Sierra
B2B SaaS with usage- and outcome-based pricing (pay per successful resolution / per conversation), multi-year enterprise contracts and high-touch implementation; model-agnostic orchestration layer on top of third-party LLMs
Four disclosed primary rounds; total ~$1.585B raised across the C/D/E (plus the ~$110M stealth-emergence round). ~16x valuation step from $1B (Feb 2024) to $15.8B (May 2026) in ~27 months, tracking the ARR ramp from ~$0 to ~$200M. The Feb-2024 ~$1B mark is press-reported and approximate; the Oct-2024, Sep-2025, and May-2026 marks are strongly sourced. No secondary or SPAC events.
Earnings, margins, COGS & capex
Sierra is a hyper-growth, pre-profitability private company. It has raised ~$1.585B across four rounds and scaled ARR from $26M (end-2024) to ~$200M (May 2026) - among the fastest enterprise-software ramps on record. Monetization is outcome/usage-based (pay per resolved conversation), which aligns price to delivered value but ties COGS to third-party LLM inference. Profitability, margins, and cash burn are not disclosed.
Income statement — where each revenue dollar goes
% of revenueOf every $1 of revenue, ~30¢ is cost of goods and ~70¢ operating expense, leaving ~0¢ of operating profit.
Revenue trend
Margins
structurally pressured by LLM inference pass-through vs pure-SaaS peers
assumed negative (growth-stage burn)
COGS structure
Primary COGS is third-party foundation-model inference (Sierra is model-agnostic, routing to OpenAI / Anthropic / Google models) plus cloud compute and high-touch professional-services/implementation labor. Outcome-based pricing means revenue scales with resolutions but so does inference cost - the gross-margin structure is the key undisclosed unknown.
Capex
Asset-light; no disclosed owned infrastructure. Compute is rented from hyperscalers and model providers, so 'capex' is effectively opex (inference + cloud). Not disclosed.
Latest earnings
n/a
No public guidance. Company messaging targets continued Fortune-50 expansion and multi-use-case agent deployment.
- ARR (May 2026)
- ~$200M
- Fortune 50 penetration
- 40%+
- Total raised
- ~$1.585B
- Post-money valuation
- $15.8B (May 2026)
- Implied ARR multiple
- ~79x trailing ARR
Growth drivers
- Land-and-expand across Fortune 50 (already 40%+ penetration) - adding new agents/use-cases per account
- Expansion into voice/telephony agents, extending TAM beyond chat into contact-center automation
- Outcome-based pricing that lets buyers justify spend against measurable deflection (Sierra cites automation/resolution rates up to ~90% in referenced deployments)
- Expansion beyond customer support into broader enterprise 'agent' workflows (the stated Series E thesis)
- Founder-led enterprise distribution + brand (Bret Taylor is also OpenAI board chair), easing top-of-funnel access to large buyers
Bull & bear
Sierra is the emerging category leader in the enterprise-agent land-grab, with a rare combination of elite founders, top-decile ARR velocity, and Fortune-50 distribution - positioned to become the standard agent platform for large enterprises.
- ~$200M ARR in ~2 years with 40%+ Fortune-50 penetration validates strong product-market fit at the highest end of the market
- Outcome-based pricing is a durable moat: buyers pay for resolutions, making ROI self-evident and expanding usage as agents improve
- Bret Taylor's dual role (Sierra CEO + OpenAI board chair) plus Clay Bavor gives unmatched access to both enterprise buyers and frontier models
- $1.585B raised and marquee backers (GV, Tiger Global, Sequoia, Benchmark, Greenoaks) fund a multi-year, category-owning expansion
- Expansion into voice/telephony and beyond support multiplies the addressable spend per customer
At ~79x ARR with undisclosed and likely thin gross margins, Sierra is priced for flawless execution while facing distribution-advantaged incumbents (Salesforce, Microsoft) and up-stack encroachment from the very model providers it depends on.
- ~79x trailing-ARR valuation ($15.8B / ~$200M) demands years of >2x growth with no misstep
- Usage-based revenue on pass-through LLM inference risks structurally lower gross margins than the SaaS comps its multiple implies
- Salesforce Agentforce and Microsoft can bundle agents into existing CRM/productivity seats with distribution Sierra cannot match
- Foundation-model labs (OpenAI/Anthropic/Google) adding native agent tooling could commoditize the orchestration layer
- High-touch implementation makes Sierra more services-heavy and slower-to-change than lighter rivals (Decagon), capping scalability
- No disclosed retention, margin, or burn metrics - the market is trusting narrative over audited unit economics
What it is worth
ARR multiple vs private/late-stage AI-application peers (no public price; DCF not meaningful pre-profit)
Margin compression from inference COGS + incumbent (Salesforce/Microsoft) distribution pressure slows growth; AI-multiple reset takes the implied value toward a 30-40x ARR range (~$6-8B), a material down-round risk.
Growth decelerates toward 1.5-2x with mid-tier gross margins; $15.8B roughly fair, next round flat-to-up modestly as it grows into the multiple.
Sustained >2x ARR growth to $400M+ within ~12-18 months with software-like (70%+) gross margins and Fortune-50 expansion supports a $20B+ valuation and eventual IPO/strategic exit at a premium.
Series E post-money is $15.8B on ~$200M ARR = ~79x trailing ARR, an extreme premium reflecting category leadership, founder pedigree, and ~5x growth. Late-stage AI application-layer peers trade roughly 20-50x forward ARR; Sierra sits well above, justified only if growth stays >2x and gross margins prove software-like. Undisclosed margins/NRR are the key valuation risk.
SWOT
Strengths
- Category-defining brand in enterprise AI agents with elite, credible founders (Bret Taylor, Clay Bavor)
- Fastest-in-class ARR ramp ($0->~$200M in ~2 years) and 40%+ Fortune-50 logo penetration
- Outcome-based pricing aligns cost to customer-realized value, a differentiated commercial wedge
- Deep, well-capitalized balance sheet (~$1.585B raised) funding multi-year enterprise land-and-expand
- Model-agnostic architecture insulates it from any single foundation-model provider
Weaknesses
- Undisclosed and likely compressed gross margins - inference COGS scale with usage-based revenue
- High-touch implementation creates services drag and slows customer-driven changes (vendor-dependency complaint vs Decagon)
- Single dominant use-case (customer service/CX) concentration; expansion into other agent workflows unproven
- Extreme ~79x ARR valuation leaves little room for execution error
- Net revenue retention, churn, and unit economics are undisclosed
Opportunities
- Voice/contact-center automation as a large incremental TAM beyond chat
- Broadening from support into sales, operations, and back-office enterprise agents
- International expansion and mid-market down-market motion beyond the Fortune 50
- Becoming the default agent orchestration layer as enterprises standardize on one vendor
- Attach to the secular shift of contact-center + BPO labor spend toward AI deflection
Threats
- Salesforce Agentforce and Microsoft embedding agents natively into CRM/Office with distribution advantage
- Well-funded pure-play rivals (Decagon, Cresta, Ada, Parloa) competing on speed/control and price
- Foundation-model providers (OpenAI, Anthropic, Google) moving up-stack into turnkey agent products, commoditizing the orchestration layer
- Margin squeeze if inference costs stay high while buyers demand lower per-resolution pricing
- Valuation/funding-market reset for AI application-layer startups
Moats, dependencies & bottlenecks
Moats
Aligns price to value and raises switching friction once agents are tuned to a customer's brand and workflows, but replicable by rivals
40%+ Fortune-50 penetration and founder credibility create trust that is slow for challengers to earn
Bret Taylor/Clay Bavor pedigree + $1.585B war chest; talent moats erode as the field matures
Volume improves agent quality, but data is largely per-customer and siloed, limiting network effects
Insulates from single-provider risk but is the exact layer foundation-model labs may absorb
Dependencies
Critical technology + COGS input Agent quality and gross margin both hinge on third-party LLM capability and inference pricing; also a potential future competitor
Inference/hosting is rented; cost and availability outside Sierra's control
Revenue concentration Growth thesis leans on expanding within a concentrated set of very large accounts
Pre-profit; a funding-market reset would raise the cost of the growth strategy
Advantages
- First-mover brand and category leadership in enterprise AI agents
- Elite founder team with OpenAI-board-level access to frontier models
- Top-decile ARR growth and 40%+ Fortune-50 penetration
- Outcome-based pricing that de-risks buyer ROI
- Deep capitalization (~$1.585B) enabling long land-and-expand horizon
Weaknesses
- Undisclosed, likely inference-compressed gross margins
- Services-heavy, slower-to-change delivery model
- Single-use-case (CX) concentration to date
- Extreme valuation multiple with thin public financial disclosure
- Structural exposure to model-provider up-stack encroachment
Bottlenecks
- Gross-margin ceiling set by third-party inference costs on usage-priced revenue
- Implementation/professional-services capacity limits deployment velocity (high-touch model)
- Scarce enterprise-AI deployment talent to staff Fortune-50 rollouts
- Dependence on foundation-model roadmap and pricing it does not control
Top signals & trends
Top signals
Roughly tripled valuation from $4.5B (Oct 2024) in ~19 months; strong investor conviction
Among the fastest enterprise-software ramps on record (crossed $100M in 7 quarters, per Sierra)
Extends reach into higher-TAM contact-center telephony beyond text chat
Prices in years of flawless growth; sensitive to any AI-multiple compression
Distribution-advantaged incumbents intensifying competition
Narrative-led valuation with limited audited unit-economics visibility
Trends
Core secular tailwind expanding Sierra's addressable labor-replacement spend
Extends TAM into telephony/contact-center automation
Threatens to commoditize the orchestration layer Sierra occupies
Distribution-led competition pressures standalone platforms
Could relieve COGS/margin pressure but also lowers barriers for competitors
Ecosystem & competitor graph
Suppliers feed the company; customers pull from it. Line thickness shows the strength of each tie (supply-chain dependency, customer earnings contribution). Hover to isolate a tie.
Private - foundation-model provider (agent reasoning); Bret Taylor chairs its board
Private - foundation-model provider (Claude models) in a model-agnostic stack
Gemini models + Google Cloud compute; also an investor via GV
Azure cloud/model access (also a competitor)
AWS cloud infrastructure for hosting/inference
GPU compute underlying the inference the platform relies on
Publicly cited Sierra customer using branded AI agents for customer service
Fintech customer deploying Sierra agents
Consumer-electronics customer for support automation
Publicly referenced Sierra deployment
Publicly cited Sierra customer (per company profiles)
Company-disclosed penetration; most individual logos undisclosed
CRM-embedded AI agents with massive install-base distribution into Service Cloud accounts
Bundles agents into Dynamics + M365; enterprise reach and pricing leverage
Now Assist / AI agents embedded in enterprise service workflows
CXone contact-center platform adding native AI/virtual-agent automation
Communications/CX platform expanding into AI agent tooling (voice/messaging)
Contact Center AI platform; also a Sierra model supplier and investor via GV
Cloud-native contact-center + agent building blocks
Private pure-play rival; competes on speed/control and CX-team workflow ownership
Private; contact-center AI for agent assist and automation
Private; automated customer-service resolution platform
Private; European voice-first AI contact-center agent platform
Private; AI support agent with strong SMB/mid-market traction