Sierra AI Valuation 2026: $15.8B, $950M Series E, and ~$200M ARR

Sierra AI is valued at $15.8B after a $950M Series E in May 2026 — up from a $1B valuation at launch just 27 months earlier — on roughly $200M in annualized revenue from enterprises paying per resolved customer-service interaction rather than per seat.

Bret Taylor left the OpenAI board chairmanship — he still holds it alongside running Sierra — and his own company just became one of the fastest-compounding valuations in enterprise software history. The pace matters more than the headline number: four rounds, four step-changes in valuation, in barely over two years, on a bet that AI agents replace the seat-based software model entirely.

Sierra AI Valuation 2026: From $1B to $15.8B in 27 Months

Sierra's valuation is $15.8B as of May 2026, set by a $950M Series E led by Tiger Global and GV, with Benchmark, Sequoia Capital, and Greenoaks Capital also participating. That is up from $10B eight months earlier and from $1B when Bret Taylor and Clay Bavor launched the company in February 2024.

The step-change pattern is the story. Fortune reported the February 2024 launch: a $110M round from Sequoia and Benchmark at close to a $1B valuation, before Sierra had a paying customer. By October 2024, CNBC reported a $175M round led by Greenoaks Capital at $4.5B. Less than a year after that, in September 2025, the valuation reached $10B on a $350M round. Compare Sierra against every other AI company repricing this fast on the AI Valuations dashboard.

The Revenue Behind the Number: ~$200M ARR on Outcome-Based Pricing

Sierra crossed $100M in annualized recurring revenue in November 2025, about 21 months after launch — a pace TechCrunch called out as among the fastest for enterprise software. By the May 2026 raise, reporting around the round put run-rate near $200M, though Sierra has not disclosed a newer figure since. At $15.8B on roughly $200M ARR, that's close to 79x revenue — a multiple that only holds up if growth keeps compounding at 2025's pace.

The mechanism behind that revenue is unusual: Sierra doesn't sell seats or API credits. It charges roughly $1.50 for every customer interaction its AI agent actually resolves, an outcome-based model Taylor has argued repeatedly is the correct pricing unit once software does the work itself rather than assisting a human who does. We break down the full mechanics — pricing tiers, margin structure, and what "resolved" means contractually — in how Sierra AI makes money. Disclosed customers span ADT, Sonos, WeightWatchers, and Casper — physical-product and subscription businesses with high support-ticket volume, which is exactly where outcome-based pricing has the clearest ROI story to sell into a procurement committee.

Sierra vs Decagon, Agentforce, and Intercom Fin on Price

Every credible competitor in enterprise AI customer agents has converged on some version of outcome- or resolution-based pricing, which makes the per-interaction rate one of the few apples-to-apples numbers across the category.

Decagon and Salesforce Agentforce price custom enterprise contracts rather than publish a flat per-resolution rate, so they are not shown on this chart.

Decagon is the closest AI-native challenger by growth trajectory, and Salesforce Agentforce and Zendesk AI are the incumbent-platform answers — Sierra's pitch against both is that it's vendor-neutral rather than bundled into a CRM or helpdesk you may not already run. Salesforce closed its acquisition of Intercom's Fin on September 10, 2026, folding a direct competitor into the same incumbent camp Sierra is positioned against.

The Bull Case: Pricing Software Like Labor, Not Seats

If Sierra is right that AI agents should be priced on completed work rather than logins, it has built the pricing model the rest of the category will eventually copy — and being first with the enterprise logos and the renewal data to prove the model works is a real moat. Outcome-based pricing also aligns Sierra's revenue directly with the value it delivers: as its agents resolve a larger share of a customer's support volume, revenue scales with them, without a separate sales motion to expand seats.

Where I Could Be Wrong

A 79x-revenue multiple prices in years of continued hypergrowth, and outcome-based pricing has a structural ceiling problem: once an enterprise's AI agent resolves nearly all of its support volume, that customer's spend flattens — there's no more "seats to add." That's the opposite of traditional SaaS expansion revenue, and it means Sierra needs a constant stream of new logo growth rather than compounding existing accounts as heavily. The other risk is competitive: Salesforce folding Fin into Agentforce and Zendesk running its own resolution-priced agent both show the incumbents can copy the pricing model without needing Sierra's valuation multiple to make the economics work for them.