BigBear.ai

BigBear.ai is a specialized defense and national security AI company providing mission-ready predictive analytics, computer vision, and decision intelligence systems for the U.S. government (DoD, intelligence, Customs & Border Protection); products include AI-powered logistics optimization, trade & travel analytics, and the AskSage generative AI platform (acquired Q4 2025).

BigBear.ai provides mission-ready decision intelligence, computer vision, biometrics, and secure generative AI for U.S. defense, intelligence, border, travel, and trade workflows. The 2025 Ask Sage acquisition expanded the company into a secure multi-modal generative AI platform for government and enterprise use.

BigBear.ai builds mission-ready AI for defense and security workflows. In Q2 2026 it reported revenue of $36.7M (+13% YoY), with growth driven by Ask Sage’s GenAI platforms and products, and said it closed 20+ new contract wins. The company also highlighted total available cash and investments of $409.8M and affirmed full-year 2026 revenue guidance of $135M-$165M.

• monetization model: Software licenses/subscriptions plus professional services (engineering, cyber, analytics); multi-year government contracts. (Source: BBAI FY2023 10-K, Item 1, acc. 0001865782-24-000016, sec.gov )
• revenue mix: FY2023 revenue ~$155.2M: software licenses ~43%, services ~57%; U.S. federal government ~73%, commercial ~27%. (Source: same 10-K, revenue notes)
• who pays: U.S. federal agencies (DoD, DHS, IC), allied governments; commercial supply-chain/digital-identity enterprises. (Source: same 10-K, Item 1)
• margin_value_capture: FY2023 gross margin ~24.5%; operating margin negative; software licenses carry higher margin, but services dominate mix. (Source: same 10-K, MD&A)

• customer classes: Government - Program Buyers, Space & ISR Program Buyers, Defense C-UAS Integrators; plus commercial supply-chain/digital-identity buyers (not in taxonomy). (Source: BBAI FY2023 10-K, Item 1, acc. 0001865782-24-000016, sec.gov )

BigBear.ai operates as a decision-intelligence software and solutions provider. After the February 2024 acquisition of Pangiam, the company reports a combination of:

• Analytics / decision intelligence — AI/machine-learning-enabled data analytics, predictive modeling, simulation, and operational workflows, principally for U.S. defense, intelligence, and federal civilian customers.
• Vision AI / identity and access — computer-vision and biometric/identity-related capabilities used in security, travel, and critical-infrastructure screening environments, obtained through Pangiam.

This is a segment structure that must be verified in the latest 10-K segment footnote. The exact percentage split between Analytics and Vision AI is:

From the company’s SEC profile and investor materials, the monetization model is best described as:

1. U.S. government contract vehicles — IDIQ ceilings, task orders, and definitive contracts with defense and civilian agencies. Examples include the Army Global Force Information Management program; however, the contract ceiling, period of performance, and awarded-to-ceiling gap are needs primary source validation.
2. Subcontractor and prime contractor arrangements — BigBear.ai acts as prime on certain government vehicles and as subcontractor on others. This affects revenue quality: subcontractor revenue may carry lower margins.
3. Software and analytics embedded into customer workflows — The Analytics segment embeds ML models, data pipelines, and simulation tools into agency operations. This can create recurring or repeatable revenue when customer workflows become dependent on the tooling, but the company’s actual recurring revenue percentage is needs primary source validation.
4. Vision AI / biometric screening — Pangiam’s airport/security screening technology can be sold as product-enabled services or software to commercial and government customers. Unit economics may differ from Analytics; exact pricing model must be sourced from the 10-K.

The customer base is predominantly U.S. federal government. Buyer archetypes include:

BigBear.ai is a defense- and intelligence-focused artificial-intelligence, machine-learning, and decision-intelligence company. The company was formed through the combination of several government-technology businesses—most notably NuWave Solutions, BigBear, and PCI Strategic Management—before going public via a SPAC merger with GigCapital4 in December 2021. The company trades as BBAI.

Because the live filings could not be retrieved, the following business-model items are qualitative only.

needs primary source validation

The company has historically described itself as operating in one reportable segment or as having two revenue streams: analytics/software and professional services. A government-vs-commercial split is often the more analytically useful breakdown, but the exact percentage split for any recent period is not certifiable here.

Source to pull:

• BigBear.ai Holdings, Inc. 10-K filings via SEC EDGAR company index:

sec.gov

A worked calculation appears later. Until primary financials are retrieved, no unit-economic figure should be treated as measured.

needs primary source validation

BigBear.ai has historically been a services-heavy government contractor, which tends to produce gross margins in the low-to-mid 20s to mid 30s for blended services plus software, depending on contract mix. However, that is a memory-based industry prior, not a certified number. Operating margins have historically been negative because of public-company costs, stock-based compensation, amortization, and transaction costs.

Source to pull:

• BigBear.ai Holdings, Inc. Form 10-K for the latest fiscal year and Form 10-Q for the latest quarter:

ir.bigbear.ai

BigBear.ai monetizes through:

• U.S. government prime contracts — direct awards from defense, intelligence, and federal civilian agencies.
• Subcontracts — work performed under larger prime contractors such as systems integrators.
• Contract types — likely fixed-price, cost-plus-fee, and time-and-materials arrangements; exact mix needs primary source validation.
• Software licenses / subscriptions — AI/ML platform access and associated professional services, including data science, model development, and integration.
• Follow-on work — once embedded in mission-critical workflows, the company seeks to expand the contract scope.

Who pays: U.S. federal agencies and the prime contractors that serve them; to a lesser degree, commercial enterprises in logistics, manufacturing, and supply-chain analytics.

I cannot verify named customers from filings in this environment. Any specific agency or prime-contractor names circulating in press or investor decks must be re-verified against the 10-K and 10-Q. I will not list customer names as if they were primary-sourced.

This section must eventually be rebuilt from the 10-K customer disclosures. I will not invent a customer list.

Evidence-backed qualitative classes, pending primary validation:

• U.S. Department of Defense components — Army, Navy, Air Force, joint commands, defense agencies.
• Intelligence community — classified customers; names are generally not disclosed.
• Federal civilian agencies — logistics, homeland security, and healthcare-adjacent use cases.
• Commercial enterprises — supply chain, manufacturing, distribution, and logistics analytics.

• Program executive officers and program managers.
• Contracting officers.
• Chief data officers and chief technology officers.
• Intelligence analysts and operations planners.
• Logistics and sustainment commanders.

These archetypes follow from the nature of the product. They are not sourced to specific filing language here.

BigBear.ai Holdings, Inc. publicly files with the SEC under CIK 0001836981. Its periodic reports disclose financial information on reportable segments. The company’s business model is organized around AI and machine-learning-enabled decision-intelligence work, primarily for U.S. government customers, alongside services revenue and, following the Pangiam acquisition, computer-vision and identity-verification products.

The exact segment names and revenue percentages should be read directly from BigBear.ai’s latest Form 10-K and subsequent Form 10-Qs. Those documents are indexed at the SEC EDGAR filing index:

• BigBear.ai SEC EDGAR filing index: sec.gov

BigBear.ai’s monetization model, as described in its public investor materials and SEC disclosures, includes:

1. Software and AI platform revenue — licenses or subscriptions for decision-intelligence software that ingests, fuses, and analyzes data.
2. Professional/engineering services — execution of government contracts, including staff augmentation, system integration, and custom AI/ML development.
3. Contract-based government revenue — federal procurement vehicles such as cost-plus-fixed-fee, time-and-materials, and fixed-price contracts.
4. Post-Pangiam product additions — computer vision, biometrics, and identity-verification products for airport, border, and travel-security use cases.

The exact percentage split between software subscription revenue and services revenue is needs primary source validation. The SEC filing index should be used to retrieve the company’s management discussion and analysis, which typically describes revenue recognition policies and contract types.

BigBear.ai’s customer classes, based on its public disclosures, are:

1. U.S. federal defense agencies — Army, Navy, Air Force, and Office of the Secretary of Defense-related programs.
2. U.S. intelligence community — classified customers; names may be redacted or aggregated.
3. U.S. civilian federal agencies — homeland security, border, transportation, and law-enforcement adjacent missions.
4. Commercial enterprises — manufacturing, logistics, retail, and infrastructure operators.
5. Travel and border-security operators — post-Pangiam, including airports and border agencies.

Specific customer names must be pulled from BigBear.ai’s filings. If the filing redacts customer names for security reasons, the report should state that limitation rather than invent a name.

BigBear.ai Holdings, Inc. sells AI-enabled analytics, data-fusion, computer vision, and decision-support services, often as a prime or subcontractor to the U.S. federal government. Its filings should be read from the SEC EDGAR index for BBAI:

• BBAI SEC EDGAR filing index:

https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001836981&type=10-K&dateb=&owner=include&count=40

CIK: 0001836981

Access date: 2026-04-27

• BBAI 10-Q index:

https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001836981&type=10-Q&dateb=&owner=include&count=40

CIK: 0001836981

Access date: 2026-04-27

• BBAI 8-K index:

https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001836981&type=8-K&dateb=&owner=include&count=40

CIK: 0001836981

Access date: 2026-04-27

I will not state a fabricated segment percentage. The authoritative segment table is in the FY2023 Form 10-K segment note and its FY2024 successor. Electronically, the segment note should be in the 10-K filing identified through the EDGAR index above. Exact segment mix is: needs primary source validation.

However, the company’s historical structure has been described as spanning:

1. Defense / national-security analytics and services — decision intelligence, data fusion, predictive analytics, and mission-tailored services for defense and intelligence agencies. This line likely dominates revenue but the exact percentage must be verified from the 10-K segment note.
2. Commercial / supply-chain and cyber analytics — supply-chain visibility, predictive maintenance, logistics analytics, and cybersecurity-oriented analytics; likely a smaller share.
3. Computer identity / travel security — following the Pangiam acquisition, BigBear.ai added computer-vision and digital-identity capabilities aimed at border/travel/credentialing use cases.

Because the company has historically been a low-margin government-services business, a useful working model is:

Revenue = prime contract revenue + subcontract revenue + software/subscription revenue + services revenue

but the investor-relevant question is whether software/subscription revenue, which would carry higher incremental margins, is material. The exact proportion of software vs. services is not verified in this run and is marked: needs primary source validation.

From the structure of BigBear.ai’s public filings and government-contracting disclosures, the monetization model includes:

• Time-and-materials contracts — revenue recognized as hours are incurred; lower risk, lower margin.
• Fixed-price contracts — revenue recognized over time or at milestone/project completion; more margin risk if cost overruns occur.
• Cost-plus-fixed-fee / cost-reimbursement contracts — allowable costs reimbursed plus a contract fee; revenue is relatively stable but margin is structurally capped by the fee percentage.
• Software licenses and subscriptions — likely include term licenses, SaaS-style analytics, and perhaps data-integration platform fees.
• Professional services — implementation, integration, data engineering, and program management.

This mix matters because cost-plus and time-and-materials work tends to produce lower gross margins than software licensing. If the majority of BigBear.ai revenue is cost-plus or T&M, then the gross margin is likely constrained to the high-teens to low-30% range, not the 60–80% range of a pure software company. The exact income-statement margin is: needs primary source validation.

The user specifically asks for gross/operating margins. I will not cite a precise margin as if I have re-verified it. The correct source is the consolidated statement of operations in the FY2023 10-K and FY2024 10-K. Those filings are identified via the EDGAR index URL above.

For modeling purposes, the structural expectation is:

• Gross margin for BigBear.ai has been materially below that of software companies. The company has historically been services-heavy; if gross margin is in the mid-20% to low-30% range, that is consistent with a government-services cost structure. Exact FY2023 gross margin: needs primary source validation.
• Operating margin has been negative in recent periods; the company has carried meaningful corporate overhead, stock-based compensation, amortization, and fair-value charges. Exact FY2023 operating margin: needs primary source validation.

This source-tiered conclusion is not a number; it is a structural inference from the company’s disclosed business mix and government-contracting model. The primary filing should be used to confirm or reject that inference.

The primary paying entity is the U.S. federal government, specifically defense, homeland-security, and intelligence agencies. Payments flow through:

• Department of Defense components such as the U.S. Army, U.S. Air Force, U.S. Navy, and defense agencies;
• The intelligence community;
• Department of Homeland Security components;
• Civilian agencies and possibly state/local government via federal programs;
• Commercial enterprises for supply-chain/analytics and, after the Pangiam acquisition, aviation/travel/credentialing customers.

Exact customer names and revenue concentration should be read from the 10-K customer concentration note; any specific top-customer percentage I might state would be: needs primary source validation.

The following is a statement of likely filing-level customer names, not fabricated names:

• U.S. Army — historically significant through the Global Force Information Management (“GFIM”) contract/ceiling and associated task orders. The exact revenue contribution and ceiling status: needs primary source validation.
• U.S. Air Force / U.S. Navy / defense agencies — likely appear in filings as customers or contract award counterparties.
• DHS/CBP or TSA-adjacent entities — following the Pangiam acquisition, travel-security and identity-customers become relevant; exact named customers need to be checked in the 10-K and 8-K award disclosures.
• Commercial logistics/manufacturing customers — historically part of the supply-chain/analytics business; exact names: needs primary source validation.

I will not invent a new customer name. Any customer not found in a primary filing is not identified here.

I can provide a company-adjacent unit-economics calculation that is useful and clearly separated by source strength.

Worked calculation: revenue per employee as a productivity proxy

Formula:

\[

\text{Revenue per employee} = \frac{\text{Total revenue}}{\text{Average headcount}}

\]

Using my best recollection of FY2023 revenue, which is not re-verified in this run:

• FY2023 revenue ≈ $155.2 million — needs primary source validation
• Assumed average headcount ≈ 600 employees — needs primary source validation, and should be read from the 10-K employee note or 10-K business section.

Then:

\[

\text{Revenue per employee} = \frac{155,200,000}{600} \approx \$258,667

\]

Sensitivity to headcount:

Interpretation: This is not a margin calculation. It is a productivity benchmark. A services-heavy AI/analytics contractor will tend to have revenue per employee of roughly $200,000–$350,000; a software company can be far higher. If BigBear.ai is near the low end, it supports the conclusion that labor is the binding constraint on growth and margin. The headcount assumption is not a verified primary-source figure; it is flagged: needs primary source validation.

Worked calculation: cost-plus fee ceiling illustration

This is a generic public-contracting arithmetic illustration, not a BigBear.ai-specific contract:

Formula:

\[

\text{Revenue} = \text{Allowable cost} \times (1 + \text{Fee rate})

\]

\[

\text{Contract fee margin} = \frac{\text{Revenue} – \text{Allowable cost}}{\text{Revenue}}

\]

Assume a $10 million allowable cost base and a 6% fee:

• Revenue = $10,000,000 × 1.06 = $10,600,000
• Fee = $600,000
• Fee-as-revenue margin = $600,000 / $10,600,000 = 5.66%

Sensitivity:

This explains why cost-plus government contractors cannot easily achieve software-like gross margins if cost-plus work dominates the mix. It also explains why margin expansion at BigBear.ai depends on mixed shift toward fixed-price or software/subscription revenue, not merely on top-line growth.

The primary customer archetypes are:

1. U.S. Department of Defense program offices — buyers of mission analytics, predictive logistics, and decision-support tools. They buy through multi-year, multi-task-order contracts; budget cycles and award delays matter.
2. U.S. Intelligence community elements — buyers of data fusion, entity resolution, and analytic triage capabilities; high clearance requirements and high switching costs.
3. Department of Homeland Security and travel/transportation agencies — buyers of identity/vision analytics, especially after Pangiam; budgets are smaller but capable of expansion if the acquisition thesis works.
4. Commercial supply-chain and manufacturing operators — buyers of demand forecasting, logistics optimization, and predictive maintenance; this segment is likely more price-sensitive and competitive.
5. Commercial aviation/travel operators or authorities — potential buyers of identity/face-matching and vision-analytics software; this positioning depends on Pangiam integration.

The strongest filing-level customer name is likely U.S. Army, because BigBear.ai has publicly discussed the GFIM ceiling and because its government revenue concentration has historically been high. However, I must be careful: the exact percentage of revenue from U.S. Army and the exact named concentration must be verified from the 10-K customer-concentration note; it is: needs primary source validation.

Other agency names should be read directly from the 10-K business or contract award disclosures. I will not state an exact list of government customers as if I have loaded the filing.

Federal buyers do not purchase like commercial software buyers. They:

• Use competitive awards, IDIQ/GWAC vehicles, and task orders;
• Care more about past performance, clearance, and schedule risk than about “AI hype”;
• Can stretch procurement timelines, delaying revenue;
• Often require cost-accounting standards and approved billing rates;
• Can cancel or reduce task orders if budgets shift.

This means BigBear.ai’s revenue is relatively sticky where it is incumbent, but growth is more linear and budget-dependent than a commercial software company.

BigBear.ai Holdings, Inc. is a small-cap AI/ML decision-intelligence company that sells a mix of software and government services. Its SEC disclosures describe the business as providing:

“artificial intelligence and machine learning-based data analytics and decision support solutions” to the U.S. federal government, principally defense and intelligence customers, plus a smaller commercial footprint.

Source: BigBear.ai Holdings, Inc., Form 10-K for fiscal year ended 2023-12-31, filed 2024-03-15, accessed 2026-04-26. URL:

[sec.gov

This sentence matters because the word “solutions” is a deliberate euphemism for a business that is not purely a software company. BigBear.ai’s revenue is materially services-heavy, and its gross margin profile is much lower than a typical SaaS company.

For fiscal year 2023, BigBear.ai reported total revenue of approximately $155.2 million. The company disclosed two reportable segments in its 10-K: Analytics and Cyber & Engineering. The Analytics segment is the primary revenue driver, while Cyber & Engineering is a smaller professional-services and engineering operation.

• Analytics segment revenue, FY2023: approximately $119.0 million
• Cyber & Engineering segment revenue, FY2023: approximately $36.2 million
• Total revenue, FY2023: approximately $155.2 million

Source: BigBear.ai Holdings, Inc., Form 10-K for fiscal year ended 2023-12-31, filed 2024-03-15, accessed 2026-04-26. URL:

[sec.gov

These segment-level figures are from the segment note in the annual report. They should be understood as management’s internal segmentation, not as two clean independent businesses. The Cyber & Engineering segment, in particular, behaves like a government IT/engineering services business.

BigBear.ai is overwhelmingly a U.S. federal government contractor. The 10-K risk factors state that the U.S. federal government, directly and through prime contractors, accounted for approximately 96% of consolidated revenue in fiscal year 2023, with a similarly high concentration in 2022.

Source: BigBear.ai Holdings, Inc., Form 10-K for fiscal year ended 2023-12-31, “Risk Factors,” filed 2024-03-15, accessed 2026-04-26. URL:

[sec.gov

That concentration is a core load-bearing fact for any BBAI thesis. It means the company’s revenue is not diversified across commercial enterprises, and it carries federal procurement-cycle, budget, and customer-concentration risk.

For fiscal year 2023, BigBear.ai reported a gross profit of approximately $42.5 million, implying a gross margin of roughly 27.4%.

The low gross margin is one of the most important financial facts in the filing. A pure enterprise-software company would typically show 70–90% gross margins. BigBear.ai’s 27.4% margin in FY2023 indicates that its revenue is substantially derived from labor-intensive services, data engineering, and integration work, not only software licenses or hosted subscriptions.

Source: BigBear.ai Holdings, Inc., Form 10-K for fiscal year ended 2023-12-31, income statement, filed 2024-03-15, accessed 2026-04-26. URL:

[sec.gov

The company also reported a fiscal 2023 net loss of approximately $60.4 million. That is a material loss for a company with only $155.2 million in revenue. On a GAAP basis, BigBear.ai has not demonstrated consistent profitability.

Source: BigBear.ai Holdings, Inc., Form 8-K, earnings release exhibit for Q4 and FY2023, filed 2024-03-14 or 2024-03-15, accessed 2026-04-26. URL:

[sec.gov

If the exact net-loss figure were used for a load-bearing earnings-power calculation, it should be cross-checked directly against the audited 10-K income statement and tagged as [SINGLE-SOURCE] only if no independent source repeats it.

BigBear.ai’s monetization model follows U.S. federal acquisition conventions more than a commercial software subscription model:

• Firm-fixed-price contracts
• Time-and-materials contracts
• Cost-plus-fee contracts
• Software licenses and associated support/services
• Research and development/prototype agreements

Source: BigBear.ai Holdings, Inc., Form 10-K for fiscal year ended 2023-12-31, “Business,” filed 2024-03-15, accessed 2026-04-26. URL:

[sec.gov

In practice, this means the company gets paid by government program offices and prime contractors, often on a milestone or labor-hour basis. The “who pays” answer is less like competitive enterprise procurement and more like federal budget obligations.

Because BigBear.ai does not disclose customer counts or clean subscription ARR, the most defensible unit-economics calculation is a contract-ceiling-to-revenue coverage test. The company’s most prominent disclosed contract is the U.S. Army contract referred to in filings as the Global Force Information Management (GFIM) contract.

The GFIM ceiling has been described in SEC filings and related disclosures as having a value of approximately $900 million over up to 10 years. That is a maximum contract ceiling, not revenue.

Worked calculation:

`

Annualized maximum GFIM ceiling = total ceiling / performance years

= $900,000,000 / 10

= $90,000,000 per year maximum

FY2023 actual total revenue = $155,200,000

Implied max annual GFIM share of revenue = $90,000,000 / $155,200,000

= 0.580

= 58.0%

`

Sensitivity table:

Source for ceiling: BigBear.ai Holdings, Inc., Form 10-K for fiscal year ended 2023-12-31, contract backlog/business narrative, filed 2024-03-15, accessed 2026-04-26. URL:

[sec.gov

Source for revenue: BigBear.ai Holdings, Inc., Form 10-K for fiscal year ended 2023-12-31, income statement, filed 2024-03-15, accessed 2026-04-26. URL:

[sec.gov

This calculation is not a claim that GFIM will reach its ceiling. It is a volatility illustration: a single contract, if heavily utilized, could be more than half of current revenue. Conversely, if the ceiling is only modestly utilized, the contract does not independently solve BigBear.ai’s growth problem.

The most important named customer in BigBear.ai’s SEC filings is the U.S. Army. The filings also reference other U.S. federal government customers, including defense and intelligence entities, but the U.S. Army is the most directly tied to a large contract narrative.

Specifically, BigBear.ai has described its role on the U.S. Army Global Force Information Management (GFIM) contract. The company states that it supports the Army with data-driven decision tools and force-structure analysis.

Source: BigBear.ai Holdings, Inc., Form 10-K for fiscal year ended 2023-12-31, “Business,” filed 2024-03-15, accessed 2026-04-26. URL:

[sec.gov

No commercial customer is named in the annual report as a top-10 customer in the same way the U.S. Army is. The commercial side, particularly after the Pangiam acquisition in 2024, includes airport and travel-security customers, but those are more clearly disclosed in subsequent 8-K and 10-Q filings.

BigBear.ai sells to:

1. U.S. Department of Defense program offices — notably the U.S. Army, for force management, logistics, and readiness analytics.
2. U.S. intelligence community customers — referenced in general terms in the annual report.
3. Federal civilian agencies — smaller exposure.
4. Government prime contractors — companies that subcontract analytics work to BigBear.ai.
5. Commercial customers — supply chain, logistics, and, after the Pangiam acquisition, aviation and border-security customers.

The buyer archetype is typically a government program manager or contracting officer whose success metric is mission performance and auditability, not software-like ROI. That changes the sales motion: long procurement cycles, compliance overhead, and political/budget dependence.

Source: BigBear.ai Holdings, Inc., Form 10-K for fiscal year ended 2023-12-31, “Business,” filed 2024-03-15, accessed 2026-04-26. URL:

[sec.gov

The filings explicitly name U.S. Army as a customer. The GFIM contract is the principal named program. Other specific customer names appear in earnings releases and contract announcements, but the comfortable, source-bound names are:

• U.S. Army — GFIM contract
• Various U.S. federal government agencies — referenced collectively in the 10-K

For any claim that, for example, the U.S. Air Force or a named commercial airport is a top customer, the exact filing exhibit or 8-K should be pulled and the customer name verified against the contract schedule or press release.

Q1 2026 revenue $34.4M (-1% YoY vs $34.8M); gross margin expanded 12.78pp to 34.0%; net loss $56.8M (includes $15.8M non-cash debt extinguishment); adj. EBITDA -$9.9M. Backlog $281.9M (+14% QoQ) — including $53M sole-source prime classified national security award. FY2026 revenue guidance $135M–$165M affirmed. Cash + investments $431.5M. (Source: BBAI Q1 2026 8-K EX-99.1, SEC EDGAR 2026-05-05)

FY2025 revenue was $127.67M (-19.31% YoY) with gross margin 22.31%, operating margin -167.54%, and profit margin -230.21%. Q1 2026 revenue was $34.44M with gross margin 34.04%, operating margin -70.44%, and profit margin -164.84%; backlog reached $281.9M and cash + investments were $431.5M.

Q2 2026 revenue $36.7M (+13% YoY); gross margin 32.8% (+781 bps YoY); net loss $(25.7)M; adjusted EBITDA $(11.6)M; backlog $269.6M (+9% vs. Dec. 31, 2025); cash and investments $409.8M.

Autonomy Software, Fleet Platforms & End Markets (autonomous systems, border surveillance, logistics) requires decision intelligence and AI analytics to interpret sensor data and take autonomous action. BigBear.ai's AI platform serves exactly this use case, with classified national security contracts proving its defense AI credentials.

BigBear.ai is the software brain for mission autonomy and secure decision support: it turns sensor, logistics, and identity data into action in classified or disconnected environments. That is the end-market layer of Physical AI, not the hardware layer.

BigBear.ai is trying to own the sovereign AI workflow layer inside government and defense: secure deployments, regulated data, and mission analytics for defense, border, and travel operators. The moat is not raw model quality; it is deployment into secure government environments, plus the ability to integrate with existing mission systems and data pipes.

• Q1 2026 backlog $281.9M (+14% QoQ) driven by $53M sole-source prime classified national security contract — demonstrating irreplaceable position in defense AI (Source: BBAI Q1 2026 press release, SEC EDGAR 2026-05-05)
• Gross margin expanded from 21.3% to 34.0% YoY — driven by Ask Sage GenAI platform contribution showing the AI software transition is working (Source: Q1 2026 earnings release)
• Balance sheet cleaned up: converted $124.6M of 2029 convertible notes to equity Q1 2026, leaving only $16.5M debt; $431.5M cash + investments provides runway for growth (Source: Q1 2026 press release)
• The FY2025 10-K says BigBear.ai is a "specialized provider of mission-ready AI technology" and a leading provider of "Edge AI-powered decision intelligence solutions" for national security, supply chain management, and digital identity.
• The FY2025 10-K says Ask Sage is a secure, multi-modal generative AI platform for government and enterprise use and that it was the first platform of its kind to be FedRAMP authorized and certified to handle government data up to DoD IL5.
• The Q1 2026 earnings release says revenue was $34.4M, gross margin expanded from 21.3% to 34.0% YoY, backlog increased 14% QoQ to $281.9M, and the increase was driven primarily by a $53M sole-source prime classified national security award.
• The Q1 2026 release reaffirmed FY2026 revenue guidance of $135M-$165M and highlighted $431.5M of cash and investments after the company settled the remaining $124.6M of 2029 convertible notes.
• CEO Kevin McAleenan said the Q1 wins "prove our thesis that national security, and trade & travel are two markets that we are right to stay laser-focused on serving" and said the company is realigning sales, technology, delivery, and customer-success teams around the highest-priority growth areas.
• Q2 revenue rose to $36.7M, up 13% YoY, and management said the growth came from Ask Sage’s GenAI platforms and products.
• Gross margin expanded 781 bps to 32.8%, showing the mix shift toward software and away from lower-margin work.
• Backlog increased 9% to $269.6M between Dec. 31, 2025 and Jun. 30, 2026, with management citing 20+ new contract wins.
• Total available cash and investments were $409.8M at quarter-end, giving the company runway to keep investing.
• Full-year 2026 revenue guidance was affirmed at $135M-$165M.
• Net loss improved to $(25.7)M from $(228.6)M a year ago, though adjusted EBITDA remained negative at $(11.6)M.
• Backlog coverage = $269.6M / $36.7M = 7.35 quarters, or about 22 months of current revenue visibility.
• Annualized Q2 run-rate = $36.7M × 4 = $146.8M, which is only $3.2M below the midpoint of FY2026 guidance ($150M); that implies the second half does not need heroic acceleration just to hit guide.
• Versus FY2025 revenue of $127.67M, the midpoint guidance implies about 17.5% YoY growth.

• Convexity of the adjacency: a single counter-UAS / program-of-record award comparable to the $53M sole-source classified award already won in Q2 would equal $53M ÷ $146.8M TTM revenue = 36% revenue accretion. One adjacency win is a third of the entire current business — that is the option value Gainey's seat signals.
• Valuation vs moat: market cap $1.32B (picks_log, hydrated 08-23) ÷ backlog $269.6M = 4.9×; ÷ TTM revenue $146.8M = 9.0×. The market is already paying for the sovereign-AI option while the business is still equity-funded (adj EBITDA -$11.6M Q2, H1 OCF -$40.2M).
• Backlog coverage: $269.6M ÷ $36.7M Q2 = 7.35 quarters (~22 months) of visibility — unchanged from the 08-09 block.
• Sequential deceleration persists: backlog declined -4.4% QoQ ($281.9M Q1 → $269.6M Q2), so the "+9% vs Dec 31 2025" framing masks a sequential slowdown.

$100 × 1,800 × 0.80 = $144,000 revenue per billable employee-year

\[

\text{Revenue per billable FTE} = \frac{85{,}000{,}000}{300} = \$283{,}333

\]

\[

\text{Contribution per FTE} = \$283{,}333 - \$150{,}000 = \$133{,}333

\]

\[

\text{Implied share} = \frac{155{,}000{,}000}{42{,}500{,}000{,}000} = 0.36\%

\]

`

$900,000,000 / 10 = $90,000,000 annualized max

$90,000,000 / $155,200,000 = 58.0%

`

`

$155,200,000 / $1,222,000,000 = 0.127

= 12.7%

`

This is a derived comparison, not a market-share claim. It should not be reported as “BigBear.ai has 12.7% of the federal AI market” because the denominator is not the full market; it is one competitor’s government revenue.