Palantir Valuation Insights (May 9, 2025)

Welcome, fellow economics, finance, and stock market enthusiasts! This is my very first post on this site, and it reflects my passion for corporate finance and valuation. In today’s brief write‐up, I’ll be sharing an analysis of Palantir and its valuation as of May 9, 2025. Moving forward, I plan to publish more analyses, valuations, and insights here—so stay tuned!

I. Company overview

 

a. Who they are?

Palantir is a U.S.-based software and analytics company founded in 2003 and publicly listed in 2020. Known initially for its ties to the U.S. intelligence community, Palantir has become an influential provider of platforms that unify and analyze massive datasets for both government and commercial clients. 

b. Flagship products (platforms)

Table 1. The flagship products of Palantir

Platform	Core use-case	Recent headline wins
Gotham	Real-time intelligence fusion for defence, law-enforcement & intel.	U.S. Army TITAN targeting trucks delivered Mar-2025.
Foundry	Data operating-system for enterprises & governments.	New roll-outs at Airbus, BP, Ferrari, Merck, NHS England.
AIP / Apollo	Secure orchestration of LLMs from cloud to edge; “Mission Autonomy” for drones & sensors.	Integrated with GothAM on TITAN; commercial pilots at Jacobs, GE, J.D. Power.

 

c. Key financial

Table 2. Palantir’s key financials overview

Metric	Latest figure
Revenue (LTM)	$ 3.12 billion
Y/Y Revenue growth	+ 33 % vs LTM to 31 Mar 2024 ( $ 2.33 b )
Segment mix	Government 55 % (≈ $ 1.72 b) / Commercial 45 % (≈ $ 1.39 b)
Adjusted after tax operating margin (Q1-25)	17.73 %
Sales / Capital (Q1-25)	2.66
Cash, cash equivalents + U.S. Treasuries	$ 5.4 billion
Market capitalisation (May 9 2025)	≈ $ 277 billion
Diluted share count drivers	164 m options (w.avg strike $ 9.62) + 6 m SARs (strike $ 56)

 

d. Future developments.

Safe Cities with Less Policing. By using advanced data analysis and real-time monitoring, Palantir suggests societies may eventually prevent crime before it escalates, reducing the reliance on large police forces.

Unmanned Battlefields. Palantir’s capabilities in integrating sensor data and coordinating autonomous systems hint at military operations with fewer on-the-ground soldiers, replaced by drone swarms or other AI-directed platforms.

“End of the Nuclear Era”. CEO Alex Karp has floated the idea that AI-driven surveillance and defense systems could eventually detect and disable threats preemptively, making nuclear arsenals less central to global security.

Golden Dome. While not officially confirmed, rumors of a next-generation “shield” or defensive architecture, fusing real-time data from satellites, sensors, and AI analytics – would align with Palantir’s role as a data-integration backbone for large-scale defense initiatives.

Though these scenarios are speculative, they reflect Palantir’s broader ambition: leveraging data fusion and AI to transform how governments and industries address safety, conflict, and strategic deterrence.

II. Intrinsic Valuation

 

Narratives and Numbers

 

1. Revenues. For a high-growth company like Palantir, valuation hinges on the “end game”: the projected size of the AI and enterprise-software markets in the terminal year and the slice of those markets Palantir can realistically capture.

Palantir groups its revenue into three regions: United States, United Kingdom, and Rest of World, but its mission orientation, security clearances, and deep defence ties make Western democracies its natural customer base. We therefore assume the United States and United Kingdom will remain its core markets, with smaller yet significant opportunities in Canada, Australia, Japan, and the European Union.

The first task, then, is to forecast the growth of the enterprise-software and AI markets across these jurisdictions – separately for commercial and government segments and to overlay Palantir’s likely market share on those projections.

Table 3. Enterprise software revenue projections (2024–2034)

Region	2024 Revenue(US $ bn)	Palantir 2024* (US $ bn)	Palantir 2024 Share*	2034 Market Size(US $ bn)
United States	420	0.95	0.226 %	1,304.50
United Kingdom	40	0.15	0.370 %	103.7
Rest (Canada, Japan, Australia, EU)	285	0.30	0.104 %	656.4
Combined total	745	1.39	0.187 %	2,064.60

  *Palantir reports revenue only by aggregated geographic segments and does not break out government versus commercial revenue within those regions. Consequently, the commercial share is estimated by allocating total revenue across regions using the same weighted geographic distribution.

Sourse: Statista, Gartner (via Next Platform), Veridion, Expert Market Research, Grand View Research

Table 3 is summurising the size of the projected software market in the 2034.

The next step is to estimate the size of the government software and AI market in these same regions.

Table 4. Projected spending on IT & software, communications and C4ISR, cyber operations, and artificial intelligence for the defence, intelligence, healthcare, and law-enforcement sectors in the listed regions in 2034.

Country / Region / Organisation	Total	Palantir Optimistic		Palantir Pessimistic
		USD Billion	Share of Total, %	USD Billion	Share of Total, %
USA	169.56	74.12	43.71%	36.42	21.48%
UK	32.63	11.00	33.72%	5.84	17.90%
NATO (NSIP)	4.54	2.27	50.00%	0.91	20.00%
EU	105.37	15.91	15.10%	7.09	6.73%
Canada	11.62	2.16	18.55%	1.03	8.90%
Australia	22.84	3.75	16.44%	1.70	7.43%
Japan	28.69	3.07	10.69%	1.63	5.67%

Sourse: Based on government data, analyst forecasts, and our best estimates.

 

Table 4 summarizes projected spending in 2034.

With the “end game” market now mapped, i.e., how each segment is expected to look ten years from today – we can frame Palantir’s potential position. We identify five illustrative market-share scenarios for both the commercial and government arenas: Small, Bigger, Decent, Massive, and Diabolical – yielding 25 combined outcomes.

Tables 4 and 5 present these scenarios: Table 4 shows Palantir’s prospective share of the total enterprise-software market, while Table 5 shows its share in the government sector.

 

 

Table 5. Expected Palantir’s commercial sector share in 2034

Expected Commercial contracts Revenues in 2034	$ Revenues in 2034 (in $Bil)
A1: Small (US/UK = 1.5%; Rest = .5% of total market)	$24
A2: Bigger (US/UK = 2.5%; Rest = .5% of total market)	$38
A3: Decent (US/UK = 3.5%; Rest = 1.5% of total market)	$59
A4: Massive (US/UK=5%; Rest = 2% of total market)	$84
A5: Diabolical (US/UK/Rest = 5% of total market)	$103

 

Table 6. Expected Palantir’s government sector share in 2034

Expected Government contracts Revenues in 2034	$ Revenues in 2034 (in $Bil)
AA1: Small (US only pessimistic)	$36
AA2: Bigger (US, UK, NATO pessimistic)	$43
AA3: Decent (US, UK, NATO optimistic)	$87
AA4: Massive (US, UK, NATO optimistic + EU, Canada, Japan, Australia pessimistic)	$99
AA5: Diabolical (US, UK, NATO,  EU, Canada, Japan, Australia optimistic)	$112

 

Palantir is likely to capture additional share in the commercial arena – recent reports point to a partnership with a global investment bank to build a next-generation financial-data platform and its AI toolkit now spans logistics-optimisation engines, predictive-maintenance models, and full supply-chain digital twins.

Even so, we expect the company to stay true to its founding mission: securing democratic institutions through government work. That mission focus will temper the pace of commercial expansion and anchors our base-case Scenario A2.

On the government side, the evidence is already visible:

       •Defense and intelligence.  New multi-year task orders under the U.S. Army TITAN and DoD Maven/Trident programs, plus NATO pilot projects for joint-operational AI.

       •Healthcare. The U.K. NHS Federated Data Platform is scaling nationwide, and a U.S. Veterans Affairs proof-of-concept is in flight.

       •Law enforcement. Expanded deployments of Foundry-powered analytics with U.S. federal and state agencies, and early adoption by select European public-safety bodies.

We therefore expect contract volume to surge in the United States, the United Kingdom and NATO, with more modest uptake in Canada, Australia, Japan and parts of the EU – many of which are choosing to build sovereign AI stacks for public workloads.

Putting these elements together, our outlook aligns with Scenario A2 / AA4:

Government growth remains the primary engine, especially in defence, intelligence, healthcare and law enforcement, while commercial wins add steady but secondary momentum.

Based on Palantir’s current position and governments’ preparations for the AI era, we expect revenues to rise markedly over the next two years, with a major surge in year three once public-sector clients embark on large-scale AI rollouts and Palantir’s expanded capacity can meet that demand.

 

2. Operating Margin. To determine Palantir’s target operating margin, we benchmark the U.S. software sector. Table 6 lays out the margin ranges that peer companies currently achieve. Our base case assumes that, once Palantir reaches maturity, its commercial business will operate efficiently enough to earn a margin above the sector median – roughly the sixth decile (labelled B3 in Table 6). Government contracts, by contrast, are constrained by procurement rules and disclosure requirements, so we cap that segment at the sector’s first-quartile margin of 26.6 percent.
 

Table 7. Target Operating Margin in year 10

Target Operating Margin
Operating Margin in 2034	Commercial Sector	Government sector	Aggregate
B1: US software First Quartile	27%	27%	27%
B2: US software median	35%	27%	29%
B3: US software 6th decile	38%	27%	30%
B4: US software 7th decile	42%	27%	31%
B5: US software Third Quartile	48%	27%	33%
B6: US software 9th decile	57%	27%	35%

 

3. Reinvestment. To gauge Palantir’s future reinvestment needs, we focus on the Sales-to-Capital ratio, which indicates how much capital is required to generate a dollar of revenue. Table 7 summarises sector benchmarks and our own adjustments. Palantir’s current ratio is relatively high at 2.66, but we expect it to fall as the company builds out capital-intensive AI infrastructure. Even so, we project Palantir will remain efficient, settling at a long-run Sales-to-Capital ratio of about 1.5. A sales-to-capital ratio of 2.50 (C7) may look attractive in a model, but sustaining that efficiency in practice is unlikely.

Table 8. Sales to Capital ratio in year 5

Scenario	Sales to Capital (from year 5)
C1: Mature	0.74
C2: All software US median	0.76
C3: 8th decile	0.96
C4: 9th decile	1.21
C5: Estimation	1.50
C6: Software application sector Prof. Damodaran	1.71
C7: Diabolical efficiency	2.50

 
4. Cost of Capital. Picture 1 shows how the cost of capital was derived. The risk-free rate is the yield on the ten-year U.S. Treasury. For the equity-risk premium we apply the implied ERP for the S&P 500 as of 1 May 2025. Country adjustments are handled by assuming no additional risk premium for the United States or the United Kingdom, while a one percent country-risk premium is added to the “Rest of World” segment and then blended into the overall ERP in proportion to revenue exposure. Beta comes from a bottom-up average of all publicly listed U.S. software peers. Palantir has no traditional bonds, so its cost of debt is estimated through a synthetic AAA rating (based on interest coverage ratio), which translates into a 0.59 percent CDS spread; that figure is used as the pre-tax cost of debt. These inputs feed into the weighted-average cost of capital. 

Country	Revenues	ERP	Weight	Weighted ERP
United States of America	2122.352	4.58%	68.13%	3.12%
United Kingdom	331.028	4.58%	10.63%	0.49%
Rest of the World	661.644	5.58%	21.24%	1.19%
Total	3115.024		100.00%	4.79%
	Bottom up unlevered beta 1.186513593 Levered Beta for equity 1.187113548 Pre-tax Cost of Debt 4.96% Tax Rate 25%
	Equity	Debt	Capital
Market Value	$            276,816.27	$       186.63	$     277,002.90
Weight in Cost of Capital	99.93%	0.07%	100.00%
Cost of Component	10.06%	3.72%	10.05%

Some might argue that the original CAPM relies on a regression‐based beta—essentially the stock’s slope relative to the market—and that this should be a key variable. Indeed, Palantir’s regression beta at the time of analysis was around 2.7. However, the R‐squared was only about 0.2, suggesting that many non‐market factors or unique risks are driving Palantir’s price. As we see below, fundamental indicators also don’t adequately explain its price. Given the company’s distinct profile and high uncertainty, that regression beta mainly reflects the market’s rapidly changing mood and momentum, struggling to interpret new global or local developments.

5. Valuation. All of the assumptions described above form the basis of the DCF valuation presented in Picture 2, which also includes the remaining model details. 

6. Sensitivity Analysis. Intrinsic value hinges on two levers: future cash flows (the numerator) and the discount rate, i.e., cost of capital (the denominator). For Palantir, we view the cost of capital as reasonably straightforward, whereas cash-flow growth is far less certain. In high-growth companies, the narrative (what story actually plays out) drives the widest swings in valuation. To test that narrative risk, we ran a sensitivity analysis across the scenario sets defined in Tables 4 and 5. Table 8 summarises the resulting per-share values for each combination. This grid lets us see (i) how the valuation shifts under alternative stories and (ii) which scenarios square with Palantir’s current market price, signalling where investors think the company is headed.

Table 8. The difference in value per share across the various story sets.

	AA1	AA2	AA3	AA4	AA5
A1	43.06	47.18	74.22	81.23	89.44
A2	56.08	60.21	87.25	94.25	102.47
A3	75.18	79.31	106.35	113.35	121.57
A4	97.8	101.89	128.93	135.93	144.14
A5	115.98	120.10	147.14	154.14	162.36

 

As Table 8 shows, shifts along the vertical axis – the commercial-sector scenarios – move the per-share valuation more than shifts along the horizontal axis, which capture government scenarios. The reason is straightforward: government contracts carry thinner margins, so even large swings in public-sector market share translate into smaller changes in free cash flow than equivalent moves on the commercial side.
The grid also indicates that the market is currently anticipating faster at least commercial growth than our base case. The earliest combination that reconciles with today’s share price is a “Diabolical” government story paired with a “Decent” commercial story; anything less optimistic on the commercial axis leaves the intrinsic value below the market quote. 
We nevertheless retain our present assumptions, keeping the per-share value as for A2/AA4 story set.
Operating margin and reinvestment are the other key determinants of FCFF. Using the margin ranges in Table 6 and the sales-to-capital scenarios in Table 7, we ran a sensitivity analysis; Table 9 summarizes the corresponding per-share values.

Table 9. The difference in value per share across the various margins and sales to capital ratios

	B1	B2	B3	B4	B5	B6
C1	65.40	74.87	77.94	82.58	90.14	99.23
C2	66.47	75.54	79.01	83.65	91.21	100.30
C3	72.64	82.11	85.17	89.81	97.38	106.46
C4	77.75	87.22	90.28	94.92	102.49	111.57
C5	81.72	91.19	94.25	98.89	106.46	115.54
C6	83.83	93.30	96.37	101.00	108.57	117.65
C7	88.69	98.16	101.22	105.86	113.43	122.51

 
The sensitivity results show limited upside because we hold the government-sector margin at roughly 27 percent across all cases. Given that constraint and our base revenue narrative (Scenario A2 / AA4), the only path to today’s market price is a commercial segment that sustains an operating margin of about 57 percent and a sales-to-capital ratio of at least 1.71.

 
…narrative shifts?
The latest quarterly report shows U.S. commercial revenue up more than 70 % and flags a potential partnership with xAI for AI-infrastructure support. Growth outside the United States, however, has been weaker than expected. To account for this, we constructed an alternative narrative in which Palantir deepens its ties with xAI. Outsourcing the infrastructure would raise the sales-to-capital ratio, because Palantir would no longer need to build AI-center capacity in-house – but it would also trim operating margins, since xAI would capture part of it. Commercial-sector revenue growth accelerates under this view, while non-U.S. clients are less likely to engage. 
The numerical impact of this narrative is presented in the end of the post. Spoiler: value is $88.53 under these circumstances.

   III. Pricing Game

Running after the market

 

1. Comparable approach.

The most straightforward way to price Palantir is to examine how the market prices comparable companies in the same sector. To standardise performance across firms, analysts typically use valuation (pricing) multiples. Because of Palantir’s specific characteristics, we avoid unconventional metrics such as price-per-user and instead rely on classical earnings and book-value multiples, benchmarking them against U.S. software peers.
To keep the sample relevant, we include only U.S. software companies with market capitalisations above $190 million; the average market cap in the group is about $35 billion. The sample spans firms with both positive and negative net income. Because Palantir is a high-growth company, we restrict the peer set to U.S. software firms with similar three-year revenue CAGRs, as reported in the Capital IQ database. Palantir’s own three-year CAGR is 23.7 %, so we include only companies whose growth falls between 15 % and 34 %.

Picture 3. Comparison of Palantir’s multiples with sector averages.

 

Even after factoring in the revenue growth analysts project for Palantir over the next three years, the stock still appears significantly overpriced on all five core multiples.

 

2. Regression approach. Another way to gauge fair price is to run a cross-sectional regression that explains today’s prices for a peer set and then apply the resulting equation to Palantir. After testing several specifications, we found that EV-to-Sales is the most reliable dependent variable for U.S. software companies comparable to Palantir.

Sample: 38 U.S. software firms with market capitalisations ranging from $231 million to $3.2 trillion.

Independent variables:

            1.         Three-year revenue CAGR (growth) – enter percentages as plain numbers

            2.         Five-year beta (market risk)

            3.         Two-year beta R² (significance of market risk in beta)

            4.         After-tax operating margin (ATOM) – enter percentages as decimals numbers

These four factors produced the strongest and most stable regression fit among the models we tested. 

Picture 4. Multiple regression of EV / Sales

It is interesting that we found that the t-statistic for the two-year beta R² is higher than for the five-year beta R², while the five-year beta itself is more significant than the two-year beta. This suggests the market prices long-term risk (captured by the longer beta) but also pays attention to shorter-term sensitivity (reflected in the two-year beta R²). In other words, the regression indicates that the market is punishing companies for higher risk (a negative coefficient on beta) while rewarding them for a higher beta R², which signals less firm-specific, non-market risk. It is also worth noting that the ROC is not statistically significant in the current sample.
Even when we replace the analyst consensus (reported by Capital IQ) with our own three-year revenue CAGR estimate of 102.4 %(Scenario A2/AA4), the regression yields the following result for Palantir:
EV/Sales = 13.58
Enterprise Value (mm) = $42,302.90 
Equity (mm) = $28,492.07
Regression Price per share = $13.52

The regression results paint an unflattering picture for Palantir’s outlook—particularly relative to its current share price. To justify today’s market price under this model, Palantir would need a three-year revenue CAGR of roughly 590 %. Moreover, the regression’s R² is only 58 %, underscoring that Palantir is a unique company whose price is driven by factors beyond the fundamental variables captured in our model. However, the limited influence of fundamentals on Palantir’s share price also suggests that the market may not be valuing the company rationally.

3. Terminal multiple method. Among the various pricing methods, the terminal (exit) multiple approach stands out as especially well‐grounded in economic and financial considerations. It explicitly addresses the “endgame,” in which the firm reaches its mature state: one projects the company’s operating metrics in that final, steady‐state year and applies a sector‐appropriate multiple. This yields a future enterprise (or equity) value, which is then discounted back at the current cost of capital – ensuring that the analysis captures both the firm’s long‐run outlook and the time value of money.
For Palantir’s pricing, we opted to employ an EV/EBIT multiple. As a first step, we identified five mature U.S. software companies – those that best capture the steady‐state outlook in this sector and reviewed their EV/EBIT multiples. As shown in the picture below, the median EV/EBIT is approximately 28.87, and we will use this figure.

Picture 5. Mature benchmark companies in the U.S. software sector

Some might note that the sample companies, though mature and software-focused, are not perfect analogs for Palantir. While a few may handle defense‐related government contracts, none offer precisely the same solutions Palantir is developing (see company overview). In practice, it is nearly impossible to find an exact counterpart for any firm, but in a pricing approach, we concentrate on how the market views and values companies that are at least somewhat comparable. Since Palantir is broadly considered—and positions itself—as a software company (rather than a defense contractor), we currently compare it to established, mature peers in the software sector.

For the terminal EBIT we took the figure of our basic (A2/AA4) narrative for the terminal year = $42,591 (mm)

Which gave us terminal EV = $1,229,727.10 (mm)

Present value of EV (default cost of capital: 10.05%) = $471,755.10 (mm)

Equity price = $457,944.28 (mm)

Price per share = $194.05

We also ran a sensitivity analysis for every scenario in Tables 4 and 5 to show how the share price changes under different assumptions and thus different terminal EBIT levels.

Table 10. Variation in share price across the alternative narratives and terminal-EBIT assumptions.

	AA1	AA2	AA3	AA4	AA5
A1	86.68	95.48	153.13	168.05	185.57
A2	112.68	121.48	179.13	194.05	211.57
A3	150.80	159.60	217.25	232.17	249.69
A4	195.9	204.66	262.31	277.23	294.75
A5	232.21	241.01	298.66	313.59	331.11

 

Our analysis shows that, in almost every scenario, the terminal-multiple approach produces a price above Palantir’s current market price. Unless the market changes how it prices mature software companies, Palantir therefore appears to be underpriced.

IV. Conclusion?

 

 

We have conducted an extensive analysis of Palantir: examining the company itself, its market environment, and forecasts for future spending and overall industry direction. Throughout this valuation and pricing exercise, we have seen just how powerful a narrative can be, particularly for a high‐growth firm with significant uncertainty. In the intrinsic valuation section, we presented a sensitivity analysis demonstrating that our beliefs about the global, market, and company outlook strongly determine the valuation and can also be one of our biggest sources of bias.
This dynamic aligns with the professor’s “Bermuda Triangle of Valuation” (Bias, Complexity, and Uncertainty): the complexity and uncertainty of global relations, AI implementation, and the identification of future winners and losers is almost overwhelming. I also recognize my own bias, since I hope for a better future – one without soldier casualties, safer cities with minimal policing, and the end of the nuclear era, so I am inclined to like this company and want it to succeed. However, we do have frameworks and checks: valuation diagnostics, the possible‐plausible‐probable lens, and other tests that help keep us grounded in reality. By walking through this valuation, I hope to have shown how and why I tried to manage these biases and ensure the final narrative is both defensible and realistic.
As we observed, the narrative also played a crucial role in Palantir’s pricing. The regression analysis suggests that its current stock price appears largely disconnected from fundamental drivers such as market risk, revenue growth, and operating margin. A simple comparables review likewise indicates that the stock trades at a premium relative to sector peers – yet the market evidently sees something different in the company. The terminal multiple approach can clarify what expectations the market truly holds for Palantir: namely, that it will eventually succeed and mature into a large, stable enterprise, justifying the elevated price.

 

The recommendation

I follow a value‐oriented investment philosophy and therefore rely heavily on intrinsic valuation, believing it offers the most accurate picture of a company’s worth. Nevertheless, that value matters little unless there are market participants willing to pay for it, so we cannot completely ignore pricing dynamics.

After reviewing our valuation and pricing analyses, I do not recommend buying Palantir at its current price if you are not pursuing an active trading strategy. For existing shareholders, my advice would be to sell at least half of your holdings, while retaining a portion. The market’s mood, momentum, and expectations (as captured by terminal multiples) could still drive the stock higher. However, as with any company (and especially with Palantir) both the narrative and any recommendations are in constant flux. I will revisit Palantir’s valuation if significant new information emerges or once its next earnings report is released.

 
Theoretical assumption
Assuming the market is efficient for mature companies and that our base narrative holds, we can infer an implied failure rate based on the drop from $194.05 to $94.25. This suggests an overall failure rate of 51.17% over 10 years, which translates to an annual failure rate of approximately 6.92%.

As promised, here is the alternative narrative in which Palantir forges a deep collaboration with xAI. Given the current global controversy surrounding xAI’s owner, I’ve titled this scenario “The Isolation”.

The detailed consequences and corresponding valuation appear in the chart below. Under this scenario, I assume “Diabolical Efficiency,” driven by a sales‐to‐capital ratio of 2.5, made possible through xAI’s infrastructure. However, I also project margins decreasing to second‐decile levels for the sector.