Similarweb Q4 2025 Miss: AI Revenue, RPO & LLM Deals
Was Similarweb’s Q4 2025 revenue miss just timing? Breakdown of AI revenue (11%), RPO backlog, NRR trends, LLM pipeline deals, and cash flow strength.
## What the “miss” obscures: a growth engine that slipped a quarter, not a thesis that broke
Similarweb’s Q4-25 headline looks like a clean “miss,” but the company’s own disclosure frames the shortfall as largely **timing-driven**—specifically, **two large LLM data-training contracts that did not close in the quarter but remain active in the pipeline**. ["two large LLM data training contracts"](0001842731-26-000009|2) ["remain active in our pipeline"](0001842731-26-000009|2)
If those deals convert, the revenue recognition simply lands later—while the market often prices it as demand destruction in real time. ["sales cycles can take longer"](0001842731-26-000009|2)
## “AI” is already real revenue—still under-modeled
The most actionable tell is that management quantified AI traction: **generative AI data and solutions reached 11% of Q4 revenue (up from 8% earlier in the year)**. ["accounted for 11% of revenue"](0001842731-26-000009|2) ["up from 8%"](0001842731-26-000009|2)
They also disclosed that **GenAI Intelligence is approaching 200 customers and roughly $3 million of ARR**, which is meaningful because it implies early monetization with room to scale (and it’s new enough that many models may still treat it as “option value”). ["approaching 200 customers"](0001842731-26-000009|2) ["ARR of approximately $3 million"](0001842731-26-000009|2)
## Contracted visibility is improving even as sentiment weakens
Despite the quarterly noise, Similarweb reported **Remaining Performance Obligations (RPO) of $288.8 million, up 17% YoY**, and expects to recognize **~69% within the next 12 months**—a disclosure that effectively says “the backlog is building.” ["RPO totaled $288.8 million"](0001842731-26-000009|2) ["up 17% YoY"](0001842731-26-000009|2) ["recognize approximately 69%"](0001842731-26-000009|2)
In parallel, they highlighted a shift toward stickier revenue: **60% of ARR is now under multi-year subscription commitments (vs. 49% last year)**, which can dampen downside risk when budgets get reshuffled. ["60% of our ARR"](0001842731-26-000009|2) ["compared to 49%"](0001842731-26-000009|2)
## The “between-the-lines” risk: retention softness is the tax for the transition
The same disclosure package flags what’s likely holding the multiple back: **NRR fell to 98% overall (103% for $100K+ ARR customers)**, and management points to tougher comps plus the nature/timing of large contracts. ["overall NRR of 98%"](0001842731-26-000009|2) ["NRR of 103%"](0001842731-26-000009|2)
They also explicitly acknowledged **commercial execution shortfalls** and launched corrective actions—upskilling GTM and creating a **dedicated team selling to LLMs**—which reads like a company moving from “product innovation” to “operationalization.” ["commercial execution shortfalls"](0001842731-26-000009|2) ["dedicated team for selling to LLMs"](0001842731-26-000009|2) ["upskilling our go-to-market organization"](0001842731-26-000009|2)
## Balance sheet + profitability: the company is funding the pivot without leverage
In Q4-25, Similarweb reported **$72.4 million of cash and cash equivalents and no outstanding debt**, alongside **positive non-GAAP operating profit**—important because it suggests the AI pivot is not forcing a “growth-at-any-cost” reset. ["$72.4 million in cash"](0001842731-26-000009|2) ["no outstanding debt"](0001842731-26-000009|2) ["non-GAAP operating profit was $3.4 million"](0001842731-26-000009|2)
They also emphasized **nine consecutive quarters of positive free cash flow**, reinforcing that they can invest while staying financially disciplined. ["ninth consecutive quarter of positive free cash flow"](0001842731-26-000009|2)
## Why this can be “alpha” (if you believe the disclosure)
The potential edge is recognizing that the quarter’s miss coincides with (1) **AI monetization already showing up in the revenue mix**, and (2) **large, complex AI/LLM deals that inherently have lumpy timing**—a setup where near-term disappointment can mask a medium-term reacceleration if conversion rates normalize. ["Revenue from generative AI data and solutions"](0001842731-26-000009|2) ["timing of two large LLM data training contracts"](0001842731-26-000009|2) ["sales cycles can take longer"](0001842731-26-000009|2)
