's multiples sit meaningfully above nearly every comparable software or AI-infrastructure peer on both a trailing and forward basis, a gap that has persisted even after the stock's roughly 30% year-to-date decline. What appears priced in is a continuation of the current growth trajectory largely without interruption: sustained 70%+ revenue growth, continued US commercial acceleration above 100%, and margin expansion holding at or near current levels for multiple years. What appears comparatively under-priced is the governance and customer-concentration discount that a similarly structured company in almost any other sector would likely carry, given the combination of permanent founder voting control and revenue concentration in politically contested government contracts.

Sell-side sentiment remains constructive on balance. Consensus ratings cluster around "Buy" or "Moderate Buy," with average price targets in the $182-192 range as of recent tracking, implying 48-58% upside from current levels. That gap between a bullish consensus and a stock still down sharply from its 52-week high reflects a genuine and unresolved debate among professional investors about whether Palantir's operating performance justifies its multiple, or whether the multiple has simply been pricing in an unusually long runway of flawless execution that leaves no margin for a single disappointing quarter.

8. Bull and Bear Scenarios

Bull case. Palantir has done what almost no other AI-era company has: proven a genuinely profitable, high-margin, high-growth business model at scale, with a Rule of 40 score (145%) that places it alongside chip manufacturers rather than typical software peers. US commercial revenue growth of 133% year over year suggests the company is successfully extending beyond its original government base into a much larger addressable market. RPO growth of 134% year over year signals multi-year revenue visibility well beyond what quarterly headline numbers alone would suggest. An $8 billion cash position with effectively no debt gives the company enormous flexibility to keep investing without external financing risk. Under this case, the stock's roughly 40% decline from its 52-week high is read as a healthy reset of an overheated multiple, not a signal that the underlying business has weakened, and continued beat-and-raise quarters gradually close the valuation gap through earnings growth rather than price decline.

Bear case. The valuation prices in years of continued flawless execution in a market that has already shown, with the stock's 40% drawdown from its peak, how violently the multiple can compress on any wobble. Insiders, including the three people who also hold permanent guaranteed voting control through the Class F structure, have been selling at a rate of up to 9.3 shares for every 1 bought over recent months. That governance structure means public shareholders have essentially no long-run mechanism to influence capital allocation or strategic direction regardless of how the stock performs, an unusual level of permanence even by founder-controlled tech company standards. Revenue concentration in politically contested government contracts, more than 40% from US government alone in the most recent quarter, exposes the business to budget and political risk that is not present in a typical enterprise software company's customer base, and has already generated real-world protest and controversy tied to specific contracts. Competitive bundling from Microsoft Fabric into existing Azure agreements threatens exactly the commercial expansion runway that current growth expectations depend on most heavily.

Both scenarios describe the same company. The question the market has not resolved is which one the current multiple is actually paying for.

9. Risks and Sensitivities

Prioritized, not exhaustive:

Valuation risk. Forward multiples several times higher than comparable high-growth software peers leave little room for a growth deceleration without significant multiple compression.
Governance structure. The Class F voting mechanism guarantees Founders a minimum 49.999999% of total voting power in perpetuity, regardless of dilution or share sales, removing the normal long-run check public shareholders have at most dual-class companies.
Insider selling pattern. A persistent, multi-month selling-to-buying ratio as high as 9.3 to 1 among insiders, including the Founders who hold permanent voting control.
Customer concentration in politically sensitive government work. US government revenue alone represented 42% of total revenue in the most recent quarter; contracts with agencies such as ICE have generated recurring public controversy.
Stock-based compensation and dilution. $684 million in fiscal 2025 SBC represents an ongoing, real dilution factor for Class A shareholders.
Competitive bundling risk. Microsoft Fabric's integration into existing Azure enterprise agreements threatens to erode differentiation at the commercial edge where growth expectations are currently highest.
Booking consistency. Sequential remaining deal value growth has been less smooth than headline revenue growth, a data point worth monitoring for early signs of demand softening.
Regulatory and political risk. AI governance regulation and data privacy scrutiny, particularly around government and defense use cases, remain live and evolving risks in multiple jurisdictions.
Execution risk on international and non-US government expansion, where growth has historically lagged the US business.

10. Conviction and Portfolio Framework

The following is offered as an interpretive framework for readers doing their own position-sizing work, not a recommendation.

Palantir is one of the few names in the current AI cycle where the operating business genuinely supports a high-conviction thesis on its own merits: real profitability, real margin expansion, real acceleration in the hardest segment to win (US commercial). That is a meaningfully different starting point than a story stock still proving out unit economics. The open question for sizing is not whether the business works. It clearly does, at least so far. The open question is whether the price already reflects years of continued flawless execution, and whether a governance structure that removes normal shareholder recourse, combined with a customer base concentrated in politically contested government work, deserves a discount the current multiple does not appear to be applying.

Where this fits in a thematic AI or software sleeve is a function of two largely separable judgments: conviction in the durability of Palantir's operating model (high, on the numbers to date) and comfort holding a security where governance influence is permanently and structurally foreclosed regardless of price or performance (a binary tolerance more than a spectrum). Position sizing decisions should weight the second judgment as heavily as the first, since it does not improve with time or price appreciation the way it might at a more conventional dual-class company.

Key monitoring points going forward:

Q2 2026 earnings (Aug 3): watch specifically for US commercial growth sustaining above 100% and any change in sequential RPO/remaining-deal-value momentum.
Insider selling-to-buying ratio trend in subsequent quarters, and whether it normalizes or continues at an elevated pace.
Any material change in the government contract mix, particularly renewal terms on the largest defense and ICE-related agreements.
Competitive response from Microsoft Fabric bundling and any measurable share loss in Palantir's stated competitive-loss commentary on earnings calls.
Multiple compression sensitivity: how the stock trades on a quarter that beats estimates but shows any deceleration in the growth rate itself.

11. Appendix: Sources and Comparables

Primary sources: Palantir Technologies Q1 2026 earnings press release and Business Update presentation (May 4, 2026); Form 10-Q for the quarter ended March 31, 2026 (SEC EDGAR); Form 4 insider trading filings (May 2026).

Secondary sources referenced: CNBC and Motley Fool Q1 2026 earnings coverage; Futurum Group and mlq.ai earnings analysis; Yahoo Finance, Robinhood, Investing.com, and StockAnalysis.com pricing and valuation data; Simply Wall St company overview; TIKR.com and BusinessQuant insider ownership tracking; 24/7 Wall St. and Yahoo Finance insider-selling coverage; RevenueMemo and WhoIsTheOwnerOf ownership-structure explainers; Motley Fool governance commentary; Lambda Finance competitive-landscape analysis; AInvest valuation commentary.

Illustrative public and private comparables:

Company	Relevant comparison
Snowflake (SNOW)	Closest public data-platform peer on forward P/S basis
C3.ai (AI)	Smaller, lower-multiple industry-specific AI applications peer
Databricks (private)	Data-plus-AI platform competitor, partly coopetitive with Palantir deployments
Microsoft (MSFT)	Fabric bundling into Azure enterprise agreements, the most cited emerging competitive threat
Anduril (private)	Defense-technology competitor, command-and-control rather than data-platform focus
Booz Allen, SAIC, CACI	Government systems-integration and accreditation competitors

12. Disclosures and Important Notices

This report was independently produced by Hawkmont Research. Hawkmont Research has not received, and does not accept, compensation from Palantir Technologies Inc., its affiliates, officers, directors, or any party with a financial interest in PLTR securities in connection with this report. Hawkmont Research holds no position in PLTR securities as of the date of publication.

This report is provided for informational and educational purposes only. It does not constitute, and should not be construed as, investment advice, a recommendation to buy or sell any security, or an offer or solicitation in any jurisdiction. Nothing in this report accounts for any individual reader's financial situation, risk tolerance, or objectives. All figures are drawn from public filings and secondary reporting as cited above and are subject to revision, restatement, or correction by the company in subsequent filings; readers should independently verify all data points against primary source filings before making any investment decision.

Forward-looking statements regarding guidance and upcoming earnings are based on company disclosures as of the date of this report and are subject to change. Past price performance is not indicative of future results.

Readers must form their own independent conclusions and are encouraged to consult a licensed financial or legal advisor before making any investment decision involving PLTR or any other security discussed in this report.