AI Market Reckoning: Why BigBear.ai and C3.ai Are Losing Ground to Chip Leaders
The artificial intelligence software market is undergoing a brutal consolidation, with two prominent players—BigBear.ai and C3.ai—struggling to maintain relevance as their revenues contract and losses mount. Meanwhile, semiconductor giant Broadcom ($AVGO) is capitalizing on explosive demand for AI chips, positioning itself as the superior investment thesis in an increasingly crowded landscape. This divergence highlights a critical shift in where artificial intelligence value is being created—and where investors should be directing their capital.
The Deteriorating Fundamentals of BigBear.ai and C3.ai
BigBear.ai presents a particularly concerning picture of decline. The company's revenue has contracted significantly over the past four years, falling from $146 million in 2021 to $128 million in 2025—a 12.3% decline in absolute terms that becomes more alarming when considering the explosive growth in the AI sector during this period. More troubling still, net losses have widened dramatically to $294 million, indicating the company is burning through capital at an accelerating rate while its top-line growth stalls.
C3.ai ($AI) faces an even more precipitous trajectory. The company's fiscal 2025 revenue of $389 million is projected to decline to just $251 million by fiscal 2028—a 35.5% revenue contraction over three years. This projection suggests that C3.ai's enterprise software solutions have failed to gain the traction necessary to compete effectively against both established software giants and nimble, well-capitalized AI startups.
Key metrics illustrating the competitive pressure:
- BigBear.ai revenue decline: $146M (2021) → $128M (2025)
- BigBear.ai net losses: $294 million (2025)
- C3.ai revenue trajectory: $389M (FY2025) → $251M (FY2028)
- Sector growth rate: Double-digit annually (industry average)
Both companies share a common problem: they are trapped in a middle ground, neither large enough to compete with Microsoft, Google, and Amazon on scale, nor nimble enough to outmaneuver specialized AI startups focused on specific use cases. Their enterprise software offerings have failed to achieve the widespread adoption necessary to achieve profitability, leaving them with deteriorating unit economics and limited paths to growth.
Broadcom's Dominant Position in AI Infrastructure
In stark contrast, Broadcom ($AVGO) is experiencing the kind of explosive growth that should capture investor attention. The company's AI chip sales surged 65% to reach $20 billion in fiscal 2025, representing a remarkable acceleration from its broader business. More impressively, Broadcom projects AI chip sales will reach $60-90 billion by fiscal 2027, implying a threefold to fourfold increase from current levels over just two years.
This growth trajectory reflects a fundamental reality: artificial intelligence requires semiconductor infrastructure. Every transformer model, every large language model training run, every AI inference operation depends on chips. Broadcom, along with competitors like Nvidia ($NVDA), sits at the critical chokepoint where AI demand translates into hardware necessity.
What makes Broadcom particularly attractive on a valuation basis is that despite this extraordinary growth, the company is trading at a reasonable 18x next year's adjusted EBITDA—a multiple that appears modest compared to the high-flying AI infrastructure stocks that have commanded 25-40x multiples. This valuation discount to growth represents a meaningful opportunity for disciplined investors.
Market Context: The AI Hardware vs. Software Divide
The divergence between BigBear.ai, C3.ai, and Broadcom reflects a broader market dynamic that has become increasingly clear over the past 18 months. The artificial intelligence boom has been primarily driven by infrastructure demand—the computational power necessary to build, train, and deploy AI models. This infrastructure layer has proven to be a more defensible, higher-margin business than enterprise AI software.
The software companies that have succeeded in the AI era have done so by either:
- Integrating AI into existing platforms (Microsoft with Copilot, Salesforce with Einstein)
- Targeting specific, mission-critical applications where AI provides exceptional ROI (specialized healthcare AI, manufacturing optimization)
- Leveraging first-mover advantages and network effects (OpenAI, Anthropic in the foundational model space)
BigBear.ai and C3.ai have failed to accomplish any of these objectives effectively. They occupy a crowded middle market of horizontal AI software platforms—a category that has proven difficult to monetize at scale. The rise of free or low-cost AI tools, combined with the ability of enterprise software leaders to rapidly integrate AI capabilities into their platforms, has compressed margins and growth prospects for pure-play AI software vendors.
Meanwhile, semiconductor manufacturers benefit from the inelasticity of demand. Data centers require chips regardless of which AI model wins the competitive battle. As enterprises race to deploy AI infrastructure, chipmakers capture value at the foundation of the entire technology stack.
Investor Implications and Market Takeaways
For equity investors, this divergence carries critical implications:
Avoid the AI Software Middle Market: Companies like BigBear.ai and C3.ai lack the scale to compete with platform leaders and the specialization to command premium valuations. Their revenue declines suggest that market adoption is deteriorating, not recovering. Further deterioration in unit economics seems likely, which could force difficult strategic decisions or restructuring.
Favor Infrastructure and Semiconductor Plays: Broadcom's growth trajectory, combined with its reasonable valuation multiple, suggests the semiconductor sector remains the most defensible way to gain exposure to AI trends. The company benefits from structural tailwinds—enterprise AI deployment, cloud computing expansion, and emerging applications like autonomous vehicles—that show no signs of abating.
Recognize Valuation Disparities: The fact that pure-play AI software companies are facing revenue declines while semiconductor companies are growing at 65%+ annually, yet trading at lower multiples, suggests market inefficiencies that disciplined investors should exploit.
The broader market lesson is that not all AI exposure is created equal. The investors who prospered during the railroad boom weren't those who invested in every startup promising to revolutionize transportation—they were those who backed the companies building the rails and locomotives. In the AI era, Broadcom and its semiconductor peers are building the rails. BigBear.ai and C3.ai are betting on applications that may or may not be built on those rails.
As enterprise AI spending accelerates over the next 24 months, this distinction will become increasingly important. Investors who have mistakenly assumed that "AI exposure" automatically means enterprise software companies are likely to face significant disappointment. The real value in artificial intelligence, at least for now, flows to those controlling the computational infrastructure upon which all AI applications ultimately depend.