Enterprise AI Market Heats Up as 'Land Grab' Begins
The enterprise AI market is undergoing a consolidation phase, with major platforms like Microsoft and Google bundling agentic assistants into their suites. Enterprise search provider Glean is positioning itself as an underlying intelligence layer for these systems. Competitors like Hebbia and Cohere are racing to differentiate on orchestration and reliability, while alternatives are gaining traction with buyers seeking transparent pricing.
- Glean recently raised $150 million in Series F funding, bringing its valuation to $7.2 billion. The company has raised a total of $765 million and surpassed $100 million in annual recurring revenue in its last fiscal year. - Competitor Hebbia, which focuses on vertical AI for finance and legal sectors, raised a $130 million Series B at a $700 million valuation. Hebbia differentiates by focusing on complex, multi-step workflows rather than general enterprise search, with pricing starting at $10,000 per year for full platform access. - Cohere, another key competitor, has raised nearly $1.5 billion and was valued at approximately $6.8–$7 billion by mid-2025. The company focuses on enterprise and government clients, emphasizing data privacy and offering models that can be deployed on any cloud or on-premises. - The shift to agentic AI is a dominant trend, with AI agents capable of reasoning, planning, and independent action expected to drive the next wave of automation. However, a significant challenge remains in scaling agentic AI projects beyond the pilot stage, with security, privacy, and compliance cited as primary barriers. - For ML Engineers, the operationalization of large language models is evolving from MLOps to the more specialized LLMOps. This new practice addresses the unique challenges of LLMs, such as managing prompts, optimizing inference-time performance, and monitoring generative behavior. - In terms of inference optimization, vLLM and TensorRT-LLM are leading engines for serving LLMs at scale. vLLM is favored for its flexibility and integration with Hugging Face, while TensorRT-LLM is chosen for maximum performance within the NVIDIA ecosystem. Stripe reportedly reduced its inference costs by 73% after migrating to vLLM. - Pricing models in enterprise AI are moving away from simple per-seat licenses towards more value-aligned strategies. These include usage-based pricing tied to metrics like API calls, and outcome-based models where fees are linked to business results such as cost savings or productivity gains. - The concept of "Shadow AI," where employees use external AI tools without IT oversight, presents both a risk and an opportunity. It signals a high demand for AI capabilities within the workforce but also creates challenges for data governance and security.
