AI labs now own 69% of enterprise AI spending, per Menlo report
The same labs selling you API access are now taking a 69% share of enterprise AI deployment spending, according to Menlo Ventures’ 2026 enterprise AI report. That’s a number worth sitting with if your product touches enterprise budgets.
Where the Enterprise Budget Is Going
Frontier labs now capture 69% of enterprise AI deployment spending, per Menlo’s survey. The counterweight: agents appear to be growing total spend rather than just reshuffling it. New budgets funded 55% of the additional AI spend Menlo tracked, which the firm reads as companies paying for work their existing software couldn’t handle.
Startups still take 61% of application-layer spending once the labs’ own apps are excluded. The app layer is competitive, but it’s not closed.
Where App-Layer Moats Actually Come From
According to Euclid VC’s analysis, the value in AI applications lands with products that own data the labs can’t replicate or purchase. Coding tools took off partly because outcomes are automatically gradable. The harder moat is outcome data that only exists when someone does the job: whether a loan defaulted, whether a claim got approved. The practical takeaway from Euclid’s piece is to list every outcome your product touches and capture job byproducts as structured data from day one.
OpenAI shipped a legal tool one week after Harvey raised at a $15.5 billion valuation, which illustrates how quickly labs can move into vertical territory once a category is visible.
The SaaS Signal Worth Watching
Stripe benchmark data covering roughly 181,000 subscription businesses shows 29 of 48 cohorts above $100,000 ARR were shrinking in October. In September, only four cohorts were shrinking. Customer churn went down over the same period, pointing toward weaker new sales rather than existing customers leaving. Only companies above $5 million ARR were still growing in October.
The writer behind this analysis suspects AI-built competitors are an easy explanation, though the data doesn’t confirm cause. What it does confirm: new logo acquisition got harder for smaller SaaS businesses this fall.
The Intelligence Economy Ownership Question
EM Capital flags a structural issue that’s easy to defer and hard to fix later. An AI trained on your best salesperson’s methods can keep using those methods after they leave. Who owns that captured expertise? If model providers control the workflow layer, they may end up owning what your company’s AI has learned. The practical advice: keep workflows portable between providers so switching models doesn’t erase accumulated institutional knowledge.
I’m founder of BizStack at Cagri Sarigoz LLC and a passionate advocate for entrepreneurs.
With over 14 years in tech, marketing, and AI, including my role as Head of SEO at CitizenShipper and co-founder of TaleBot at Intale AI, I’m dedicated to sharing genuine, useful product insights and tips.
At BizStack, I aim to cut through the digital noise to provide clear, actionable advice.
And more than all else, I’m a father to a (always) little girl and a husband.
Contact me at [email protected] for assistance.
