Defense AI platforms are software systems that ingest classified military and intelligence data — satellite imagery, signals intelligence (SIGINT), sensor feeds, battlefield telemetry — and produce actionable outputs: targeting solutions, predictive analytics, autonomous decision support, and intelligence analysis. The unit of commerce is the government contract, typically structured as a ceiling-value agreement (an upper spending limit that agencies can draw against over multiple years) or a firm-fixed-price task order. The Pentagon requested $13.4 billion for AI and autonomy in FY2026, the first dedicated budget line of its kind. The global defense AI market was $11.5 billion in 2025 and is projected to reach $35.8 billion by 2034 at a 13.4% CAGR, per GM Insights. Four US-listed companies compete primarily as AI platform vendors to this buyer base: Palantir ($365B market cap, $1.63B quarterly revenue), C3.ai ($1.6B market cap, $53M quarterly revenue), BigBear.ai ($2.4B market cap, $34M quarterly revenue), and Leidos ($16.2B market cap, $4.4B quarterly revenue — though Leidos is primarily a systems integrator with AI as one capability within a much larger services business).
The Pentagon is spending $13.4 billion on AI and autonomy in FY2026. The four listed players absorb very different shares: Palantir books $858M/quarter in government revenue across DoD and intelligence, with a $10B Army platform ceiling. C3.ai has a $500M DoD agreement ceiling but only $53M/quarter total revenue across all customers. BigBear.ai runs at $34M/quarter with a $282M backlog weighted toward national security. Leidos earns $4.4B/quarter but is a diversified IT integrator; its defense AI work is one thread in a much larger fabric.
These companies sell software platforms — not hardware, not troops, not consulting hours (with the exception of Leidos, which sells all of these). The core product takes raw military data and makes it usable for human or machine decision-makers. Money flows through several contract structures:
Palantir and C3.ai sell proprietary software platforms where the incremental cost of serving an additional user is near zero (high gross margins — Palantir at 88% adjusted, C3.ai at 37% non-GAAP). BigBear.ai mixes proprietary software with project-based work (34% gross margin, up from 21% a year ago after its Ask Sage acquisition). Leidos is a services business with 11.5% operating margins — it employs tens of thousands of cleared personnel who deliver IT integration, and AI is a capability embedded in those services rather than a standalone platform.
Sources: Palantir Q1-2026 8-K (May 3, 2026); C3.ai FQ3-2026 press release (Mar 2026); BigBear.ai Q1-2026 press release (May 2026); Leidos Q1-2026 8-K (May 6, 2026).
| Contract / Program | Buyer | Ceiling or Value | Type | Status |
|---|---|---|---|---|
| Army Software & Data Platform (PLTR) | U.S. Army | $10.0B | IDIQ, multi-year | contracted 2025; task orders flowing |
| Maven Smart System (PLTR) | DoD cross-service + NATO | $1.3B cumulative | Prototype OTA + mods | contracted $480M (2024) + $99.8M expansion + $795M mod |
| Army Vantage (PLTR) | U.S. Army | $618.9M | Follow-on, $401M base | contracted Dec 2024 |
| TITAN Phase 3 (PLTR) | U.S. Army | $178.4M | Prototype | contracted 10 ground stations |
| DoD Production OTA (AI) | DoD-wide (MDA primary) | $500M | 5-year OTA ceiling | contracted 2021; MDA task orders flowing |
| MACRO II (LDOS) | U.S. Army | $869M | 5-year award | contracted AI-enabled battlefield systems |
| Sole-source classified award (BBAI) | Undisclosed (national security) | $53M | Prime classified | contracted Q1 2026 |
Palantir's total remaining deal value (RDV) was $11.8 billion at Q1-2026 end (+98% YoY). Remaining performance obligations (RPO) were $4.5 billion (+134% YoY). Total Q1 contract bookings (TCV) were $2.4 billion (+61% YoY). These figures span government and commercial; Palantir does not separately disclose government-only RDV or RPO.
BigBear.ai's total backlog was $281.9 million at Q1-2026 end (+14% QoQ), with ~$75 million in Q1 wins concentrated in national security.
Leidos's total backlog was $48.4 billion (funded: $9.6B; unfunded: $38.8B). The Defense segment alone had $12.6B in backlog. Trailing-twelve-month book-to-bill was 1.1x. Q1 net bookings were $3.3 billion.
Sources: CDO Magazine (FY2026 Pentagon AI budget breakdown); C3.ai FQ3-2026 press release; GM Insights defense AI market report (2025); GovConWire (CDAO awards); Palantir Q1-2026 8-K; BigBear.ai Q1-2026 press release; Leidos Q1-2026 8-K.
The supply of defense AI platforms is constrained by three bottlenecks:
Private competitors: Anduril (valued at $12.7B est.) builds autonomous weapons systems with an AI-first architecture. Shield AI (valued at $12.7B est.) builds autonomous drone piloting software. Both are private and not available for public-market investment.
Sources: CDO Magazine; GovConWire; Defense One (clearance processing timelines); Palantir 10-K FY2025; C3.ai FQ3-2026 press release; BigBear.ai Q1-2026 press release.
| Measure | Demand side | Supply side |
|---|---|---|
| Pentagon AI budget (FY2026) | $13.4B appropriated | Top 5 defense AI firms hold 37.9% market share globally est. |
| Software & integration slice | $1.2B (FY2026) | Palantir gov't alone: $3.4B annualized |
| Global defense AI market | $11.5B (2025) → $35.8B (2034) est. | Fragmented across dozens of vendors globally |
| Palantir total RDV (all customers) | $11.8B contracted | Delivering at $1.63B/quarter run-rate |
| Leidos total backlog | $48.4B (funded $9.6B) | Delivering at $4.4B/quarter run-rate (~2.7 years funded coverage) |
| C3.ai federal bookings mix | 55% of bookings (growing 134% YoY) | $53M/quarter total revenue (all customers) |
| New entrants (CDAO commercial AI awards) | $200M to Anthropic, OpenAI, Google, xAI | Not yet operational on classified networks |
Pricing direction: Defense AI contract values are rising. Palantir's Maven program grew from $480M (May 2024) to $1.3B cumulative (by 2025) through modifications. Palantir's U.S. government revenue grew 84% YoY in Q1-2026. C3.ai's federal bookings grew 134% YoY. The Pentagon's $13.4B AI budget is a new line item — prior-year AI spending was embedded and not separately tracked, making growth rates difficult to calculate.
Sources: Palantir Q1-2026 8-K; C3.ai FQ3-2026 press release; GM Insights; CDO Magazine; DefenseScoop (Maven contract increases).
| Metric | PLTR | AI | BBAI | LDOS |
|---|---|---|---|---|
| Market cap | $364.8B | $1.6B | $2.4B | $16.2B |
| Stock price | $152.17 | $11.18 | $5.11 | $126.82 |
| Shares out | ~2,398M | ~145M | 477M | ~126M |
| Total revenue (latest Q) | $1,633M | $53.3M | $34.4M | $4,400M |
| Revenue growth (YoY) | +85% | declining (FY guidance $247-251M) | -1% | +4% |
| Gov't revenue (latest Q) | $858M (53% of rev) | ~55% of bookings ($ not disclosed) | Majority est. ~85%+ | ~$4,200M est. ~95% |
| Adjusted operating margin | 60% | Neg (non-GAAP loss ~$56-64M/Q) | Neg (adj. EBITDA -$9.9M) | 14% (adj. EBITDA) |
| GAAP net income (latest Q) | $871M | Net loss | -$56.8M | $335M |
| Backlog / RDV | $11.8B (RDV); $4.5B (RPO) | Not disclosed | $281.9M | $48.4B total; $9.6B funded |
| Cash + investments | $8.0B | $621.9M | $431.5M | $457M |
| Total debt | ~$0 | ~$0 | $16.6M | $6.3B |
| Gross margin | ~88% adj. | 37% non-GAAP | 34% GAAP | ~27% implied est. |
| Adj. free cash flow (latest Q) | $925M | Negative | Negative ($18M op. cash burn) | $270M |
| Key defense contract ceiling | $10B (Army) | $500M (DoD OTA) | $53M classified (Q1) | $869M (MACRO II) |
| Defense AI purity | Pure-play platform (53% gov't) | Enterprise AI with defense pivot (~55% of bookings) | Pure-play defense/intel AI | Diversified IT integrator; AI is one capability |
Palantir is 200x+ larger than BBAI and AI by market cap but generates 20-50x their revenue. PLTR and LDOS are profitable and cash-generative; AI and BBAI burn cash. Palantir has the highest margins (88% gross, 60% operating). Leidos has the largest absolute backlog ($48.4B) but operates at services-business margins (14% EBITDA). BBAI settled $124.6M of convertible debt via equity conversion in Jan 2026 and has $16.6M of debt remaining against $431.5M of cash and investments. C3.ai has $622M of cash; full-year FY2026 non-GAAP operating loss is guided to $220-228M.
Sources: Palantir Q1-2026 8-K; C3.ai FQ3-2026 press release; BigBear.ai Q1-2026 press release; Leidos Q1-2026 8-K; StockAnalysis.com (market data, Jun 2 2026).
What it costs to own each company, stated as arithmetic. Prices as of June 2, 2026.
| Valuation metric | PLTR | AI | BBAI | LDOS |
|---|---|---|---|---|
| Market cap | $364.8B | $1.6B | $2.4B | $16.2B |
| Enterprise value (mkt cap - cash + debt) | ~$356.8B | ~$1.0B | ~$2.0B | ~$22.0B |
| Price / trailing revenue (annualized last Q) | 55.9x | 7.5x | 17.5x | 0.92x |
| Price / FY2026 guided revenue | 47.6x ($7.66B) | 6.5x ($249M mid) | 16.0x ($150M mid) | 0.89x ($18.2B mid) |
| Price / trailing GAAP net income (ann.) | 104.7x | Neg | Neg | 12.1x |
| Price / adj. free cash flow (ann.) | 98.7x | Neg | Neg | 15.0x |
| EV / gov't revenue (ann.) | 103.9x | N/A (not disclosed) | ~14.5x est. | ~1.3x est. |
In cash terms: To buy all of Palantir at $365B, the company would need to generate ~$18.2B/year of owner cash to produce a 5% annual yield on purchase price — that is 2.8x its current annualized revenue of $6.5B and 4.9x its current annualized adjusted free cash flow of $3.7B. For Leidos at $16.2B, a 5% yield requires $810M/year — below its trailing annualized net income of $1.34B. For BigBear.ai, the $2.4B price corresponds to a company generating $138M/year in revenue at negative margins. For C3.ai, the $1.6B price corresponds to $213M/year in revenue (FY2026 annualized) with $220M+ in annual operating losses; the $622M cash pile provides roughly 3 years of runway at the current burn rate.
Palantir's FY2026 revenue guidance of $7.65-7.66B implies 71% growth and a forward P/S of ~48x. Rule of 40 score (revenue growth % + operating margin %): 145. Leidos trades at under 1x revenue; government services contractors historically trade between 0.7-1.5x revenue. est.
Sources: market data Jun 2, 2026 (StockAnalysis.com, CompaniesMarketCap.com); company filings as cited above.