Chapter 3

The Cleared Moat

Booz Allen's defensibility rests on four things that are hard to copy: a security-cleared workforce, incumbency on thousands of task orders, revenue spread so thin no single contract matters, and the position of the federal government's largest AI provider. The clearest evidence that this holds is the recompete win rate — roughly nine in ten every year for six years — and a National Security core that kept growing about 2% while Civil collapsed in fiscal 2026. The moat is real. It is also narrowing at the edges, and this chapter measures both.

What the moat is made of

Booz Allen competes for U.S. government work against enterprise-software firms, global technology providers, the large defense primes, pure-play services contractors, and startups — and it says it wins on technical capability, past performance, its relationships with customers, pricing, and its ability to recruit and retain skilled, security-cleared talent [1]. None of those is a patent. Three of them are structural barriers that take a rival years to reproduce.

The first is clearances. About 77% of the roughly 31,500-person workforce holds a security clearance [1] — a share that has risen since the 2010 IPO, when 71% held clearances and a quarter of the firm sat at the Top Secret/Sensitive Compartmented Information level [2]. A cleared analyst cannot be hired on a day's notice; the vetting is expensive, slow, and government-controlled, which is exactly what makes an installed base of cleared staff a barrier rather than a cost line.

The second is spread. Booz Allen delivered under 5,026 contracts and task orders in fiscal 2026, and about 84% of revenue came from 2,426 active task orders under IDIQ vehicles [3]. The largest single task order was about 4% of revenue and the largest definite contract about 1% [3]. No one lost re-compete sinks the year; the franchise is an aggregation of thousands of small mission footholds, not a handful of mega-programs.

Sources: FY2026 Form 10-K — Our People and Culture and Contracts [1] [3]; win rate p.9 [4]; AI projects p.7 [5].

The third barrier is technical position, and it is the one management is spending against hardest. Booz Allen describes itself as the federal government's largest AI provider, with roughly 400 active AI projects, and frames its VoLT strategy — Velocity, Leadership, Technology — around building specialized products at speed through venture investments and partnerships with hyperscalers and defense-tech firms [5]. That claim is harder to verify from the outside than a clearance count, but it is consistent with where the growth has been.

Nine in ten recompetes, six years running

The single most useful moat gauge Booz Allen discloses is its win rate on re-competed contracts — the work already held, put back out to bid. If incumbency were weak, this is where it would show. It does not. The recompete win rate has sat between 88% and 92% every year since fiscal 2021, and was 89% in fiscal 2026 [4].

Sources: win rates as reported in each fiscal year's Form 10-K — FY2026 and FY2025 [4]; FY2024 and FY2023 [6]; FY2022 and FY2021 [7].

The gap between the two lines is the point. Retaining held work at ~90% is a very different business from winning new work at 56–66%. It says the franchise's defensive perimeter is strong even when its offensive one is ordinary, and it is why the company can lose a competitive-bid cycle and still grow — the base rarely leaks. The new-business win rate is more volatile and has drifted down from 66% in fiscal 2023 to 57% in fiscal 2026 [4], which matters more for the growth case than for the moat.

The national-security core held

The fiscal 2026 revenue decline was not a franchise failure; it was a customer failure, and a specific one. Defense revenue rose to $6.07 billion and Intelligence to $1.90 billion, together about 2% higher, while Civil and Commercial fell 22% to $3.25 billion [8]. The National Security core has grown in each of the last three years; the shock was concentrated where federal-civilian budgets were cut.

Source: FY2026 Form 10-K, Note 3 Revenue — Disaggregation by customer type [8].

Two disclosures reinforce that the core is intact. Total backlog reached a record $38.2 billion at fiscal 2026 year-end, still growing even as revenue shrank [9]. And the demand pull sits with the highest-priority missions: the Department of Veterans Affairs alone was the single largest customer at 10% of revenue [4], and the defense and intelligence work is anchored to warfighter, cyber, and intelligence-community programs that are less exposed to civilian-efficiency cuts.

Where the moat is thinning

A durable moat is not a static one, and three cracks are visible. First, the recompete win rate, while high, slipped from 92% in fiscal 2024 and 2025 to 89% in fiscal 2026 [4]. Three points is within the historical band, but it is a downtick in the metric that most defines the franchise, in the same year the environment turned — worth watching rather than dismissing.

Second, procurement reform cuts against the shape of Booz Allen's book. A fiscal 2026 executive order, "Promoting Efficiency, Accountability, and Performance in Federal Contracting," directs agencies to increase the use of firm-fixed-price contracts for new awards to the maximum extent consistent with law [10]. Cost-reimbursable work was 59% of fiscal 2026 revenue [8]; a durable shift toward fixed-price awards would move risk onto the contractor and pressure the margin structure, however gradually.

Third, the offense is softer than the defense. The new-business win rate has fallen to 57%, and the growth engine now leans on the National Security core continuing to expand while Civil finds a floor. The moat protects what Booz Allen already has; it does not, on its own, guarantee the next leg of growth.

Scale among the named rivals

Booz Allen does not name its competitors in the fiscal 2026 filing, but its own compensation peer group and the broader federal-services set place it against a familiar field. Among the pure-play federal contractors, it is the second-largest by revenue — behind Leidos, ahead of CACI, SAIC, Parsons, and Maximus — and its operating margin, even in a trough year, sits in the upper half of the group.

Sources: most recent Form 10-K per company (fiscal year-ends differ): Booz Allen (year ended March 2026) [11]; Leidos (Jan 2025) [12]; CACI (Jun 2025) [13]; SAIC (Jan 2025) [14]; Parsons (Dec 2024) [15]; Maximus (Sep 2024) [16].

The scale table carries a caveat: these firms are not identical. Leidos is larger but more diversified, carrying health, commercial, and security-products lines outside the pure government-services model [12], and Maximus is weighted toward health and human-services business-process work rather than defense and intelligence. The head-to-head national-security rivals are CACI and SAIC, and Booz Allen is bigger than both. Size is not itself a moat, but in a market where past performance and the ability to field cleared teams at scale are award criteria, being the second-largest player is a structural advantage that compounds with the clearance and incumbency barriers above.

What the evidence supports

The weight of the evidence favours a durable franchise over an eroding one. The recompete win rate — the cleanest external read on incumbency — has held near 90% for six years and through the fiscal 2026 shock; the National Security core grew while Civil fell; backlog set a record; and Booz Allen enters the AI-procurement wave as the largest federal provider and the second-largest pure-play by revenue. That is the raw material of a business the market is pricing as impaired.

The strongest fact against the read is that the moat's best metrics are lagging and its softer ones are turning: the recompete rate ticked down to 89%, the new-business win rate has fallen to 57%, and a procurement shift toward fixed-price awards presses on a book that is still 59% cost-reimbursable. What would change the read is a second consecutive year of falling recompete win rates, or National Security growth stalling — either would suggest the erosion is structural rather than a single hard year. On the evidence to date, the moat is intact and the core is holding; the burden is on the bear to show that the fiscal 2026 cracks widen rather than close.