C3.ai
B2B subscription + consumption software; sells a horizontal AI platform (C3 AI Platform) plus prebuilt/generative/agentic AI applications, largely co-sold through hyperscaler alliances (Microsoft Azure, AWS, Google Cloud) and the Baker Hughes energy JV; concentrated in large enterprise + US government/defense accounts
Earnings, margins, COGS & capex
FY2026 (ended Apr 30, 2026) was a reset year: revenue fell to $250.3M from $389.1M in FY2025, driven by the transition away from large multi-year subscription 'elephant' deals toward consumption-based pricing, deals slipping, and admitted weak sales execution in North America and Europe. Gross margin also deleveraged as fixed hosting/delivery cost hit lower revenue: GAAP gross margin fell to ~31% ($77.4M) and non-GAAP to ~46% ($116.2M), from ~61%/~70% in FY2025. GAAP loss from operations widened to $(498.5)M (inflated by restructuring and stock-based comp); non-GAAP operating loss was $(217.8)M. GAAP/non-GAAP net loss per share were $(3.35)/$(1.35). The company launched a restructuring (~26% headcount cut) targeting ~$135M annualized non-GAAP cost savings and a path to non-GAAP profitability, and founder Thomas Siebel resumed the CEO role (effective May 8, 2026) after a 2025 health-related leave, with Stephen Ehikian continuing as President. FY2027 guidance is telling: revenue $210-240M — still declining — with a narrower non-GAAP operating loss of $(128)-$(160)M. Balance sheet remains a key asset: ~$575M cash/investments and no debt.
Income statement — where each revenue dollar goes
% of revenueOf every $1 of revenue, ~0¢ is cost of goods and ~0¢ operating expense, leaving ~100¢ of operating profit.
Revenue trend
Margins
sharp compression on consumption-mix shift + revenue deleverage against fixed hosting/delivery cost
deleveraged hard — the ~70% software-model margin is NOT holding at current volume; recovery depends on revenue stabilizing
sharply worse YoY on revenue decline + restructuring/SBC charges
FY2027 guided to $(128)-$(160)M loss — narrower but still a loss
COGS structure
Primarily cloud-hosting/compute for the platform and delivery, plus support and prioritized-engineering-services labor. FY2026 gross margin compressed materially (non-GAAP ~46% vs ~70% prior year) because those largely fixed hosting/delivery costs did not scale down with the ~36% revenue decline — so gross margin is now a problem too, not just the top line. The consumption-pricing shift and hyperscaler co-sell economics are added swing factors.
Capex
Low — no meaningful physical plant; the company does not build data centers (rides Azure/AWS/GCP). Capex is a low single-digit percentage of revenue (office/IT), immaterial to the investment case; cash burn is driven by operating loss, not capex.
Latest earnings
Q4 revenue in the guided range ($51.6M vs $48-52M) and Q4 non-GAAP loss from operations of $(54.4)M beat the guided $(56)-$(64)M; but full-year growth was sharply negative and FY2027 guidance points revenue lower again — the Street treated FY26 as a major reset, not a beat. Q4 GAAP loss from operations was $(121.2)M.
Company guided FY2026 revenue $246.7-$250.7M (landed $250.3M) and Q4 $48-$52M (landed $51.6M). FY2027 guidance: total revenue $210.0-$240.0M (a further ~4-16% decline) and non-GAAP loss from operations $(128.0)-$(160.0)M (narrower than FY2026's $(217.8)M); Q1 FY2027 revenue $50.0-$54.0M, non-GAAP loss $(40.5)-$(48.5)M. Message is a restructuring-driven path toward non-GAAP profitability, but a return to top-line growth is a later-year aspiration, not FY2027 guidance.
- FY2026 revenue
- $250.3M (-36% YoY)
- FY2026 gross margin
- ~31% GAAP / ~46% non-GAAP
- FY2026 non-GAAP loss from ops
- $(217.8)M
- FY2026 GAAP loss from ops
- $(498.5)M
- FY2026 EPS
- $(3.35) GAAP / $(1.35) non-GAAP
- Cash + investments
- ~$575.4M, no debt
- FY2027 revenue guide
- $210-240M (still declining)
- Restructuring
- ~26% headcount cut; ~$135M targeted annualized non-GAAP cost/cash-burn reduction
Growth drivers
- Agentic AI / generative-AI product cycle — C3 is pivoting its narrative to autonomous 'agentic' enterprise applications, positioned as the FY2028+ growth vector (FY2027 revenue is still guided down)
- Consumption-based pricing — intended to lower adoption friction and expand seats/usage over time (near-term it depressed reported revenue by removing large prepaid contract recognition)
- Hyperscaler co-sell — alliances with Microsoft, AWS and Google Cloud; ~73% of FY2025 agreements ran through partners/collaborations
- US government & defense — sustained demand across Air Force/DoD/intelligence programs, a relatively durable vertical (restructuring explicitly re-focuses on defense AI)
- Energy JV — Baker Hughes alliance (oil & gas + broader energy AI apps)
- Industry diversification — FY2025 revenue spanned 19 industries with non-oil-&-gas revenue up ~48% YoY, reducing single-vertical dependence
Reported financials — SEC EDGAR
Audited GAAP figures pulled from SEC filings · latest filing 2026-06-24. The audited primary-source spine — not financial advice.
Revenue — annual (GAAP)
Margins & balance sheet — FY’26
Bull & bear
A cash-rich, founder-led turnaround trading near distressed multiples: if Siebel stabilizes sales execution, laps the pricing transition, restores gross-margin leverage, and lands the agentic-AI product cycle, C3 can bend back to growth and non-GAAP profitability off a ~$575M net-cash base with minimal dilution risk.
- Downside is cushioned by ~$575M cash and no debt — roughly $3.7-4/share of net cash under a ~$1.4B cap, so a large share of the equity is asset-backed rather than pure growth optionality
- The revenue decline is substantially self-inflicted (pricing transition + sales execution), which is fixable, not a demand collapse — management still cites strong customer validation and 19-industry breadth
- Restructuring removes ~$135M of annualized cost; FY2027 guidance already narrows the non-GAAP loss to $(128)-$(160)M, and a credible break-even path can re-rate a beaten-down name fast
- Distribution moat via Microsoft/AWS/Google co-sell and the Baker Hughes JV is intact — these did not walk away during the bad year
- Founder Thomas Siebel has done enterprise-software scale-and-exit before (Siebel Systems -> Oracle); his return signals conviction and hands-on control
- Agentic AI is the right wave to ride, and C3's application-layer positioning could differentiate from raw-model and infra vendors if execution follows
A structurally challenged application-platform vendor losing the enterprise-AI race to hyperscalers and Palantir, with a broken growth model (revenue still guided down in FY2027), collapsed gross margins, heavy losses, key-man dependency, and a real risk the equity drifts toward its cash value as the business keeps shrinking.
- A 36% revenue decline (Q3 -46%) and FY2027 guided down again ($210-240M) is not a rounding error — it signals the platform may be losing relevance as hyperscalers bundle AI and Palantir wins the enterprise/gov narrative
- Gross margin collapsed too: non-GAAP GM 70%->46%, GAAP 61%->31% — so the 'the model is fine, just fix the top line' bull leg is weaker than it looks
- Burning ~$218M non-GAAP operating loss on $250M revenue; even guided FY2027 is a $(128)-$(160)M loss on lower revenue — the model isn't self-evidently profitable on a shrinking base
- Key-man risk is acute — the thesis leans on one 70-something founder whose 2025 health leave forced a CEO scramble; that's a live governance overhang
- Consumption pricing is an unproven bet that already blew up reported revenue; there's no guarantee usage expansion offsets the lost prepaid contract economics
- Persistent heavy stock-based compensation dilutes holders and masks the true cost structure in non-GAAP figures
- With a ~$1.4B cap on ~$575M cash, the market is pricing barely ~$825M of enterprise value onto a still-declining business — a value trap unless growth actually returns
What it is worth
EV/Sales on a distressed, net-cash-heavy turnaround: market cap ~$1.4B less ~$575M net cash = ~$825M enterprise value on $250.3M FY2026 revenue = ~3.3x trailing EV/Sales (~5.6x P/S on market cap); on FY2027 guided revenue of $210-240M it is ~3.4-3.9x forward EV/Sales. Not a DCF case while FCF is deeply negative — the equity is priced as cash floor + turnaround optionality.
Revenue keeps shrinking past FY2027 and losses persist; EV compresses toward cash value (~$575M), pulling the equity toward its net-cash floor (~$3.7-4/share area) with take-under risk — a value trap despite the strong balance sheet.
Revenue lands near FY2027 guidance ($210-240M) and cost cuts narrow losses toward the guided $(128)-$(160)M; equity stays range-bound around net cash plus a modest software multiple (~3x EV/Sales), roughly consistent with the current ~$1.4B cap.
Turnaround works: revenue stabilizes then reaccelerates past FY2027 and non-GAAP profitability is reached; the market re-rates toward ~4-6x EV/Sales on a growing base, implying materially higher equity value than the ~$1.4B cap — plus optionality as an acquisition target.
With revenue down 36% and guided lower again in FY2027, plus gross margin now deleveraged (non-GAAP ~46%), C3 trades at a fraction of profitable-grower Palantir's multiple. The floor is anchored by ~$575M net cash (~$3.7-4/share); the ceiling depends on whether Siebel restores growth and reaches non-GAAP profitability. Ranges below are scenario illustrations, not price targets — not financial advice.
SWOT
Strengths
- Fortress balance sheet — ~$575M cash/investments, zero debt — years of runway even at elevated burn, and optionality to fund a turnaround without dilution at a depressed price
- Genuine enterprise/government install base and reference customers (Baker Hughes, US Air Force/DoD, major energy and industrial names) that are hard for challengers to replicate
- Deep hyperscaler co-sell alliances (Microsoft, AWS, Google Cloud) providing distribution the company can't fund alone
- Founder-operator back at the helm — Thomas Siebel (built and sold Siebel Systems to Oracle) is a credible enterprise-software turnaround executor
- Non-GAAP loss narrowing in guidance (FY2027 $(128)-$(160)M vs FY2026 $(217.8)M) as the ~$135M cost cut lands
Weaknesses
- FY2026 revenue fell 36% (Q3 -46%) and FY2027 is guided down again — the core growth thesis broke and hasn't inflected
- Gross margin deleveraged hard — non-GAAP GM fell to ~46% (from ~70%) and GAAP to ~31% — the 'unit economics work' claim no longer holds at current volume
- Chronic GAAP unprofitability and heavy cash burn; non-GAAP op loss ~$218M on $250M revenue
- Self-inflicted execution failure — management admitted weak North America/Europe sales execution and a mishandled pricing-model transition
- High historical stock-based compensation diluting shareholders and flattering non-GAAP metrics
- Leadership instability in 2025-2026 (CEO health leave, handoff to Ehikian, then Siebel's return) — governance/succession overhang for a founder-dependent company
- Perceived crowding-out by hyperscalers' own AI platforms and by Palantir in the government/enterprise-AI narrative
Opportunities
- Agentic AI wave — if C3's agentic enterprise apps land, it re-enters a large and expanding TAM with a differentiated application layer
- Consumption pricing, once lapped, could compound via usage/seat expansion and lower the barrier to new logos
- Restructuring to non-GAAP profitability would remove the 'melting ice cube' bear framing and support a re-rate
- Government/defense AI spending remains a structural tailwind for a cleared, referenceable vendor — restructuring explicitly re-centers on defense AI
- M&A optionality — with ~$575M cash and a ~$1.4B cap, C3 is both a potential consolidator of small AI-app startups and, conversely, a plausible take-private/strategic target
Threats
- Hyperscalers (Microsoft, AWS, Google) and Palantir competing directly for the same enterprise-AI budgets, often bundled or with stronger momentum
- Continued revenue decline (FY2027 guided lower) could push valuation toward net-cash value, inviting activist or take-under scenarios
- Enterprise-AI buyers building in-house on foundation models + open tooling, disintermediating the application-platform layer
- Customer/vertical concentration risk (energy via Baker Hughes JV — government) — a single large renewal miss moves the P&L
- Execution risk on a turnaround that depends heavily on one founder whose 2025 health issues remain a stated overhang
Moats, dependencies & bottlenecks
Moats
Real, contracted distribution C3 can't self-fund — but it's rented reach, and the same hyperscalers also compete with their own AI platforms; ~73% of FY2025 deals ran through partners
Cleared, referenceable deployments (defense, energy, utilities) are sticky and hard to replicate, but concentration cuts both ways
Weak-to-Moderate Differentiation eroding as foundation-model + open tooling let buyers build in-house; the application layer must keep proving it's cheaper/faster than DIY
Strong (as a defensive moat) ~$575M net cash buys time to fix execution — a durability advantage, not a competitive one
Dependencies
Infrastructure + go-to-market C3 runs on and co-sells through the hyperscalers, who are simultaneously its largest competitive threat; ~73% of FY2025 deals ran through partners
Customer + channel (energy vertical) A major energy-vertical revenue and pipeline source — renewal terms and vertical concentration are watch items
Customer concentration Durable demand and the restructuring's re-focus area, but exposed to procurement cycles, continuing resolutions, and competition from Palantir
Anthropic, et al., via hyperscalers) C3's generative/agentic apps ride third-party LLMs; pricing, access, and capability shifts flow through its product economics
Key person / leadership Turnaround is founder-dependent; his 2025 health leave and 2026 return are a stated governance overhang
Advantages
- ~$575M cash, no debt — rare downside protection and self-funding runway for a turnaround at a depressed valuation
- Contracted hyperscaler distribution (Microsoft/AWS/Google) and the Baker Hughes JV still in place
- Proven, cleared enterprise/government deployments that are difficult and slow for challengers to earn
- Founder with a prior enterprise-software scale-and-exit track record now directly in control
- ~$135M cost cut already narrowing the guided non-GAAP loss into FY2027
Weaknesses
- FY2026 revenue down 36% and FY2027 guided down again; growth thesis broken and demand durability in question
- Gross margin collapsed (non-GAAP 70%->46%, GAAP 61%->31%) — profitability harder than the 'top-line only' framing implied
- Large ongoing operating losses and cash burn; profitability unproven on a shrinking base
- Self-inflicted execution and pricing-transition missteps
- Heavy stock-based compensation and dilution
- Key-man / leadership-continuity risk and 2025-2026 governance churn
- Strategic squeeze between hyperscalers (partners and rivals) and Palantir
Bottlenecks
- Sales execution — management's own diagnosis; conversion of pipeline in North America and Europe is the near-term gating factor
- Lapping the consumption-pricing transition before reported revenue can reliably grow again (FY2027 is still guided down)
- Restoring gross-margin leverage — non-GAAP GM at ~46% needs revenue to scale back up against fixed hosting cost
- Proving non-GAAP profitability post-restructuring without further starving growth investment
- Product differentiation vs hyperscaler-native AI and Palantir at the point of enterprise buying decisions
- Re-establishing management/board stability and succession clarity around a founder-led turnaround
Top signals & trends
Top signals
Mixed / cautiously bullish · Signals conviction and hands-on turnaround control, but concentrates key-man risk and follows leadership churn
Bullish on cost, bearish on demand · Credible path to narrower non-GAAP losses, but a 26% cut also implies management sees a smaller near-term business
The core signal that broke the growth story, and guidance says it keeps shrinking in FY2027 before any hoped-for inflection
Margins deleveraged with volume — recovery is gated on revenue re-scaling, not just cost cuts
Modestly bullish · Cost discipline is showing up faster than the top line
Value-trap watch · Small implied enterprise value — either deep-value setup or a business drifting toward cash worth
Right wave, unproven execution; narrative-heavy until bookings confirm it
Trends
Positive if captured · Expands TAM and gives C3 a fresh product wedge — but hyperscalers and startups chase the same wave
Compresses the independent application-platform layer C3 occupies
Disintermediation risk for prebuilt-app vendors
Structural tailwind for cleared, referenceable vendors — shared with Palantir; C3's restructuring re-centers here
Industry-wide shift; long-term expansion vehicle but the near-term reported-revenue headwind that broke C3's FY2026
Negative near-term · Loss-making, decelerating AI stories are out of favor vs profitable compounders
Ecosystem & competitor graph
Suppliers feed the company; customers pull from it. Line thickness shows the strength of each tie (supply-chain dependency, customer earnings contribution). Hover to isolate a tie.
Primary cloud infrastructure host and co-sell channel
Cloud infrastructure and marketplace/co-sell partner
Cloud infrastructure and co-sell partner
Anthropic, et al., accessed via hyperscalers) LLM/generative capability underpinning C3's generative + agentic applications
Anchor energy customer and JV partner
US Air Force / Department of Defense / intelligence agencies Major government/defense customer base for AI applications; restructuring re-centers on defense AI
Energy-sector enterprise customer (representative of the oil & gas vertical)
Energy customer using C3 for asset/operations AI (representative)
Utility and manufacturing deployments spanning C3's ~19-industry FY2025 base
The defining enterprise/government-AI competitor and narrative winner — profitable, fast-growing, directly overlapping in defense and industrial AI
Both C3's top distribution partner (Azure) and its largest competitive threat via Azure AI/Copilot/agent platforms
Cloud partner and rival — SageMaker/Bedrock and AWS-native AI services compete for the same build-vs-buy budget
Partner and rival via Vertex AI and Gemini-based enterprise agents
Data-cloud platform increasingly pushing AI/ML app development and agents on governed enterprise data
Agentforce and enterprise AI apps compete for the applied-AI budget in large accounts
watsonx enterprise/government AI, consulting-led — overlaps in regulated and public-sector deployments
Private (not listed) — data + AI lakehouse and model/agent tooling; a major build-side alternative to C3's platform
Private (not listed) — enterprise AI/agent platform competing for applied-ML workloads