Thomas Siebel's Second Act: Vision or Delusion?
Thomas Siebel is a software legend—he built Siebel Systems into the dominant CRM platform of the 1990s before selling to Oracle. After a near-fatal elephant attack in Tanzania and years of recovery, Siebel returned with C3.ai in 2009, betting that AI would transform enterprises just as CRM had. By 2020, C3.ai's IPO valued the company at $10 billion, and Siebel's vision seemed vindicated. But by 2025, the stock trades at $3 billion—a 70% decline—as investors question whether C3.ai has a sustainable business model or if it's being commoditized by hyperscalers.
Siebel's pitch remains compelling: enterprises need industry-specific AI applications, not just infrastructure. C3.ai's platform offers 40+ pre-built apps for predictive maintenance (oil & gas), supply chain optimization (manufacturing), anti-money laundering (financial services), and readiness analytics (defense). These aren't generic ChatGPT wrappers—they're verticalized solutions trained on industry data and embedded into enterprise workflows. The problem: Microsoft, Google, and AWS are building similar capabilities, and they have distribution, capital, and customer relationships that C3.ai can't match. Siebel argues C3.ai's multi-cloud strategy (works on Azure, AWS, Google) creates optionality—but skeptics counter that enterprises will simply buy AI from their existing cloud provider. The bear case: C3.ai is a feature, not a product, and will be crushed by hyperscaler competition.
Business Model & Competitive Moat
C3.ai sells subscriptions to its C3 AI Platform and pre-built applications. C3 AI Platform provides infrastructure for building, deploying, and operating enterprise AI applications—model training, data integration, governance, monitoring. C3 AI Applications are industry-specific solutions like CRM Next (customer engagement), Production Schedule Optimization (manufacturing), and Readiness (defense logistics). Revenue comes from subscription licenses based on data volume, users, or compute consumption. The company targets Fortune 500 enterprises in energy, manufacturing, financial services, defense, and healthcare.
C3.ai's supposed competitive advantages are debatable: Industry expertise—15+ years of vertical focus creates domain knowledge in energy, manufacturing, defense; pre-built applications—40+ apps reduce time-to-value vs. building from scratch; model-agnostic platform—works with any ML framework or cloud provider; data integration—connects to legacy enterprise systems (SAP, Oracle, historians); governance and explainability—enterprise-grade compliance and auditability. However, these advantages are fragile—hyperscalers are adding vertical expertise, open-source tools (LangChain, LlamaIndex) democratize app development, and Microsoft Copilot embeds AI directly into workflows. C3.ai's moat may be a mirage.
Financial Performance
C3.ai's financials tell a story of growth without profitability—a dangerous combination in 2025's market:
- •Revenue (FY2025 guidance): $350-370M, up 20-24% YoY; growth accelerating but from small base
- •Subscription Revenue: 90%+ of total; recurring but short contract durations (1-2 years) create uncertainty
- •Operating Margin: -15% despite cost cuts; GAAP losses persist after 15 years in business
- •Free Cash Flow: Negative $30-50M annually; cash burn continues despite revenue growth
- •Cash Position: $800M+ cash and investments; provides ~5 years of runway at current burn
- •Customer Concentration: Baker Hughes 24%, top 10 ~60%—alarming dependency
- •Retention: Net dollar retention undisclosed but expansion/contraction concerns persist
C3.ai guides to break-even profitability "soon" but has made similar promises for years. The path to profitability requires either gross margin expansion (currently ~70%, good for software) or massive scale (10x revenue) to absorb fixed costs. Neither seems imminent given competitive pressures.
Growth Catalysts
- •Generative AI Tailwind: Enterprise AI budgets exploding; C3.ai positioned to capture share
- •Federal Government Expansion: Defense contracts (Readiness app) could scale rapidly with DoD adoption
- •Partnership Momentum: Microsoft, Google, AWS partnerships providing distribution and co-selling
- •Industry Verticalization: Deep expertise in oil & gas, manufacturing could create sticky customers
- •International Expansion: Currently ~80% U.S. revenue; Europe, Asia untapped markets
- •M&A Optionality: Siebel owns 20%+ stake; acquisition by Microsoft, Oracle, or private equity possible
- •Profitability Achievement: Break-even would remove major investor concern and re-rate valuation
Risks & Challenges
- •Hyperscaler Competition: Microsoft, Google, AWS offering competing AI platforms with superior distribution and capital
- •Commoditization: Open-source tools and OpenAI APIs making custom AI development accessible without C3.ai
- •Customer Concentration: Baker Hughes 24% of revenue; loss would be catastrophic
- •Profitability: 15 years unprofitable raises existential questions; when does burn stop?
- •Execution Risk: Enterprise AI projects are complex, long sales cycles, high implementation costs
- •Valuation: 8x sales is expensive for unprofitable software with uncertain competitive position
- •Founder Risk: Siebel is 72; succession unclear, and company is deeply tied to his vision and relationships
Competitive Landscape
C3.ai competes with hyperscalers and specialized AI vendors. Microsoft offers Azure AI, Copilot, and industry solutions—massive distribution advantage. Google Cloud has Vertex AI and industry-specific models. AWS provides SageMaker and AI services. Palantir (PLTR) competes in defense and government AI with better financials ($2.5B revenue, profitable). Databricks (private) offers data + AI platform. Snowflake (SNOW) adding AI capabilities. Pure-play enterprise AI is a crowded, brutal market.
| Company | Market Cap | Revenue | Profitability | Key Strength | AI Positioning |
|---|---|---|---|---|---|
| C3.ai (AI) | $3B | $350M | Unprofitable | Vertical apps | Enterprise AI platform |
| Palantir (PLTR) | $140B | $2.5B | Profitable | Defense/Gov | AI-powered analytics |
| Snowflake (SNOW) | $50B | $3B | Unprofitable | Data platform | Adding AI layers |
| Databricks | $43B* | $2B+ | Near breakeven | Data + AI | Unified analytics |
| Microsoft | $3T | $250B+ | Highly profitable | Everything | Copilot embedded |
C3.ai is the smallest, least profitable player competing against giants. Palantir's success (profitable, $140B market cap) shows enterprise AI can work—but Palantir has government contracts and operational leverage C3.ai lacks. The market is asking: why doesn't Microsoft just build C3.ai's apps?
Who Is This Stock Suitable For?
Perfect For
- ✓High-risk, high-reward speculators comfortable with total loss
- ✓AI thematic investors wanting pure-play enterprise AI exposure
- ✓Believers in Thomas Siebel's vision and execution
- ✓M&A arbitrage players betting on acquisition by hyperscaler
- ✓Contrarians betting market has over-discounted challenges
Less Suitable For
- ✗Risk-averse investors (unprofitable, concentrated customers, existential competitive threats)
- ✗Income investors (no dividend, burning cash)
- ✗Value investors (8x sales for unprofitable company is expensive)
- ✗Those seeking near-term profitability (path remains unclear)
- ✗Conservative portfolios (volatility and binary outcomes)
Investment Thesis
C3.ai is a binary bet: either Thomas Siebel is right that vertical AI applications create a defensible business, or hyperscalers commoditize the category and C3.ai becomes irrelevant. The bull case hinges on C3.ai achieving scale in defense, energy, and manufacturing—markets where domain expertise and regulatory compliance matter more than raw infrastructure. If Baker Hughes, Shell, and the Department of Defense standardize on C3.ai apps, the company could reach $1B+ revenue and profitability, justifying a $10-15B valuation (3-5x current). The generative AI boom provides tailwinds, and partnerships with Microsoft/Google/AWS offer distribution.
The bear case is more compelling: Microsoft will embed AI into Dynamics, Power Platform, and Azure, eliminating the need for C3.ai. OpenAI APIs and open-source tools democratize custom AI development. Baker Hughes churn would crater revenue. Profitability remains elusive, and cash burn continues. At 8x sales for an unprofitable company with existential competitive threats, the valuation offers no margin of safety. This is a speculative trade, not an investment. For those betting on Siebel's vision, position sizing is critical—this is a 1-2% portfolio allocation at most, with acknowledgment that total loss is possible.