There’s a lot of chatter in the news these days about the AI bubble. Most of it is because of the circular nature of the deals among the foundational model providers like OpenAI and Anthropic, and cloud providers (Microsoft, Amazon) and NVIDIA.
OpenAI recently published a report called “The state of enterprise AI” where they said:
The picture that emerges is clear: enterprise AI adoption is accelerating not just in breadth, but in depth. It is reshaping how people work, how teams collaborate, and how organizations build and deliver products.
AI use in enterprises is both scaling and maturing: activity is up eight-fold in weekly messages, with workers sending 30% more, and structured workflows rising 19x. More advanced reasoning is being integrated— with token usage up 320x—signaling a shift from quick questions to deeper, repeatable work across both breadth and depth.
Investors at Menlo Ventures are also seeing positive signs in their data, especially when it comes to the tech space outside the frontier labs:
The concerns aren’t unfounded given the magnitude of the numbers being thrown around. But the demand side tells a different story: Our latest market data shows broad adoption, real revenue, and productivity gains at scale, signaling a boom versus a bubble.
AI has been hyped in the enterprise for the last three years. From deploying quickly-built chatbots, to outfitting those bots with RAG search, and more recently, to trying to shift towards agentic AI. What Menlo Venture’s report “The State of Generative AI in the Enterprise” says is that companies are moving away from rolling their own AI solutions internally, to buying.
In 2024, [confidence that teams could handle everything in-house] still showed in the data: 47% of AI solutions were built internally, 53% purchased. Today, 76% of AI use cases are purchased rather than built internally. Despite continued strong investments in internal builds, ready-made AI solutions are reaching production more quickly and demonstrating immediate value while enterprise tech stacks continue to mature.
Also startups offering AI solutions are winning the wallet share:
At the AI application layer, startups have pulled decisively ahead. This year, according to our data, they captured nearly $2 in revenue for every $1 earned by incumbents—63% of the market, up from 36% last year when enterprises still held the lead.
On paper, this shouldn’t be happening. Incumbents have entrenched distribution, data moats, deep enterprise relationships, scaled sales teams, and massive balance sheets. Yet, in practice, AI-native startups are out-executing much larger competitors across some of the fastest-growing app categories.
How? They cite three reasons:
- Product and engineering: Startups win the coding category because they ship faster and stay model‑agnostic, which let Cursor beat Copilot on repo context, multi‑file edits, diff approvals, and natural language commands—and that momentum pulled it into the enterprise.
- Sales: Teams choose Clay and Actively because they own the off‑CRM work—research, personalization, and enrichment—and become the interface reps actually use, with a clear path to replacing the system of record.
- Finance and operations: Accuracy requirements stall incumbents, creating space for Rillet, Campfire, and Numeric to build AI‑first ERPs with real‑time automation and win downmarket where speed matters.
There’s a lot more in the report, so it’s worth a full read.
