Science Applications International Corporation (SAIC)

Deep Dive Research Report

Listing: Nasdaq Global Select Market, ticker SAIC (transferred from NYSE effective March 5, 2024, per the Nasdaq trader notice) Headquarters: Reston, Virginia Fiscal year end: Friday closest to January 31 (FY2026 ended January 30, 2026) Report date: September 3, 2026 Most recent reported period: Q2 FY2027, released August 31, 2026

1. What The Company Does

SAIC sells the labour, engineering judgment and software of roughly 23,000 cleared people to the United States federal government. It does not build fighter jets, satellites or missiles. It builds, runs, modernises and integrates the systems around them - the networks that carry intelligence, the simulations that test a missile before it is fired, the data platforms that let an analyst find a needle, the cloud that hosts an Air Force's applications, the software that decides whether a truck at the border gets pulled over.

Per the FY2026 Form 10-K, approximately 98% of revenue comes from the U.S. government, split roughly 52% Department of War (the filing uses the department's post-2025 name) and 46% intelligence and other federal agencies, with only a residual slice from commercial, state, local and international customers. The company holds approximately 1,700 active contracts and task orders. Over 5,800 of its employees - about 25% of the workforce - are military veterans or active duty, and the majority hold active security clearances.

The value proposition is deceptively simple and structurally hard to replicate: the government cannot buy this work from a normal IT company. To do it you need people who have been cleared to Top Secret/SCI (a process that can take a year or more and cannot be accelerated with money), a facility security clearance, a government-audited accounting system that can survive a DCAA audit of indirect cost pools, a past-performance record on prior contracts, and the institutional knowledge of how a specific agency's legacy systems actually work after two decades of accretion. Accenture and IBM can compete for the commodity end of this. Almost nobody can compete for the classified end.

Founding and the two identities

The original Science Applications International Corporation was founded in 1969 by Dr. J. Robert Beyster, who built it around an employee-ownership model - stock distributed broadly to staff, an internal market for the shares, and a set of operating principles designed to make technical staff behave like owners. Per the Beyster Institute at UC San Diego's Rady School, the company started with a $500 contract and early work studying the effects of underwater explosions for the Navy. By Beyster's retirement in 2004 it had grown past 40,000 employees and roughly $8 billion in revenue.

The company that trades as SAIC today is the younger half of a 2013 split. On September 27, 2013, SAIC, Inc. separated into two companies: the parent renamed itself Leidos and kept the capability-development, product and platform businesses; the roughly $4 billion services unit was spun off and kept the SAIC name and heritage.

The reason for the split is the single most important structural fact about this company, and it explains its economics to this day. Federal acquisition rules contain organizational conflict of interest (OCI) provisions: a contractor that advises an agency on what to buy, or that writes the requirements or evaluates the bids, is generally barred from selling the thing being bought. As the company grew, its advisory and systems-engineering work was progressively locking it out of bidding on the hardware and platform work, and vice versa. Splitting the company freed each half to bid where the other could not.

The consequence is that SAIC deliberately sits on the services and integration side of that line. It is an "unbiased" advisor and integrator by design. That is why it has almost no product IP of the kind Lockheed or Northrop carry, why its revenue is overwhelmingly labour-based, and why - as we will see - its recent strategic project has been trying to get less labour-based without violating the boundary that gave it a business in the first place.

What the work actually looks like

Take the U.S. Army Corps of Engineers programme called RITS (Revolutionary Information Technology Services), which SAIC won in 2021 and lost in 2025-26. Under RITS, SAIC ran the IT environment supporting roughly 37,000 users across 43 geographic districts and about 1,500 field and project offices worldwide. That means service desks, network operations, endpoint management, cybersecurity monitoring, application hosting, and the migration of legacy systems to cloud - executed under a task order against a GSA vehicle, staffed by badged SAIC employees sitting in Corps facilities, billed largely on a cost-reimbursable basis with a fee on top.

Now take the other end of the portfolio. SAIC holds a position on the COMET multiple-award IDIQ with the Missile and Space Intelligence Center, announced subsequent to Q2 FY2027 and estimated at $14 billion of ceiling value across all awardees. Under COMET, per the August 31, 2026 earnings release, SAIC would "provide expertise to develop and maintain hardware, software, systems, and military intelligence capabilities" - foreign missile system exploitation and threat modelling, done inside a SCIF, by people who cannot describe their work to their families.

Both are SAIC. The first is the business the company is trying to shrink. The second is the business it is trying to grow. The gap between those two sentences is the entire investment story.

2. Business Segments

SAIC reorganised into its current two reportable segments in February 2024. Note that the reporting structure (two segments) is distinct from the operating structure (business groups), which was consolidated separately in November 2025 - discussed at the end of this section.

Revenue mix for FY2026 (ended January 30, 2026): Defense and Intelligence ~77%, Civilian ~23%.

2.1 Defense and Intelligence (~77% of revenue)

What it does. This segment serves the Department of War and the U.S. Intelligence Community. The work spans four broad clusters. First, mission systems engineering and integration: helping the Navy, Air Force and Army specify, integrate, test and sustain the software and electronics inside weapons systems and command-and-control environments. Second, training and simulation: building live, virtual and constructive (LVC) simulation environments so services can rehearse operations without expending ordnance. Third, space and intelligence: ground systems for satellite constellations, signal and imagery exploitation tooling, data platforms for all-source analysis, and technical advisory work for space acquisition programmes. Fourth, enterprise and mission IT for defence customers: networks, cloud migration, cybersecurity and service management on military installations.

Named programmes give the flavour: the Air Force HOPE 2.0 award (~$928M, disclosed Q2 FY2026); a Navy training programme (~$202M, Q2 FY2026); Navy airborne electronic warfare test and evaluation support (~$130M over five years, Q2 FY2027); an Army systems-engineering, LVC and modelling-and-simulation contract supporting all service branches (~$330M over five years, Q2 FY2027); a $249M contract supporting the Naval Air Warfare Center Weapons Division instrumentation platform; and a ~$400M five-year recompete win with an unnamed U.S. intelligence agency announced August 5, 2026. At the TD Cowen conference, CFO Prabu Natarajan named cloud-based command and control for NORAD airspace visualisation, radar modernisation under GMAS, and CEDA advisory work supporting space procurement as differentiated growth vectors.

The core capability. Two things took years to build and cannot be bought. The first is the cleared workforce and the facility infrastructure: SCIF space, a facility clearance, and thousands of people already adjudicated at TS/SCI with polygraph. When an intelligence customer needs 200 cleared engineers, the binding constraint is not price, it is whether a bidder can actually produce them. The second is incumbency knowledge: on a 15-year programme, the incumbent knows the undocumented interfaces, the political landmines and the customer's actual (as opposed to written) requirements. This is why SAIC's recompete win rate outside commoditised enterprise IT runs at 85-90% and exceeded 90% in Q2 FY2027 (Q4 FY2026 and Q2 FY2027 calls).

Why it exists separately. Beyond the obvious customer split, the segments have genuinely different economics and contracting cultures. Defense and Intelligence is dominated by cost-reimbursement contracting - the government pays allowable costs plus a fee, which caps margin but also caps risk. That produces high-single-digit adjusted operating margins. Classified work also carries physical, personnel and audit overheads that Civilian work does not.

Competitive position within the segment. SAIC competes here against Leidos, CACI, Booz Allen, Peraton, ManTech, Parsons and the services arms of the platform primes (General Dynamics IT, Northrop, Lockheed). It wins on incumbency, on cleared-workforce depth, and on being platform-agnostic - it is not trying to sell the customer its own hardware. It loses when the work is commoditised network operations, where lower-cost integrators and the large commercial IT firms bid aggressively.

Role in the group. This is the ballast and the scale engine, but it is not the margin engine. Management's framing across FY2026 and FY2027 has been that within D&I there are two very different businesses stapled together: a differentiated mission business earning attractive returns, and a commoditised defence enterprise IT business earning far less. The strategy is to grow the first and let the second run off.

2.2 Civilian (~23% of revenue)

What it does. Federal civilian agencies plus a modest state-and-local practice. Customers include the Department of Homeland Security and Customs and Border Protection, the Department of State, the Treasury, NASA, the Government Accountability Office, veterans' health IT, and county governments. Work spans citizen-facing service delivery (benefits and claims processing), border security systems, enterprise IT modernisation, cloud migration and application development.

Recent named awards: a Treasury cloud services award (~$728M, Q2 FY2026); a State Department Vanguard award (~$547M, Q2 FY2026); an Orange County IT services contract (~$164M, Q2 FY2026); a GAO IT solutions contract (~$95M, new work, Q4 FY2026); a DHS award (~$200M over five years, Q1 FY2027); and, announced after Q2 FY2027, a ~$740M five-year DHS recompete under which SAIC provides full-scale operations and maintenance for the Customs and Border Protection systems that assess security risk from travellers and cargo entering the United States.

The core capability. The Civilian segment's distinctive skill is running high-volume, citizen-facing transaction systems where failure is politically visible. A border risk-assessment system that goes down halts commerce at ports of entry. A benefits system that miscalculates eligibility produces congressional hearings. SAIC has been on the State Department's Vanguard programme since 2010 - fifteen years of running an agency's global enterprise IT - and holds branded assets like eligibilityNOW, an AI-driven claims and benefit-eligibility tool.

Why it exists separately. The contracting economics are materially different. Civilian agencies have moved further and faster toward firm-fixed-price and outcome-based contracting than defence customers have. CFO Prabu Natarajan said on the Q2 FY2027 call (August 31, 2026) that civilian customers are "further along than defense and intelligence customers" in adopting outcome-oriented contracting. Fixed price transfers execution risk to the contractor - and with it, the upside from doing the job for less than the price. That is why Civilian earns a materially higher adjusted operating margin than D&I, and why management repeatedly points to it as the proof-of-concept for where the whole company could go.

On the Q1 FY2027 call (June 1, 2026), Natarajan described Civilian margins running around 15%, up from a prior 12%-12.5% level, and said he expected "this business to operate consistently at 15% or so."

Competitive position. Here the competitive set widens to include the commercial IT giants. Accenture Federal Services, Deloitte, IBM and General Dynamics IT all bid civilian enterprise IT hard, and Accenture Federal is the firm that took RITS away from SAIC. SAIC wins where mission knowledge and agency incumbency matter (CBP risk assessment, State's global estate) and loses where the work is generic systems integration priced against commercial benchmarks.

Role in the group. The margin proof point and the strategic template. Roughly a quarter of revenue, but management's rhetoric treats it as roughly half the strategy.

Segment comparison

The operating structure sits underneath this

On November 13, 2025, SAIC consolidated five business groups into three, effective January 31, 2026: Army and Navy merged into the Army Navy Group (ANG) under Barbara Supplee; Air Force & Combatant Commands merged with Space and Intelligence into the Air Force, Space and Intelligence group (AFSI) under Vinnie DiFronzo; the Civilian Business Group continued under Srini Attili. Three executives - Josh Jackson (Army), David Ray (Space and Intelligence) and Lauren Knausenberger (Chief Innovation Officer) - departed. Then-interim CEO Jim Reagan said the changes were intended to "align our investments more closely with those opportunities" and to optimise for "speed, flexibility and efficiency."

The reorganisation is worth noting for what it says about cost structure: SAIC removed a layer of very senior, very expensive P&L leadership three weeks after a CEO change.

3. Products and Business Detail

3.1 The capability catalogue

SAIC organises its offerings into four stacks, per its corporate site.

Mission IT. The differentiated end. Includes CJADC2 (Combined Joint All-Domain Command and Control) work - stitching sensors, shooters and decision systems across services into a single picture; data and AI services; digital transformation; and quantum technologies. This is where the company wants to concentrate.

Enterprise IT. Service management, cloud, cybersecurity and digital workplace. SAIC's branded cloud platform here is CloudScend. This is the stack management has been deliberately shrinking: on the Q4 FY2026 call (March 16, 2026), CEO Jim Reagan indicated enterprise IT exposure would decline from roughly 17% of revenue in FY2025 to about 10% in FY2027.

Engineering Services. Digital engineering (model-based systems engineering, digital twins, simulation of a system before metal is cut) and systems integration and delivery. This is the oldest capability in the house and traces directly to the Beyster-era physics and analysis work.

Professional Services. Transformation and consulting - the advisory work that OCI rules make possible precisely because SAIC does not sell the platforms.

The 10-K frames the same portfolio around five strategic imperatives: Undersea Dominance, Border of the Future, Citizen Experience, All Domain Warfighting, and Next Generation Space. These are how the company chooses where to invest bid-and-proposal dollars.

3.2 Named software assets

SAIC has more named IP than most pure services firms, largely acquired.

Tenjin - launched January 2023, a low-code-to-full-code AI/ML development and orchestration platform powered by Dataiku. It handles model development, training, deployment, automation, data preparation and visualisation. Its purpose is to let an agency with limited data-science staff stand up production ML without building the toolchain itself.

Koverse Data Platform (KDP) - acquired 2021. A security-first data management and governance platform that enforces attribute-based access control (ABAC), delivering what SAIC markets as "Zero Trust for data." This matters more than it sounds. In classified environments, the hard problem is not storing data, it is proving that a given analyst may see a given record given their clearance, citizenship, need-to-know and the data's originator caveats. ABAC does that at the row and cell level. Koverse also supports data synchronisation across locations while preserving owner controls - useful in disconnected or bandwidth-constrained environments.

In November 2023 SAIC combined the two, adding natural language processing to Tenjin and bundling it with Koverse to create a text-, speech- and imagery-capable all-source analysis backbone.

RAG-R - a retrieval-augmented-generation framework for large language models, marketed for secure deployment with real-time knowledge updates and structured reasoning in defence settings. This is SAIC's answer to the obvious government question: how do you use an LLM when you cannot send data to a commercial API and you need the model to cite its sources?

eligibilityNOW - AI-driven claims processing and benefit-delivery tooling for civilian benefits agencies.

SilverEdge - acquired October 15, 2025 for a preliminary purchase price of $203 million net of $6 million cash acquired (reported elsewhere as ~$205M). SilverEdge brings a software platform and agentic AI products for the national security sector, plus classified AI-enabled network capability. On the Q3 FY2026 call (December 4, 2025), Reagan called it evidence of "our ability to invest in differentiated IP capable of solving customer problems," adding: "I am wildly enthusiastic about what SilverEdge is going to be able to do, not just as a standalone part of our business, but as we integrate it into the portfolio." On the Q4 FY2026 call he noted the classified AI-enabled network capability is "extensible beyond original customer base" - which is the acquisition thesis in one phrase. SilverEdge contributed $20 million of revenue in Q2 FY2027.

3.3 The delivery model, not a manufacturing process

SAIC has essentially no factories. Its "manufacturing" constraints are three.

Clearances. The rate at which the company can grow a classified programme is set by clearance adjudication throughput, not by recruiting. A cleared engineer is a scarce, poachable asset with a market price, and industry-wide wage inflation in the cleared market is a direct margin input.

Facilities. SCIF space is government-accredited, expensive and slow to build. Capacity here is a genuine bid discriminator.

Bid-and-proposal capital. A large federal recompete costs millions of dollars and months of senior staff time to bid. This is why "selectivity" is a real strategy and not a euphemism: on the Q1 FY2027 call, Reagan noted the pipeline had contracted roughly 25% year over year, "predominantly in enterprise IT, reflecting our selectivity in this part of the market." Management targets new-business win rates of at least 30% by bidding fewer, better-qualified opportunities, and reported submitting roughly $21 billion of proposals in the first three quarters of FY2026 with a target of over $30 billion of FY2027 submissions (Q3 FY2026 call).

Reported capital expenditure is small - roughly $35 million budgeted for FY2027, with about $25 million spent in the first half - and is described as flexible depending on customer production demands.

3.4 Geography

SAIC's geography is essentially the United States federal footprint: the Washington DC region (Reston headquarters), Huntsville (Army missile and space), Colorado Springs and Los Angeles (Space Force), Norfolk and San Diego (Navy), Ogden and Dayton (Air Force sustainment), plus deployed and overseas support at customer sites. International revenue is a residual within the "commercial, state/local and international" bucket that makes up roughly 2% of the total. Where international work exists it is generally in support of a U.S. customer - for example, a $16.8 million Navy contract to support Saudi C4ISR operations reported in March 2026 - or delivered through a global agency estate like the State Department's, which SAIC has served under Vanguard since 2010.

3.5 Milestones that changed the business

4. Customers

Who buys

Ninety-eight percent of SAIC's revenue is the U.S. government, as prime or subcontractor. Roughly 52% is the Department of War; roughly 46% is the Intelligence Community and other federal agencies. Within that, the named buying institutions are the Army, Navy, Air Force, Space Force, combatant commands, unnamed intelligence agencies (SAIC's press releases routinely say "a U.S. intelligence agency" because the customer's identity is classified), DHS and Customs and Border Protection, the Department of State, the Treasury, NASA, the GAO, the Army Corps of Engineers, and veterans' health organisations. State and local work is small; Orange County, California is one named account.

There is no single commercial customer of consequence. There is no consumer.

Who inside the customer decides, and how

Three separate people have to say yes, and they optimise for different things.

The contracting officer (KO) owns the legal award. Their institutional incentive is defensibility - an award that survives a GAO protest. This produces a strong bias toward documented, mechanical evaluation criteria and toward incumbents with clean past performance.

The programme manager and technical evaluation board score the technical proposal and past performance. They care whether the bidder actually understands the mission, and whether the named key personnel exist and are available.

The budget holder controls the money and, critically, the timing. Under a continuing resolution, funds arrive late and in increments, which is why funded backlog ($3.8 billion at Q2 FY2027) is a small fraction of total backlog ($22.1 billion): the government has committed to the scope but not yet obligated the cash.

Sales cycles are long. A major recompete typically runs 12-24 months from draft RFP to award, and can be extended by months more if a losing bidder protests to the GAO - as SAIC itself did on RITS, unsuccessfully.

Why they buy from SAIC

Four specific reasons, in rough order of weight.

Cleared people who already exist. For classified work, this is close to dispositive. Proposals are evaluated partly on whether the bidder can staff the programme on day one.

Incumbency and transition risk. Fifteen years on State's Vanguard programme means SAIC knows an estate that no successor could learn quickly. Agencies are genuinely afraid of transition, and that fear is worth a great deal to an incumbent. It is also why SAIC's non-enterprise-IT recompete win rate sits at 85-90%.

Vendor neutrality. Because of the 2013 OCI-driven split, SAIC does not sell competing hardware or platforms. When an agency wants advice about what to buy, that neutrality is a legal prerequisite.

Compliance infrastructure. A government-audited accounting system, CMMC-track cybersecurity posture, facility clearances, and a track record of surviving DCAA audits. This is table stakes that takes years and real money to establish - and it is the reason a well-capitalised commercial software company cannot simply decide to enter.

Switching costs

High for mission work, low for commodity work, and the gap between those two is the central fact of SAIC's current situation.

On a classified mission programme, switching means re-badging or replacing hundreds of cleared staff, re-accrediting systems, and accepting a real risk of capability gap. Agencies do it rarely.

On generic enterprise IT - running a network, staffing a service desk, reselling cloud - the work is specifiable, comparable and re-competeable, and there are many capable bidders. SAIC lost RITS to Accenture Federal on exactly this basis. It also chose not to defend some of this ground: it no-bid the low-margin portion of the Air Force Cloud One recompete, and later voluntarily dismissed its protest of the Cloud One architecture award that went to Leidos.

CFO Prabu Natarajan put the strategic distinction memorably at the TD Cowen conference:

"We're focused on being more selective... we want you to transform. We're not here just to run the network."

Concentration

Customer concentration is extreme by ordinary corporate standards - one buyer, the U.S. government, is essentially all of it - but that is the industry, not the company. The more meaningful concentration risks are programme concentration and agency concentration. A single programme roll-off can move the whole company: the RITS loss alone is a ~$200 million annual revenue headwind. On the Q1 FY2026 call, management noted that roughly 70% of Civilian revenue came from its top five agencies.

This concentration is best read as a reflection of the barrier to entry rather than a symptom of weak diversification. Every competitor has the same customer.

Contract structure and what it means for predictability

FY2026 contract mix, per the 10-K: cost-reimbursement 62%, time-and-materials 22%, firm-fixed-price 16%.

Cost-plus means SAIC bills allowable costs and earns a fee. Margin is structurally capped, cost overruns are largely passed through, and revenue tracks headcount and customer funding. Time-and-materials is similar in spirit - a billing rate per labour hour. Together, 84% of revenue is essentially a rate-times-hours business, which is why revenue is so sensitive to how fast the government pushes money onto contracts, and why the phrase "on-contract growth" appears in every earnings call.

Firm-fixed-price is different: SAIC quotes a price, and keeps the difference if it delivers for less. It is where operating leverage lives, and where execution failures bite (management disclosed a fixed-price space programme that produced a $3-5 million negative EBITDA impact in Q1 FY2026). Management said on the Q2 FY2027 call that fixed price is currently 15%-18% of sales but that roughly one-third of the pipeline is fixed-price, and that programme managers and contract teams are being trained for the shift.

Backlog provides the revenue bridge: $22.1 billion total at Q2 FY2027, of which $3.8 billion funded. Management's planning assumption, disclosed on the Q2 FY2026 call, is 70%-90% backlog conversion in a given year, plus on-contract growth, plus new business.

5. Competitive Landscape

The structure of the industry

Federal services is a fragmented oligopoly with three distinct tiers, and SAIC straddles two of them uncomfortably.

Tier 1 - the platform primes' services arms. Lockheed Martin, Northrop Grumman, RTX and General Dynamics (through GDIT) all sell services alongside hardware. They bring balance-sheet scale and platform adjacency, but they carry OCI baggage - they cannot advise on procurements involving their own products.

Tier 2 - the pure-play federal services and engineering firms. Leidos, CACI, Booz Allen Hamilton, Parsons, Peraton, ManTech, Amentum, Serco. This is SAIC's true peer set. Competition here is intense, largely price-and-past-performance driven, and the differentiation is real but narrow: each firm has agencies where it is entrenched and agencies where it is a perpetual challenger.

Tier 3 - the commercial IT and consulting firms coming down-market. Accenture (through Accenture Federal Services), Deloitte, IBM, AECOM. They bring commercial delivery methods, global scale and genuinely lower cost structures on standardisable work. They are the reason SAIC is exiting commoditised enterprise IT - Accenture Federal took RITS, a contract SAIC had held since 2021, and the GAO denied SAIC's protest on May 19, 2026, clearing Accenture to begin work on the successor programme (CASTLE-NET).

The 10-K names all of these: General Dynamics, Lockheed Martin, Northrop Grumman and RTX among diversified contractors; Booz Allen Hamilton, CACI International, Leidos Holdings, ManTech International, Parsons, Peraton and Serco Group among focused IT and engineering firms; and Accenture, AECOM, Amentum Holdings, Deloitte and IBM among commercial IT providers.

Competitor comparison

Market caps are approximate peer-size references only, sourced from stockanalysis.com as of early September 2026.

Where SAIC wins and where it loses

Wins. Recompetes on differentiated mission work - management has repeatedly cited 85%-90% win rates outside enterprise IT, and above 90% in Q2 FY2027. New business in mission areas, where CFO Natarajan said on the Q4 FY2026 call that non-enterprise-IT new-business win rates "approached 50%." Advisory work where OCI neutrality is required. Programmes where the customer's real constraint is cleared headcount.

Loses. Commoditised, price-driven enterprise IT against Accenture Federal, GDIT and Peraton. Anything where the customer is consolidating many contracts into a single mega-vehicle and then competing task orders on price - the State Department's Evolve consolidation is exactly this shape, replacing Vanguard 2.2.1 plus 22 other contracts. And, historically, it has lost on growth to Leidos and CACI, both of which have compounded faster since the 2013 split.

Barriers to entry - real, but asymmetric

The barriers are high and they are the reason this industry has such durable incumbents: facility and personnel clearances, government-compliant accounting, past-performance history, and the sheer capital cost of bidding. A new entrant cannot buy its way in at speed; it must acquire an incumbent.

But the barriers protect the industry, not any individual company's revenue base. Inside the moat, roughly ten firms with identical qualifications compete for the same work every five years, and a recompete loss transfers revenue instantly and completely. SAIC's RITS loss is the case in point: no technological disruption, no strategic error of the dramatic kind, just a competitor with a better price-technical tradeoff on a re-bid of work SAIC had been performing.

There is no moat narrative here that survives contact with the data at the company level. The correct description is: high industry barriers, low individual-contract stickiness on commoditised work, high individual-contract stickiness on classified mission work, and a company deliberately trying to move its mix from the first to the second.

Structural shifts underway

Three matter. Consolidation of contract vehicles - agencies are collapsing dozens of contracts into large multiple-award IDIQs (Evolve, COMET), which increases the number of task-order competitions and reduces the value of any single win. The commercial firms coming down-market with lower cost structures on standardisable work. And the shift toward outcome-based and fixed-price contracting, which rewards firms willing and able to take delivery risk and punishes pure labour arbitrage. Natarajan warned on the Q2 FY2026 call that pure labour-based service models face vulnerability and that the industry must "convert more of labor based into differentiated tech offerings."

6. Industry

What drives demand

Federal services demand is a function of appropriated budgets and, more immediately, of outlay velocity - how fast the Treasury actually pushes obligated money onto contracts. These are different things, and the second has been the binding constraint through FY2026.

The primary drivers are: national security threat perception (which sets the defence topline), the pace of legacy-system replacement in civilian agencies, the migration of government workloads to cloud, cybersecurity mandates, and increasingly the adoption of AI in government workflows. A secondary but powerful driver is administrative disruption: reorganisations, workforce reductions and procurement-policy changes at agencies, which paradoxically slow near-term contract awards even when they raise long-term demand for outsourced capability.

Size and trajectory

IT is the largest single federal contracting category, at over $100 billion of annual federal spending, rising to over $150 billion when state, local and education buyers are included. For scale, the federal government committed roughly $755 billion on contracts in FY2024. SAIC therefore competes for a slice of a market where its share of the addressable federal IT and technical-services pool is in the low-to-mid single digits - there is no share ceiling constraining growth, only competitive intensity.

The FY2027 defence budget request is on the order of $1.5 trillion per CSIS analysis, with outlays under those plans estimated to peak in FY2028. The topline is not the problem. The problem is enactment: Congress has repeatedly failed to pass full-year appropriations on time, and per the Committee for a Responsible Federal Budget, the House and Senate have passed competing continuing resolutions extending funding into December.

Where SAIC sits in the chain

SAIC sits between the agency and the technology. It buys nothing of consequence except labour and some resold software and cloud, and sells integration, engineering judgment and operations. It is upstream of nobody and downstream of the OEMs whose products it integrates - which is precisely the neutral position the 2013 split created. In the value chain it captures the integration margin, not the product margin, which is why its structural profitability is a fraction of a platform prime's.

Regulation is the industry

Federal services is shaped by regulation more than by technology. The Federal Acquisition Regulation (FAR) and its defence supplement govern contract types, allowable costs and competition requirements. Organizational conflict of interest rules define which companies may bid on what. DCAA audits test the legitimacy of indirect cost pools and can result in disallowed costs years after the fact. Personnel security clearance adjudication gates the workforce. CMMC cybersecurity certification is becoming a prerequisite for defence work. And the GAO bid-protest process means any award can be frozen for months - a mechanism SAIC has used (RITS, Cloud One) and had used against it.

Two live regulatory-adjacent dynamics matter. First, the current administration's procurement guidelines and government-efficiency initiatives have slowed RFP issuance and award decisions - management has flagged this on every call for six quarters. Second, the push toward fixed-price and outcome-based contracting shifts risk to contractors and, over time, should widen the gap between operators who can execute and those who cannot.

Cyclicality

Federal services is counter-cyclical to the economy and pro-cyclical to the budget cycle. In a recession, federal spending is stable or rises; SAIC's revenue barely notices. But it is highly sensitive to appropriations timing, which has its own cycle unrelated to GDP:

• A continuing resolution freezes spending at prior-year levels and blocks new-start programmes, deferring awards.
• A government shutdown stops work on unfunded contracts. SAIC quantified the Q3 FY2026 shutdown as roughly a one-point revenue headwind and, on the Q2 FY2026 call, said a month-long shutdown historically produced under 1% revenue impact with recovery within billing cycles and minimal cash impact.
• Administration transitions produce 12-24 months of procurement slowdown as political appointees arrive and review priorities. This is exactly what SAIC experienced through FY2026.

Tailwinds and headwinds

Tailwinds: a rising defence topline; the sheer accumulated technical debt in federal civilian systems; AI adoption creating genuinely new spend categories; cyber mandates; and the ongoing shift of classified workloads to accredited cloud, which requires integrators who can operate at that classification level.

Headwinds: chronic continuing-resolution governance that delays awards; a procurement workforce that has itself been reduced, slowing evaluations; commercial IT firms with lower cost structures attacking the commodity tier; consolidation of contracts into large IDIQs that intensifies price competition at the task-order level; and cleared-labour wage inflation that compresses margins in a cost-plus world.

7. Growth Triggers

Drawn from the six most recent earnings calls: Q1 FY2026 (June 2, 2025), Q2 FY2026 (September 4, 2025), Q3 FY2026 (December 4, 2025), Q4 FY2026 (March 16, 2026), Q1 FY2027 (June 1, 2026), Q2 FY2027 (August 31, 2026).

• Project ORBIT cost transformation reaching full run-rate. A three-year programme targeting approximately $150 million of annual run-rate savings, drawn from roughly 3,500 employee-generated ideas spanning procurement, recruiting, process simplification and automation including AI tooling. Roughly $100 million is earmarked for reinvestment in growth and capacity; roughly $50 million supports margin expansion. First flagged as a "$100 million-plus" indirect-cost opportunity on the Q3 FY2026 call (December 4, 2025), named and quantified on the Q1 FY2027 call (June 1, 2026), and expanded on at Q2 FY2027 (August 31, 2026). Repeated across three consecutive calls.

  "3.5 thousand ideas from across the company which our dedicated team is now analyzing and prioritizing for execution." - James C. Reagan, CEO, Q1 FY2027 call, June 1, 2026

• Margin path to mid-10% in FY2028 and approximately 11% by FY2030 as ORBIT matures (Q2 FY2027 call, August 31, 2026). Management framed FY2027 as "a year of commitment" and FY2028 as "a year of implementation" for the ORBIT rollout.

• Shift of the bid pipeline toward fixed-price work. Fixed price is currently 15%-18% of sales; approximately one-third of the pipeline is fixed price (Q2 FY2027 call, August 31, 2026). Management is training programme managers and contract teams for the transition and points to Civilian's ~15% operating margin as the proxy for what mix shift can deliver.

• Ramp of recently-won programmes that under-delivered in FY2026. Programmes won in FY2025-FY2026 that ramped slower than planned generated approximately $350 million in FY2026, are assumed to contribute approximately $500 million in FY2027, and carry a potential run-rate above $800 million (Q4 FY2026 call, March 16, 2026; reiterated Q1 FY2027 and Q2 FY2027). Named examples from the TD Cowen conference: Air Force TENCAP (~$1B), Navy ATSO (~$75M/year), Army OSINT (~$75M/year). Repeated across three calls.

• State Department Evolve vehicle expansion. Evolve is a $10 billion ceiling, seven-year IDIQ consolidating the incumbent Vanguard scope (approximately $250 million annually for SAIC) plus 22 other State contracts. SAIC was selected on four of five work streams, positioning it to compete for scope beyond what it holds today (Q1 FY2027 call, June 1, 2026).

  "a $10 billion ceiling over 7 years" - Prabu Natarajan, CFO, Q1 FY2027 call, June 1, 2026

• COMET IDIQ position with the Missile and Space Intelligence Center. SAIC won a position on the estimated $14 billion multiple-award IDIQ, disclosed subsequent to the close of Q2 FY2027 (announced with Q2 FY2027 results, August 31, 2026). This is a hunting licence rather than backlog - value accrues only as task orders are competed and won.

• DHS Customs and Border Protection recompete, ~$740 million over five years, awarded after the close of Q2 FY2027 (August 31, 2026). SAIC will provide full-scale operations and maintenance for the systems that assess security risk from travellers and cargo entering the country.

• Intelligence and space award momentum. Over $1.6 billion of intelligence and space awards booked in the first half of FY2027, described by management as well ahead of recent trends, including a ~$400 million five-year intelligence agency recompete (Q2 FY2027 call, August 31, 2026), following $970 million of seven-year space and intelligence recompetes disclosed at Q1 FY2027 (June 1, 2026).

• SilverEdge integration and cross-selling. The classified AI-enabled network capability acquired in October 2025 is described as extensible beyond the original customer base (Q4 FY2026 call, March 16, 2026), and contributed $20 million of revenue in Q2 FY2027.

  "I am wildly enthusiastic about what SilverEdge is going to be able to do, not just as a standalone part of our business, but as we integrate it into the portfolio." - Jim Reagan, Interim CEO, Q3 FY2026 call, December 4, 2025

• Enterprise IT run-off completing. Enterprise IT exposure declining from approximately 17% of revenue in FY2025 to an expected ~10% in FY2027 (Q4 FY2026 call, March 16, 2026). Management's argument is that once the drag is out of the base, reported growth converges to the growth of the differentiated core.

• Proposal volume expansion. Approximately $21 billion of proposals submitted in the first three quarters of FY2026, with a target of over $30 billion of submissions in FY2027 (Q3 FY2026 call, December 4, 2025).

• Strategic portfolio review with an update due at the December earnings call. CEO Reagan said management is identifying "the intersection of our strongest right to win and our greatest growth potential," including potential M&A, while cautioning against expecting "a sharp change in its identity or business focus" (Q2 FY2027 call, August 31, 2026).

• Tuck-in M&A capacity. Management has stated it is pursuing tuck-in acquisitions aligned with mission-critical areas, with net leverage at 3.0x-3.1x described as within the targeted range and trending downward (Q1 FY2027 and Q2 FY2027 calls).

Trigger summary

8. Key Risks

8.1 Recompete concentration: a single loss transfers hundreds of millions instantly

Mechanism. Roughly 84% of revenue is cost-plus or time-and-materials work performed under contracts that expire and are re-competed, typically on five-to-ten-year cycles. When a recompete is lost, the revenue does not decline - it stops, in full, on a known date, and transfers to the winner. There is no gradual erosion to manage around.

The RITS loss is the working example. SAIC held the Army Corps of Engineers' Revolutionary Information Technology Services task order supporting roughly 37,000 users; Accenture Federal won the successor CASTLE-NET; the GAO denied SAIC's protest on May 19, 2026. CEO Reagan quantified it on the Q1 FY2027 call:

"a $200 million headwind beginning in Q3" - James C. Reagan, CEO, Q1 FY2027 call, June 1, 2026

He noted it contributes roughly three points of organic growth pressure across Q3 and Q4 combined.

Calibration: high probability, moderate-to-severe magnitude, recurring. Management has stated a typical recompete run-rate of 10%-15% of revenue per year, with individual programmes in the 1%-3% revenue range coming up over any twelve-month window. Some losses are certain; the question is only which. The specific forward exposure management named on the Q4 FY2026 call is the State Department Vanguard programme, where SAIC has been the incumbent for fifteen years, with most of the revenue impact expected in FY2028. SAIC secured positions on four of five Evolve work streams, which mitigates but does not eliminate the risk - a position on an IDIQ is a right to bid, not revenue.

8.2 Appropriations timing: the company cannot control when it gets paid to grow

Mechanism. Because most work is rate-times-hours, revenue growth on existing contracts ("on-contract growth") requires the customer to obligate more funding onto the contract. Under a continuing resolution, agencies cannot start new programmes and are slow to increase funding on existing ones. Contracting officer turnover and new procurement review processes compound the delay. The result: SAIC can hold every contract, win new work, and still see revenue contract - which is precisely what happened through FY2026.

Then-CEO Toni Townes-Whitley described the mechanism plainly on the Q2 FY2026 call:

"We are seeing a more challenging environment than we had previously forecasted."

She attributed it to slower conversion of on-contract growth, increased programme disruptions and delays in new business awards, naming Army Transformation Initiative programmes (including S3I in Huntsville), Treasury's tCloud initiative and certain Space Force programmes. CFO Natarajan added on the same call: "We have to assume that things get worse before they get better."

Calibration: high probability, moderate drag, persistent. Management's own FY2027 planning assumption, stated on the Q2 FY2027 call, is that the government begins its next fiscal year under a continuing resolution and that no material improvement in the procurement environment occurs. That is a company assuming the headwind persists rather than one betting on relief.

8.3 The margin story depends on a mix shift that is only partly in management's control

Mechanism. Management's path to ~11% margins by FY2030 rests on three legs: ORBIT cost savings, running off low-margin enterprise IT, and shifting toward fixed-price work. The third leg depends on customers choosing to buy that way. Management itself acknowledged on the Q2 FY2027 call that civilian customers are "further along than defense and intelligence customers" in adopting outcome-oriented contracting - and Defense and Intelligence is roughly 77% of revenue. If defence customers stay on cost-plus, the mix shift stalls in the segment that matters most.

There is a second-order risk here: fixed price cuts both ways. SAIC disclosed a fixed-price space programme on the Q1 FY2026 call that produced a $3-5 million negative EBITDA impact. Taking more fixed-price work with programme managers who have spent careers on cost-plus is an execution risk the company is explicitly training for, which is itself an admission that the capability is not yet uniform.

Calibration: moderate probability, moderate magnitude, multi-year.

8.4 Management discontinuity at the top

Mechanism. Toni Townes-Whitley became CEO in October 2023 and departed October 23, 2025 after two years, during which she pursued a portfolio pivot toward mission and enterprise IT. James C. Reagan, a director since January 2023, was named interim CEO the same day and made permanent February 17, 2026. Three weeks after the interim appointment, the company consolidated five business groups into three and three senior executives departed.

Per the FY2026 proxy statement, Reagan is 67, has been a CPA licensed in Virginia since 1982, and previously served as Executive Vice President and Chief Financial Officer of Leidos Holdings from July 2015 to June 2021, as CFO of Vencore, Inc., as CFO of PAE, Inc., as CFO of Aspect Communications until 2005, and in senior roles at MCI Telecommunications. He holds a BBA in accounting and finance from the College of William & Mary and an MBA from Loyola University Maryland. Board Chair Donna S. Morea, 71, is the independent non-executive Chair and is described in the proxy as a former President of CGI Group.

The specific risk is twofold. First, a CFO-turned-CEO is temperamentally well-suited to a cost-and-margin programme like ORBIT and less obviously suited to reversing a multi-year growth deficit - and growth, not margin, is the company's actual problem. Second, at 67 with a stated portfolio review underway, the tenure horizon is uncertain, which raises the odds of another strategic reset within a few years.

Calibration: moderate probability, moderate magnitude. Note the mitigating evidence in Section 11: Reagan and the CFO both bought stock in the open market in December 2025.

8.5 Structural exposure to commercial IT firms in the commodity tier

Mechanism. Accenture Federal, Deloitte, IBM and GDIT bid standardisable enterprise IT with lower cost structures and commercial delivery methods. SAIC's response has been to exit rather than to compete: no-bidding the low-margin portion of Air Force Cloud One, dismissing its protest of the Cloud One architecture award that went to Leidos, and shrinking enterprise IT from ~17% of revenue in FY2025 to a targeted ~10% in FY2027. The Q4 FY2026 call described this as "addition by subtraction," with the Cloud One decision alone creating a ~$200 million headwind and a $60 million year-over-year Q4 reduction.

The risk is that the boundary keeps moving. If work that is currently "differentiated mission" becomes standardisable through better tooling and AI, the commodity tier expands and SAIC's exit strategy becomes a treadmill.

Calibration: moderate probability, high magnitude over five-plus years. Natarajan effectively conceded the mechanism at the TD Cowen conference by describing the exited work as earning 6%-7% margins against 12%-14% on differentiated offerings - the price of not competing there is real, and the assumption that the 12%-14% work stays protected is an assumption.

8.6 Bid-and-proposal selectivity is shrinking the funnel before it grows the wins

Mechanism. On the Q1 FY2027 call, Reagan noted the pipeline had contracted approximately 25% year over year, predominantly in enterprise IT, "reflecting our selectivity in this part of the market." Selectivity raises win rates, which management has demonstrated. But a smaller pipeline with a higher win rate produces the same or less absolute new business unless win rates rise faster than the funnel shrinks. Q2 FY2027 book-to-bill was 0.6 for the quarter and 0.8 on a trailing twelve-month basis - below the 1.0 replacement level, and well below the 1.5 quarterly and 1.4 year-to-date figures posted a year earlier in Q2 FY2026.

Management expects to finish FY2027 near 1.0x book-to-bill. That is a forecast, not an outcome.

Calibration: high probability of continued lumpy bookings, moderate magnitude. Federal bookings are genuinely lumpy - Q2 FY2026 printed 1.5 and Q4 FY2026 printed 0.3 - so a single quarter proves little. A trailing-twelve-month figure sitting at 0.8 for consecutive quarters would be a different matter.

8.7 Cleared labour cost inflation in a cost-plus world

Mechanism. On cost-plus contracts, SAIC bills allowable costs plus fee. Rising cleared-engineer wages therefore pass through to the customer on direct labour - but the fee is typically negotiated as a percentage or a fixed amount, and the indirect cost pools (recruiting, security processing, facilities) are subject to competitive rate pressure. Meanwhile on fixed-price work, wage inflation comes straight out of margin. As the company deliberately shifts toward fixed price, its exposure to cleared-labour inflation rises.

Calibration: high probability, low-to-moderate magnitude. This is a persistent drag rather than an event.

9. Walk The Talk

The six calls used:

1. Q1 FY2026 - June 2, 2025 (Toni Townes-Whitley, CEO; Prabu Natarajan, CFO)
2. Q2 FY2026 - September 4, 2025 (Townes-Whitley; Natarajan)
3. Q3 FY2026 - December 4, 2025 (Jim Reagan, Interim CEO; Natarajan)
4. Q4 FY2026 - March 16, 2026 (Reagan; Natarajan)
5. Q1 FY2027 - June 1, 2026 (Reagan; Natarajan)
6. Q2 FY2027 - August 31, 2026 (Reagan; Natarajan)

The most recent is three days old. This is a complete six-call record spanning a CEO change, which makes it unusually informative: it lets you separate one management team's credibility from another's.

The Townes-Whitley period: a guide that did not survive one quarter

On June 2, 2025, management reaffirmed FY2026 guidance of $7.6-$7.75 billion of revenue, implying roughly 2.5% organic growth at the midpoint, with adjusted EBITDA margin of 9.4%-9.6%, adjusted diluted EPS of $9.10-$9.30, and free cash flow of $510-$530 million. Townes-Whitley described a portfolio pivot toward mission and enterprise IT aligned with administration priorities, targeted a 1.2 trailing-twelve-month book-to-bill in coming quarters, and estimated the revenue impact of government efficiency initiatives at less than 1%.

Three months later, on September 4, 2025, the revenue guide was cut to $7.25-$7.325 billion - from roughly 2.5% organic growth to a 2%-3% organic contraction. That is a swing of roughly five points of growth in a single quarter, on a business with a $23 billion backlog. Adjusted EBITDA guidance came down from $715-$735 million to $680-$690 million.

"We are seeing a more challenging environment than we had previously forecasted." - Toni Townes-Whitley, CEO, Q2 FY2026 call, September 4, 2025

"We have to assume that things get worse before they get better." - Prabu Natarajan, CFO, same call

The explanation offered - slower on-contract growth conversion, programme disruptions, delayed awards - was almost certainly accurate. But the size of the miss against a guide reaffirmed twelve weeks earlier is the credibility problem, and the "less than 1%" efficiency-initiative estimate from June did not survive contact with September. It is worth noting what was raised in the same breath: adjusted EPS guidance went up from $9.10-$9.30 to $9.40-$9.60 and free cash flow went up from $510-$530 million to over $550 million, even as revenue was cut. Margin and cash held while the top line broke. That pattern repeats for the next four quarters and is the single most durable fact about this management team.

Also on that September call, management gave a preliminary FY2027 framing of 0%-3% revenue growth with 9.5%-9.7% adjusted EBITDA margin. Hold that thought.

Townes-Whitley departed on October 23, 2025, seven weeks later.

The Reagan period: a different promise, consistently kept

On December 4, 2025, with Reagan as interim CEO, the company raised FY2026 adjusted EPS guidance by $0.40, attributed to "strong program performance year to date" and a lower effective tax rate assumed at roughly 10%. Reagan disclosed the SilverEdge acquisition (closed October 15, 2025) and, critically, put a number on the cost opportunity for the first time: "over $100 million in annual spend that we're actively working to reinvest into higher ROI areas," with the promise of "margins approaching 10% in the near term, with additional potential upside in FY27." He also laid out three priorities - sharpening execution to create investment capacity, deploying financial resources more efficiently, and prioritising yield and bid quality in business development. He stated an intention to repurchase approximately $500 million of shares in each of FY2026 and FY2027.

On March 16, 2026, FY2026 closed with revenue of $7.26 billion - inside the September guide of $7.25-$7.325 billion - with a 9.7% adjusted EBITDA margin and free cash flow of $577 million against a guide of "greater than $550 million," a beat management characterised as exceeding the guide by 10%. Adjusted diluted EPS came in at $10.75 against a September guide of $9.40-$9.60 and a December guide of roughly $9.80-$10.00. That is a very large EPS beat, materially helped by an effective tax rate of roughly 14% versus a normalised ~23%.

Two things stand out from that call. First, the promise on margin was kept: December's "margins approaching 10%" became a delivered 9.7% for the year and 10.3% in Q4. Second, the FY2027 guide was cut hard relative to September's preliminary framing: from 0%-3% growth to a 2%-4% organic contraction, with revenue of $7.0-$7.2 billion, adjusted EBITDA margin of 9.9%-10.1% and adjusted EPS of $9.50-$9.70. Management named the causes - $400 million of recompete headwinds spread across all four quarters, the Cloud One no-bid, and slow-ramping wins - and framed the exits as deliberate:

Reagan described the strategy as "addition by subtraction," deliberately no-bidding low-margin Cloud One work. - Q4 FY2026 call, March 16, 2026

The intellectually honest reading is that Reagan reset the bar low on his first full guide as permanent CEO. Whether that is prudence or sandbagging is answered by what happened next.

FY2027: two consecutive raises

On June 1, 2026, Q1 FY2027 delivered adjusted EBITDA margin of 11.6% - described by Reagan as "record margin" - and adjusted diluted EPS of $3.23 against $1.92 in the prior-year quarter. Adjusted EPS guidance was raised roughly 4% to $9.90-$10.10 and adjusted EBITDA guidance to $720-$730 million, while revenue guidance was held. Book-to-bill was 1.1 for the quarter and 1.0 trailing twelve months. Reagan also delivered the bad news cleanly: the RITS recompete was lost, the protest denied, and a $200 million headwind would begin in Q3. Project ORBIT was named and the 3,500-idea crowdsourcing effort described.

On August 31, 2026, Q2 FY2027 produced 5.3% organic growth - against a full-year guide that had assumed 4%-2% contraction three months earlier - with on-contract growth of 9%, "well ahead of plan." Guidance was raised again, substantially: revenue to $7.2-$7.3 billion (organic growth of -2% to 0%, versus -4% to -2% prior), adjusted EBITDA to $750-$755 million, and adjusted diluted EPS to $10.65-$10.75. Recompete win rate exceeded 90%. Free cash flow guidance held at greater than $600 million.

"I am proud of our team's performance this quarter, delivering solid organic growth and double-digit margins as we continue to execute with discipline." - Jim Reagan, CEO, August 31, 2026

Set against the March guide of $9.50-$9.70, the August guide of $10.65-$10.75 represents an increase of roughly 11% in five months, on a business whose revenue is still expected to be flat-to-down for the year.

Promise versus outcome

Assessment

This is a management team that does what it says on cost, margin and cash, and that has systematically failed to predict its own revenue.

Across all six calls without exception, every margin, EPS and free-cash-flow commitment was met or exceeded, often substantially. Across the same six calls, revenue guidance was cut twice by large amounts (September 2025 and March 2026) and the drivers of the cuts - award timing, on-contract growth conversion, recompete outcomes - were the same drivers each time. The company appears to have genuine control over its cost structure and essentially no forecasting control over its top line.

Two caveats keep this from being a clean endorsement. First, the FY2026 EPS beat was materially assisted by an effective tax rate of roughly 14% against a normalised ~23% - that is a real benefit to shareholders but it is not operational execution, and the FY2027 guide correctly normalises it. Second, the March 2026 FY2027 guide was set low enough that two raises still leave revenue guidance at flat-to-down for the year. A cynic would call that expectation management; a fair-minded reader would note that Q2's 5.3% organic growth and 9% on-contract growth are real outcomes, not accounting, and that the RITS loss was disclosed with specificity rather than buried.

The honest verdict: credible on what they control, unreliable on what they forecast. If you are underwriting margin expansion and free cash flow, this team has earned the benefit of the doubt across six consecutive quarters. If you are underwriting revenue growth, six quarters of evidence says wait for it to appear in results before believing the guide - and note that the two raises in FY2027 came from the same team that had just cut the FY2027 bar in March.

10. Shareholder Friendliness Index

Dividends. SAIC has paid a quarterly cash dividend of $0.37 per share without interruption and without change across FY2024, FY2025 and FY2026, and it remains $0.37 for the dividends declared payable April 24, 2026 and October 23, 2026 - an annualised $1.48 per share, flat for at least three full fiscal years. Cash dividends paid were $79 million in FY2024, $75 million in FY2025 and $70 million in FY2026; the declining total against a flat per-share rate is simply the shrinking share count. The company states it intends to continue paying quarterly dividends subject to board discretion. The dividend is a token, deliberately: management's stated capital-return priority is repurchase, not distribution, and there is no evidence of any intention to grow the per-share rate.

Buybacks and dilution. The board authorised repurchases of up to $1.2 billion effective December 16, 2024, with no expiration date, replacing the prior programme. Actual execution over the last three fiscal years: $357 million (~3.3 million shares) in FY2024, $527 million (~4.2 million shares) in FY2025, and $422 million (~4.0 million shares) in FY2026 - roughly $1.3 billion of repurchase over three years, all disclosed in the respective fourth-quarter earnings releases. Fiscal 2027 is continuing at pace: $175 million of repurchases in Q1 FY2027 and $90 million in Q2 FY2027 per the quarterly releases, with a full-year target of approximately $400 million stated by CEO Reagan on the June 1, 2026 call. That target is below the ~$500 million per year Reagan indicated on the December 4, 2025 call, so the buyback is being scaled to cash generation rather than to a promise. (MoatMap's disclosure feed shows two US$110 million and US$98 million repurchase-related rows filed August 31, 2026, which reflect the quarterly report's different measurement basis rather than a second programme; the $90 million plan-repurchase figure in the earnings release is the cleaner number for the quarter.)

The dilution answer is unambiguous. Weighted-average diluted shares were 53.7 million in FY2024 and 42.8 million in Q2 FY2027 - a reduction of roughly 20% in under three years, comfortably outrunning equity-compensation issuance. The share count is genuinely shrinking, not merely offsetting grants. Net leverage sits at 3.0x-3.1x, described by management as within the targeted range and trending down, so the buyback is being funded from free cash flow rather than from balance-sheet expansion.

Verdict: Returns Capital. A ~20% reduction in diluted share count over three fiscal years, funded from free cash flow while deleveraging, is the clearest evidence in the file - though shareholders should note the dividend has been frozen at $0.37 for three years and the repurchase pace for FY2027 was quietly reset below the figure management named in December 2025.

11. Insider Activities

Source: SEC Form 4 filings for Science Applications International Corp (CIK 0001571123), cross-referenced against the MoatMap disclosure database. A disclosure note: the MoatMap block supplied for this report captures three transactions in the trailing twelve months (two DiFronzo gift legs and one McCarthy tax-withholding disposition) and does not capture three open-market purchases that appear in Form 4 filings within the same window - the December 16, 2025 buys by the interim CEO and the CFO, and an April 9, 2026 buy by the EVP of the Civilian business. Those are included below and cited to their filings, because they are the only transactions in the period that carry any signal at all.

Recent transactions (most recent first)

Buys - reading the signal

There were three open-market purchases in the trailing twelve months, and two of them happened on the same day.

On December 16, 2025, both the interim CEO and the sitting CFO bought SAIC stock in the open market. James Reagan purchased 1,000 shares at $100.17; Prabu Natarajan purchased 2,000 shares at approximately $100.49, increasing his direct holding by about 40%. This is a very bullish signal. Three things make it stronger than the dollar amounts suggest.

First, it is cluster buying. Two of the three most senior executives in the company transacted on the same day, twelve days after the December 4 earnings call at which Reagan had first quantified the "over $100 million" cost-reduction opportunity that later became Project ORBIT. They bought after telling the market the plan and before there was any evidence it would work.

Second, the context was maximum uncertainty. Reagan had been interim CEO for eight weeks. The permanent appointment would not come until February 17, 2026. The company had just cut FY2026 revenue guidance sharply in September and would cut FY2027 guidance again in March. Buying into that is not a routine gesture.

Third, Natarajan's purchase was large relative to his own position. A 40% increase in a CFO's direct holding is not a token. In salary terms, roughly $201,000 is a meaningful fraction of a year's base pay for a federal-services CFO - months, not weeks.

The third purchase, Attili's 100 shares at $94.96 in April 2026, is genuinely token - roughly $9,500 against a 17,656-share holding - and reads as a symbolic gesture rather than a conviction trade. It is worth noting only because it fits a pattern: this management team's small buys are the kind executives make when they want to be seen owning stock, and the December buys are the kind they make when they think it is cheap.

Sells - working out the why

There were no discretionary open-market sales by SAIC insiders in the trailing twelve months that this research could locate.

Every disposition in the period falls into one of two non-signalling categories:

• Tax withholding on vested equity (transaction code F). Kathleen McCarthy's 1,233 shares at $114.35 in June 2026 and Prabu Natarajan's 1,685 and 837 shares in April 2026 are all shares delivered to the company to satisfy tax liability on vesting restricted stock, not shares sold into the market. This is mechanical, occurs on the vesting date, and carries no information about the executive's view. Reason: disclosed in the filings themselves as payment of tax liability by delivering securities.
• Bona fide gifts. Vincent DiFronzo's two 2,682-share legs on July 2, 2026 are recorded at $0.00 per share and disclosed as bona fide gifts - one from an indirectly held trust, one from directly held shares. This is estate or charitable planning, not a market exit. Reason: disclosed as a gift in the Form 4.

No 10b5-1 plan sales, no block trades, no sponsor exits, no diversification programmes appear in the window.

Net assessment

Insiders at SAIC were net buyers over the last twelve months, in a period when the business was contracting and the leadership was in transition. Three open-market purchases, zero discretionary sales, and the only two dispositions of consequence being mechanical tax withholding and a family gift.

The activity is concentrated, but concentrated in the right two people. The CEO and CFO buying on the same day, eight weeks into an interim tenure and two months before the permanent appointment, is the most informative single data point in this section. It is also worth noting what changed: Reagan had been on the board since January 2023 without, on the available record, buying in the open market until he took the operating job. Natarajan increased his direct stake by 40% in one transaction.

The counterweight is size. In absolute dollars this is roughly $310,000 of buying across three transactions by executives whose annual compensation runs into the millions and who receive large annual equity grants (Reagan's April 2026 award alone was 31,543 shares). These are not career-defining bets. They are signals, and they should be read as signals rather than as capital allocation.

Plain-language read: bullish signal, with the caveat that the amounts are modest. The absence of any discretionary selling across a year of guidance cuts, a CEO departure and a major recompete loss is arguably as informative as the buying itself. Nobody at this company took the opportunity to get out.

12. Scenarios

Bull case

Project ORBIT turns out to be what Reagan claims it is - a structural re-engineering rather than another cost-cutting round. The 3,500 crowdsourced ideas convert into genuinely re-designed procurement, hiring and back-office processes, and the roughly $150 million of run-rate savings arrives on schedule. Critically, the two-thirds earmarked for reinvestment goes where management said it would: into bid-and-proposal capacity on differentiated mission work, into capacity on existing contracts, and into price competitiveness on the bids SAIC actually wants to win. Margins reach the mid-10s in FY2028 and approach 11% by FY2030, and because the share count keeps shrinking, earnings per share compounds faster than the business does.

Meanwhile the top line stops being the problem. The FY2027 recompete headwinds - RITS gone in Q3, Cloud One already out of the base, enterprise IT down to roughly a tenth of revenue - finish washing out of the comparison, and what remains is a portfolio of intelligence, space, defence mission and civilian work growing at the rate management says its core grows. The slow-ramping cohort from FY2025-FY2026 - TENCAP, ATSO, Army OSINT - finally reaches its potential run-rate above $800 million rather than the $500 million assumed. On-contract growth, which printed 9% in Q2 FY2027 against a 2%-3% full-year plan, turns out to reflect a genuinely improved outlay environment rather than a single good quarter.

The vehicle positions convert. COMET's $14 billion ceiling and Evolve's $10 billion ceiling are hunting licences today; in the bull case SAIC wins a disproportionate share of the task orders on both, because on COMET it brings the missile-intelligence heritage that MSIC actually needs, and on Evolve it brings fifteen years of knowing the State Department's global estate. The Vanguard recompete exposure that management flagged for FY2028 resolves in SAIC's favour or is more than offset by Evolve scope it does not hold today.

And the fixed-price shift works. Civilian's roughly 15% operating margin proves to be a template rather than an exception. As the pipeline's one-third fixed-price weighting flows into the revenue base, SAIC starts capturing the upside of executing efficiently instead of passing every saving back to the customer. The company that emerges is smaller in headcount, higher in margin, and finally growing - and by then it has retired another chunk of its share count at prices that look cheap in hindsight.

Base case

Nothing dramatic happens, which is roughly what management has guided for.

FY2027 lands where the August guide points: revenue flat to slightly down, adjusted EBITDA margin in the low-to-mid 10s, free cash flow above $600 million, and adjusted EPS materially above where the year started. The RITS roll-off hits Q3 and Q4 as advertised, so the second half looks worse than the first, and management has already told the market to expect high-9% margins in the back half because of targeted investment. Book-to-bill finishes near 1.0 - enough to hold the backlog, not enough to inflect it.

Project ORBIT delivers something, but less cleanly than the plan. Some of the savings prove to be normal cost discipline that would have happened anyway; some of the reinvestment goes into holding price on recompetes rather than into new capability. Margins improve toward the mid-10s by FY2028 because cost programmes in labour businesses usually do work, at least for a while. The 11% by FY2030 target stays a target.

The government contracting environment stays roughly as it is: continuing resolutions at the start of each fiscal year, slow award decisions, occasional shutdown noise, an adequate but not generous flow of money onto contracts. SAIC keeps winning the recompetes that matter - the 85%-90% win rate holds - and keeps losing the commodity ones it has decided not to fight for. Enterprise IT completes its run-off. The strategic portfolio review announced for the December 2026 call produces a modest divestiture and a tuck-in acquisition or two rather than a transformation, consistent with Reagan's own guidance that there will be no "sharp change in its identity or business focus."

Revenue growth turns modestly positive in FY2028 as the headwinds clear, and the buyback keeps shrinking the share count at something like the recent pace. The story remains "cost and cash, not growth," and the market keeps pricing it that way.

Bear case

The recompete arithmetic keeps going the wrong way. Vanguard - the State Department programme SAIC has held since 2010, flagged by management as the primary FY2027 recompete exposure with the revenue impact landing in FY2028 - goes to a competitor, or the Evolve consolidation reallocates scope such that SAIC's four work-stream positions produce far less than the incumbent base they replace. RITS was $200 million; Vanguard is comparable in scale and has fifteen years of institutional weight behind it. Two large incumbencies lost in three years, in a business where a loss transfers revenue in full on a known date, means the top line does not stabilise in FY2028 either. The 85%-90% recompete win rate is a statistic about many small contracts; it says nothing about whether the four or five programmes that actually matter renew.

The margin story then collides with the revenue story. ORBIT's savings are real but they were always going to be spent - two-thirds by design - and if the revenue base is shrinking, the fixed indirect cost pools that ORBIT is attacking have to be cut faster than planned just to stand still. Management's stated intent to reinvest $100 million into growth becomes an intent to protect margin instead. Mid-10s in FY2028 slips. The 11% target quietly disappears from the deck.

The fixed-price pivot goes badly. Roughly a third of the pipeline is fixed price and management is training programme managers for it, which is another way of saying most of them have never done it. A federal services company that has spent a decade on cost-plus takes on a handful of large firm-fixed-price programmes and discovers, as SAIC already did on one space programme in FY2026, that fixed price transfers real risk. A single meaningful overrun on a large fixed-price award wipes out several quarters of ORBIT savings, and does it publicly.

And the boundary keeps moving. The commodity tier that SAIC is exiting - the tier where Accenture Federal, GDIT and Peraton win on price - expands, as AI-assisted delivery makes more of what SAIC currently calls "differentiated mission work" specifiable and comparable. The company spends five years retreating up the value chain and finds the value chain retreating faster. Meanwhile the pipeline has already contracted 25% year over year in the name of selectivity, and a smaller funnel with a good win rate turns out to produce a smaller company.

Underneath all of it, appropriations dysfunction persists. Continuing resolutions become the permanent state, new-start programmes stay frozen, and on-contract growth - the mechanism by which a rate-times-hours business grows without winning anything new - stays structurally suppressed. SAIC ends up with excellent cost discipline, a shrinking share count, and a business that is slowly getting smaller. That is not a catastrophe. It is a value trap.