Bret Taylor’s Sierra closed a $950M round led by Tiger Global and GV1 on May 4, 2026, pushing its post-money valuation past $15 billion,1 up from $10 billion eight months earlier when Greenoaks Capital led a $350M round.1 Total capital raised now exceeds $1 billion.1 The company claims 40%+ of the Fortune 50 as customers,1 with agents handling billions of interactions across mortgage refinancing, insurance claims, and retail returns.
The Valuation Math
Sierra hit $100M ARR in late November 2025,2 roughly 21 months from founding, then crossed $150M ARR by early February 2026 according to Sierra’s “Year Two in Review”.3 A $50M ARR increase in a single quarter.3 At $150M ARR3 and a $15B+ post-money,1 the implied revenue multiple is approximately 100x, the same ratio as the prior $10B round at $100M ARR.1 The multiple has held as ARR scaled, which is not typical.
Locking In the Fortune 50
The 40% Fortune 50 figure1 carries weight in enterprise procurement conversations. Sierra’s “Year Two in Review” puts aggregate reach at 95% of US shoppers,3 50% of families in healthcare, 70% of the fintech value chain, and 25% of European banking.3 Named accounts include ADT, Cigna, DIRECTV, Gap, Rocket Mortgage, SiriusXM, Redfin, Wayfair, Nordstrom, and Nubank.3
For enterprise CX and operations teams, the build-vs-buy decision has shifted. Sierra has two years of production history on Fortune 50 workloads, an outcome-based SLA structure, and reference customers across every vertical a procurement committee will ask about. Building an equivalent internal platform requires at minimum 18 to 24 months, a team with production-LLM-ops depth, and full internal ownership of every incident. The vendor’s reference customer list is now the strongest argument against building.
Concentration cuts both directions. A platform-level failure at Sierra would be visible across a material fraction of US consumer finance, retail, and insurance simultaneously. That is Sierra’s liability, but it is also what keeps existing customers in place: workflows embedded this deeply do not migrate on short timelines.
In April 2026, Sierra launched Ghostwriter,4 which builds agents from natural-language descriptions. As the AI agent marketplace matures, generating agents on demand is a different category of entrenchment than running them: the customer’s institutional knowledge starts to live inside Sierra’s abstractions.
The Governance Tension
Taylor co-founded Sierra in early 2024 alongside Clay Bavor, who spent 18 years at Google leading Gmail and Drive. Taylor is also chairman of OpenAI’s board and previously served as Salesforce co-CEO.2
The Labs Go Direct
Anthropic and OpenAI both launched joint ventures for enterprise AI services the same week as Sierra’s announcement.5 Salesforce Agentforce, Decagon, Intercom, and Cresta occupy parts of the adjacent stack, but vertical integration from the model layer is structurally different from competition within the integration layer. The labs have the models, the API relationships, and now direct enterprise contracts to assemble end-to-end offerings.
Sierra’s position is the standard one for any infrastructure vendor facing vertical integration from a supplier. Production history buys time; so does customer entrenchment that takes longer to unwind than a procurement cycle. The labs still need to build the operational depth Sierra has been accumulating since February 2024: the incident response procedures, the integration tooling, the compliance infrastructure that comes from running Fortune 50 workloads over two years.
The $1 billion in capital1 is designed to widen that operational gap before the labs’ direct offerings mature. The 100x revenue multiple is Tiger Global and GV betting it will.
Frequently Asked Questions
How does Sierra’s per-resolution pricing compare to Salesforce Agentforce or Decagon?
Salesforce Agentforce and Decagon largely use per-seat or per-interaction subscription billing. Sierra’s per-completion model aligns cost with agent output, but creates a non-obvious dynamic: as agents improve and resolve issues on fewer attempts, billable events per customer can decline. Revenue growth then depends on ticket volume scaling faster than accuracy, unless Sierra introduces tiered resolution pricing.
What sectors does Sierra’s Fortune 50 footprint not yet cover?
Named customers and disclosed use cases cluster in consumer-facing verticals — retail, insurance, banking, telecom, and mortgage lending. There is no public presence in manufacturing, defense, logistics, or government, sectors where agent infrastructure would trigger regulatory frameworks (ITAR, FedRAMP, sector-specific data residency) Sierra has not addressed.
Who else is on Sierra’s cap table beyond Tiger Global and GV?
The investor syndicate also includes Sequoia, Benchmark, ICONIQ Capital, and Thrive Capital, alongside the disclosed Greenoaks-led prior round. The cap table spans growth, multi-stage, and crossover funds, which is consistent with Sierra having leverage to set terms rather than consolidating around a single lead — a factor in the 100x multiple holding between the $10B and $15B rounds.
What governance obligations does Taylor’s OpenAI chairmanship create if the companies compete for the same contract?
A board chairman owes fiduciary duties — loyalty and care — to OpenAI’s shareholders. If OpenAI’s enterprise JV bids against Sierra for a Fortune 50 agent contract, Taylor would need to recuse himself from OpenAI’s deliberations on that deal and from any Sierra-specific information sharing about model access terms or the enterprise roadmap. Neither company has disclosed whether such recusal protocols exist.