
Key Points
- Quality of the business — FY25 revenue +56.2%, 157% NDR, 62% adj. op margin, Rule of 40 of 155.
- AIP flywheel — US Commercial 4.9x in two years, bookings still outrunning revenue.
- Bookings divergence — TCV +49% vs revenue +93%, non-US Commercial TCV −13%, backlog 0.80x.
- Political risk — DGSI/NHS sovereignty pushback plus sole-source federal exposure.
- Valuation call — SELL, US$150 TP, 14% downside; fair P/E 43x vs 108x FY26; price embeds a 40% decade CAGR
1. Executive Summary
Company overview. Palantir sells operational software to institutions that must act on messy, sensitive, multi-source data: Gotham (defence and intelligence), Foundry (the commercial analogue, which builds the "ontology" — a digital twin of a company's processes), AIP (the 2023 layer that lets AI agents take governed actions against that model) and Apollo (deployment into any environment, including classified enclaves). Embedded forward-deployed engineers install it. It is sold as a committed annual subscription with metered consumption inside the allowance.
Financial Performance. Revenue reached $4,475 million (+56.2% YoY), significantly exceeding the initial February guidance of $3,741–$3,757 million and delivering a 19.4% beat against full-year estimates. The US market, which now accounts for 74% of total revenue, was the primary growth engine. US revenue surged 75% to $3,320 million, anchored by strong US Commercial adoption, which jumped 109% to $1,465 million. US Government revenue also posted strong growth, rising 55% to $1,855 million. Overall, government clients still comprise 54% of the total revenue mix. The absolute standout metric was bookings. Total Contract Value (TCV) skyrocketed 128% to $10.8 billion, pushing the company's full-year Rule of 40 score to 106%.
Strong growth and company moat: US Commercial is compounding at ~121% p.a. towards guided revenue of >$3.4 billion, and here the bookings corroborate it — US Commercial TCV +153%, while remaining deal value rose 124% to $6.24 billion. Net dollar retention reached 157%, up from 139% two quarters ago. The installed base alone delivers 57% growth with no new customers.
The bears: The bookings do not corroborate the pace of revenue growth. Total Contract Value grew +49% against +93% revenue growth, while non-US Commercial TCV fell 13% in Q2. GAAP backlog is 0.80x trailing 12-month revenue, down from 0.94x last quarter. Revenue is being recognised faster than it is being replaced. Politics is a live cost on both flanks. Europe is moving from slow to closed — France's DGSI is migrating to ChapsVision, while the UK is weighing an NHS exit clause. In the US, roughly half of Palantir's ~$3.2 billion of federal obligations since 2024 were awarded without competition, a structure exposed to the next administration.
Current price embeds a 40% revenue CAGR for a decade — 20% above consensus in 2027 and 36% above in 2028. As such, we initiate a SELL call for Palantir Technologies with a target price of US$150. This is a criticism of the price, not the business. Palantir can beat its own guidance by 6% every quarter for the next three years and the shares would still be expensive.
2. Company Overview
Palantir sells four things that are usually described as one.
Gotham is the original product: a data-integration and analysis platform for defence and intelligence customers, built around the problem of fusing classified, multi-source, differently-permissioned data into a single operational picture. Its moat is substantially accreditation — the years and cost required to obtain and maintain authority-to-operate at classification levels — and workflow entrenchment.
Foundry is the commercial analogue: an operational data platform that builds a digital twin (Palantir's "ontology") of an enterprise's physical and business processes, so that supply chains, manufacturing lines, claims processes or clinical pathways can be simulated and acted upon in one system.
AIP (Artificial Intelligence Platform), launched 2023, is the layer that made the last three years happen. It connects large language models to the ontology with governance, permissioning, lineage and human-in-the-loop controls, so that an AI agent can take actions against enterprise systems rather than merely produce text. This is the product that took US commercial revenue from $702m in FY2024 to a guided $3.42bn+ in FY2026 — a 4.9x increase in two years.
Essentially, AIP is the catalyst that transformed Palantir from a high-friction, bespoke government contractor into a hyper-scalable commercial software vendor. With this, AIP has shortened Palantir’s adoption cycle and improved the scalability of its commercial model. f. Deploying its legacy Foundry platform required highly technical, top-down enterprise sales motions that could take months or even years to close. AIP has fundamentally disrupted this model.
Apollo is the deployment and continuous-delivery infrastructure that lets all of the above run identically in a cloud region, an on-premise data centre, a classified enclave, or a ship at sea. It is the least-discussed and arguably most strategically important asset, because it is what allows Palantir to serve the customers that hyperscaler-native competitors structurally cannot.
Wrapped around all four is the forward-deployed engineer (FDE) model — Palantir embeds engineers inside the customer to build the first workflows. Historically the bear case against Palantir; now, in an environment where enterprises cannot hire AI implementation talent, it is the acquisition moat.
At the centre of all four products, however, lies Palantir's deepest and strongest moat: the Ontology. Ontology transcends traditional, read-only data management by functioning as a bidirectional, operational digital twin of an enterprise. Unlike legacy databases or basic semantic layers, which primarily aggregate information for human analysis, the Ontology intrinsically maps physical assets, logical relationships and executable actions into a unified architecture. This proprietary framework allows users and AI agents not only to query data, but also to safely "write back" operational decisions directly into underlying systems of record, while remaining bounded by strict enterprise security guardrails.
Figure 1: Palantir business overview

49% of Palantir’s revenue came from US Commercial, with the remainder coming from US Government and international customers. International revenue fell from 26.9% of total revenue to 18.7% in a single year, an 8.2 percentage-point decline in its contribution, while US Commercial gained 8.9 percentage points to become the largest single segment. On current trajectories — with Commercial compounding at 110% versus 79% for Government — Commercial should overtake Government as the larger segment during the second half of 2026. US Commercial is supported by bookings growing faster than revenue, while US Government sits behind a non-US Commercial bookings line that contracted 13% year-over-year in the quarter.
Table 1: Segment performance
|
Segment |
Q2 2026 |
Q2 2025 (est.) |
Growth |
% of revenue |
Contribution to +93% |
|
US Commercial |
$764.0m |
$306.8m |
+149% |
39.5% |
+45.6pp |
|
US Government |
$809.0m |
$425.8m |
+90% |
41.8% |
+38.2pp |
|
International Commercial |
$181.4m |
$142.8m |
+27% |
9.4% |
+3.8pp |
|
International Government |
$181.0m |
$127.0m |
+42% |
9.4% |
+5.4pp |
|
Total |
$1,935.5m |
$1,002.5m |
+93% |
100% |
+93pp |
Source: Palantir Report. Data as of 20 August 2026.
Table 2: Aggregation
|
Aggregation |
Q2 2026 |
Growth |
Mix |
|
United States |
$1,573m |
+115% |
81.3% |
|
International |
$362.4m |
+34% |
18.7% |
|
Government (total) |
$990.0m |
+79% |
51.1% |
|
Commercial (total) |
$945.4m |
+110% |
48.8% |
Source: Palantir Report. Data as of 20 August 2026.
Table 3: Backlog
|
Quarter |
Total TCV |
TCV y/y |
US Comm TCV |
y/y |
Non-US-Comm TCV |
y/y |
Revenue y/y |
TCV/Revenue |
|
Q3 2025 |
$2,760m |
+151% |
$1,310m |
+342% |
$1,450m |
+81% |
+63% |
2.34x |
|
Q4 2025 |
$4,262m |
+138% |
$1,344m |
+67% |
$2,918m |
+196% |
+70% |
3.03x |
|
Q1 2026 |
$2,410m |
+61% |
$1,176m |
+45% |
$1,234m |
+80% |
+85% |
1.48x |
|
Q2 2026 |
$3,373m |
+49% |
$2,132m |
+153% |
$1,241m |
−13% |
+93% |
1.74x |
Source: Palantir Report. Data as of 20 August 2026.
From FY25 reported financials: In FY2025, Palantir delivered exceptional top-line growth, with total revenue reaching $4.475 billion, representing a 56% increase year-over-year. This performance was heavily driven by the company's US operations, which surged 75% to $3.32 billion. The standout segment was US Commercial, which skyrocketed 109% year-over-year to $1.465 billion, demonstrating strong enterprise adoption of its platforms.
Profitability inflected sharply during the year, proving the scalability of its software model. Palantir generated GAAP operating income of $1.414 billion, reflecting a 32% margin. The company posted full-year GAAP net income of $1.625 billion, achieving an outstanding Rule of 40 score of 127% by the fourth quarter. It closed the fiscal year with a record-setting Total Contract Value (TCV) of $4.26 billion in Q4 alone.
3. Industry Overview
Palantir does not have a single, true one-to-one publicly traded competitor that mirrors its exact end-to-end architecture. This lack of a direct comparable is precisely why Palantir commands a unique premium valuation — it straddles the line between a defence prime contractor, an enterprise data infrastructure provider and an AI application layer.
When enterprises run a standard Request for Proposal (RFP) for data aggregation and analytics, Palantir often finds itself in bake-offs against Microsoft or Snowflake. However, when a company or government agency requires an IL-6 certified platform (the highest DoD security clearance) that can actively write back operational decisions to legacy ERPs such as SAP in real time using governed AI agents, the list of viable competitors shrinks to almost zero.
For instance, Microsoft Fabric is a cloud-native, highly interconnected SaaS platform. It relies on constant external telemetry, cloud-based APIs, continuous commercial updates and distributed control planes. To achieve IL-6, Microsoft must effectively "sever" Fabric from the commercial internet and rebuild its control plane to operate entirely offline. This requires stripping out internet-dependent features and retrofitting the platform to function without its standard commercial tethers — a massive engineering hurdle. Palantir, by contrast, built its platforms (such as Gotham and Foundry) from day one with the capability to be deployed on disconnected, air-gapped networks and tactical edge environments.
However, the caveat is that for enterprises that do not need a complex, bidirectional digital twin to run their business — which represents the majority of US commercial companies and is Palantir's fastest-growing business segment — Microsoft Fabric could actually be a "good enough" competitor. Microsoft is aggressively climbing up the technology stack to commoditise Palantir's core moat: the Ontology. Microsoft is developing features such as Fabric IQ, which functions as a metadata and semantic layer sitting on top of OneLake data. While it currently lacks the deep, bidirectional write-back capabilities and complex action frameworks of Palantir's Ontology, it is a clear signal of Microsoft's roadmap. If Microsoft successfully matures Fabric's semantic capabilities to allow enterprise AI agents to execute reliable workflows natively within the Azure ecosystem, Palantir's unique value proposition will face severe commoditisation pressure.
4. Investment Thesis
The US commercial AIP flywheel (reflected in expectations)
Palantir's US Commercial business has compounded from $702 million (FY2024) to $1,465 million (FY2025, +109%) to guided revenue of >$3,424 million (FY2026, +134%) — a 4.9x increase in two years.
The metric that captures this expansion is net dollar retention: 139% (Q4 2025) → ~150% (Q1 2026) → 157% (Q2 2026). At 157%, the installed base alone would deliver 57% revenue growth with zero new customers. This is rarely seen outside of hyper-growth, consumption-based software companies, such as Snowflake during its IPO phase, where customers rapidly scale their computing usage once the platform is installed. This provides strong evidence of customer expansion after adoption.
Once customers see the value of the Artificial Intelligence Platform (AIP), they rapidly expand their seat counts, roll the software out to new departments and execute more use cases, driving their annual spend up aggressively within a 12-month window.
Critically, (unlike the government segment) , US commercial bookings are growing faster than US Commercial revenue (+153% TCV vs +149% revenue). The engine is replenishing itself.
AIP is the strongest catalyst for this. After launching in 2023, US Commercial revenue grew from $702 million in FY2024 to a guided $3.42 billion+ in FY2026 — a 4.9x increase in two years.
Tyson Foods generated over $200 million in cost savings across 20 use cases within two years. Specifically, it increased average truck capacity utilisation from 46% to 87% after implementing Palantir. Palantir embedded engineers with Airbus to build "Skywise", an open aviation data platform powered by Foundry. Skywise is now used to monitor fleet health and coordinate predictive maintenance for almost 12,000 connected aircraft globally, drastically reducing operational delays and streamlining production efficiency across Airbus's own manufacturing lines.
We expect US Commercial revenue to be the main revenue contributor in the coming years. Its differentiated offerings and these characteristics support a premium to the software-sector median, although the appropriate premium depends on whether current customer expansion and margin levels can be sustained..
Figure 2: Net Retention Rates showing that existing customers are renewing larger value

Figure 3: Palantir gaining traction from US commercial customers

The slowing government franchise (Negative)
Government accounted for 51.1% of Q2 2026 revenue ($990 million, +79% y/y). The structural change over the past eighteen months is that Palantir has moved from winning individual programmes to owning procurement vehicles, such as the US Army Enterprise Service Agreement (July/August 2025): a ten-year contract worth up to $10 billion, consolidating dozens of separate Palantir contracts into a single instrument. The agreement reduces near-term procurement friction and may support further task orders, although its US$10 billion ceiling should not be treated as committed revenue or backlog. The August 2026 DoD directive further reinforced this position, with up to $243.9 million committed through March 2027 and instructions to identify additional spending through December 2028.
As mentioned earlier, compared with its commercial business, Palantir possesses a stronger moat in the government segment. Although US defence spending is expected to increase over the coming years, the segment faces several headwinds. The bookings do not fully support the pace of revenue growth: non-US Commercial TCV declined 13% year-over-year in Q2 2026 and grew only 17.5% across H1 2026, compared with government revenue growth of approximately 79%. Some of this can be explained by ESA accounting, where task orders are recognised without generating new headline TCV, but not all of it. The trend is also consistent with the deterioration in GAAP Remaining Performance Obligations (RPO), which fell to 0.80x revenue from 0.94x previously. If Palantir cannot secure sufficiently large government contracts each year, this could weigh on the company's overall growth given the segment's significant contribution to total revenue, and could be particularly detrimental for a company trading at such elevated valuations.
Sovereign AI. The European pushback is no longer merely rhetorical; it is actively affecting pipeline visibility and contract renewals as governments increasingly prioritise digital sovereignty over pure technical capability. In June 2026, it was confirmed that France's domestic intelligence agency (DGSI) is migrating away from Palantir, unwinding a relationship that began after the 2015 Paris attacks. The agency is shifting its mass-data analysis capabilities to a domestic provider, ChapsVision. Following sustained parliamentary concerns over data sovereignty and dependence on a single US defence contractor, NHS England is also officially reviewing exit strategies and evaluating alternative UK suppliers.
Figure 4: Government revenue growth slowing down

What you are paying for at current prices (Negative)
The growth story is compelling. The market expects Palantir to grow its earnings by 83% in FY2026 and 50% in FY2027. Its five-year earnings CAGR is close to 52%, ranking the company among the fastest-growing companies in the US at its current scale. In Q2 2026, the company delivered a Rule of 40 score of 155 (93% revenue growth and a 62% adjusted operating margin), accelerating from 145 in the previous quarter and ranking among the top tier of its peers.
But the real question is whether the price investors are paying today justifies that level of growth.
At the current level of US$175, the shares are currently trading at a forward P/E of 108x for FY2026, 75x for FY2027 and 51x for FY2028. Palantir’s superior revenue growth, customer expansion and operating margins support a premium to the software sector, but not necessarily the premium implied by its current share price.
Our model incorporates assumptions such as revenue growing at a 50% CAGR over the next five years and a 38% CAGR over the next ten years — fairly optimistic assumptions — alongside a WACC of 10.5% and a 4.5% terminal growth rate. This translates into a fair P/E of 43x. Even so, our modelled price arrives at US$150, 14% below the current share price of US$175.
The valuation therefore depends not only on its revenue growth, but also on Palantir converting that growth into durable free cash flow per share.
A slower margin expansion, continued share-based compensation, or weaker customer expansion would dampened investors’ conviction even if headline revenue remains strong.
Simply put, even under a more optimistic scenario, investors are paying too much at current prices. We also studied a previous success case — Nvidia in 2020 — when its revenue was expected to grow at a 47% CAGR over the following five years and the company traded at a forward P/E of 44x, broadly in line with our fair P/E. *However, the Nvidia analogy suffers from survivorship bias. There are many more companies in the past that traded at such valuations but never recovered.
Figure 5: Palantir tops peers in Rule of 40

Table 4: Scenario analysis
|
Bear |
Base |
Bull |
|
|
Next 10 year revenue growth CAGR |
29% |
37% |
40% |
|
Target Price (USD) |
53 |
150 |
172 |
Figure 6: Nvidia used to trade at average forward P/E of 43x when future revenue growth is 50%

5. Upside risk to our call
The beat-and-raise cadence continues and consensus chases. Palantir has beaten its own quarterly guidance by a mean of +6.3% across five observations, with the magnitude widening (+5.9% → +6.5% → +7.6%). Full-year guidance was raised 19.4% through FY2025 and has already been raised 13.4% in FY2026 with two quarters to go. If Q3 2026 revenue prints above $2.30bn (versus $2.16bn guided) and FY2026 guidance moves toward $8.5bn, the stock can rise.
We are wrong about the bookings. Our central non-consensus point is the TCV/RPO divergence. The legitimate rebuttal is that the $10bn Army Enterprise Service Agreement changed the accounting geography of government bookings — task orders may generate revenue without headline TCV events.
6. Initiate SELL; Target Price of USD 150
Palantir Technologies is an exceptional business being sold at an unexceptional risk-reward.
93% revenue growth at a $7.7 billion annualised run rate, 157% net dollar retention, an 86% adjusted gross margin, a 62% adjusted operating margin and a Rule of 40 score of 155% highlight a company with an operating profile that is arguably unmatched in the software sector.
However, the decisive fact is that even assuming a 37% revenue CAGR over the next 10 years, the current share price still appears expensive. Based on the current forward P/E, we believe the stock is overvalued. We derive a target price of US$150, implying downside potential of 14% from the current price, based on a fair P/E multiple of 43x.
Table 5: Valuation
|
USD mil |
2025A |
2026E |
2027E |
2028E |
|
Revenue |
4,475.40 |
8,145.20 |
12,232.20 |
17,507.20 |
|
Growth (% y/y) |
56.20% |
82.00% |
50.20% |
43.10% |
|
EPS |
0.6 |
1.62 |
2.33 |
3.43 |
|
Growth (% y/y) |
224.00% |
154.66% |
43.70% |
47.12% |
|
P/E |
279.24x |
108.08x |
75.19x |
51.11x |
|
Fair P/E |
44 |
|||
|
Target Price |
150 |
|||
|
Upside Potential |
-15.3% |
|||
|
Source: Bloomberg Finance L.P., iFAST compilations. Data as of 19 August 2026. |
||||
Declaration:
This research report was prepared with the assistance of artificial intelligence (AI) tools. iFAST Financial Pte Ltd does not rely exclusively on AI for content generation; the content of this report – including all investment theses, ratings, price targets and conclusions – has been independently reviewed and verified by the research analyst(s) to ensure accuracy and professional integrity.
For specific disclosure, at the time of publication of this report, IFPL (via its connected and associated entities) and the analyst who produced this report hold a NIL position in the abovementioned securities.

