Sail Research

What Sail Research does

Sail Research (doing business as "Sail") is an AI infrastructure company founded in 2026 by former Apple engineers Neil Movva and Samir Menon and headquartered in San Francisco, California. The company operates an inference platform that serves open-source AI models — including Kimi-K2.6, GLM-5.2, gpt-oss-120b, Nemotron 3 Super 120B, and Gemma 4 31B — through APIs that are compatible with the OpenAI and Anthropic interfaces, with a deliberate architectural emphasis on throughput and cost efficiency rather than latency. Sail discloses unit economics of 6–35x lower cost per query and up to 90% token savings versus incumbents, and reports processing trillions of tokens per week with a headcount of 101–250 employees.

Beyond the core inference API, Sail's product surface includes Sailboxes (persistent Linux virtual machines designed for long-horizon AI agents, featuring a live VM migration capability), Voyages (a telemetry and observability layer for agentic workloads), Completion Windows (tiered pricing that lets customers trade latency for cost), and an Enterprise Tier offering HIPAA, SOC 2, and GDPR compliance with SLAs and volume pricing. The company employs a dual go-to-market: a product-led growth motion with a $5/month free credit tier and self-serve developer onboarding, paired with enterprise contracts for regulated industries. Current named customers are startups including Quadrillion Labs, Parallel Web Systems, Detail.dev, and Jack and Jill; Sail is privately held, backed by Sequoia and Kleiner Perkins, and was valued at $450M in its combined $80M Seed and Series A.

Name

Sail Research

Legal name

Sail Research Co.

Website

https://sailresearch.com

Company type

Private

Founded year

2026

Operating status

Operating

Headcount range

101–250 employees

Short description

Sail Research is an AI infrastructure company founded in 2026 by former Apple engineers that runs a cost-optimized inference platform serving open-source models via OpenAI/Anthropic-compatible APIs, plus Sailboxes (persistent Linux VMs) and Voyages (observability) for long-horizon AI agents, targeting startups and regulated enterprises.

Ownership category

akta.pro rank

Product category

AI Inference Infrastructure

NAICS

Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (518), Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (5182), Computer Systems Design Services (541512), Other Computer Related Services (541519)

SIC

Services-Computer Programming, Data Processing, Etc. (7370), Services-Computer Integrated Systems Design (7373), Services-Computer Programming Services (7371)

akta.pro primary industry

Model Hosting, Serving & Inference Platforms (HDAAACAB)

akta.pro secondary industries

Model Serving, Inference & Deployment Platforms (APIs, Edge/On-Prem) (HDAEANAC), Model Compression & Optimization (Quantization, Distillation, Pruning) (HDAAACAL)

Keywords

• AI inference platform
• LLM inference API
• Agent infrastructure
• Cloud sandbox VMs
• Open-source model serving

Headquarters

HQ city

San Francisco

HQ country

United States

HQ region

North America

Offices1 record

Markets served

GTM type

B2B

Offering type

Software

Cost components

Technology or R&D, Infrastructure, Personnel, Marketing or Sales, Operations

Revenue model

1. Inference API - Token-based Usage: Usage-based token pricing for inference API. Pay-per-token model with completion window tiers determining price. Input tokens, cached input tokens, and output tokens billed at different rates per model. Free $5 monthly credits for self-serve tier.
2. Sailboxes - Active Resource Billing: Sailboxes charge only for CPU, memory, and disk actually used while running. No charge during sleep, pause, checkpoint, or cold-start. Creation has one-time charge based on size (S or M). Observed usage billing with ceiling-based sizing.
3. Enterprise Custom Contracts: Volume pricing, HIPAA-compliant region-locked datacenters, uptime and latency SLAs, early access to features, dedicated support. Billed monthly in arrears.

Pricing tiers

Go-to-market motion1 record

Distribution channels2 records

Marketing channels5 records

Core offering

Sail Research operates a cost-optimized AI inference platform serving leading open-source LLMs (GLM-5.2, Kimi-K2.6, gpt-oss-120b, Nemotron 3 Super 120B, Gemma 4 31B IT, DeepSeek V4 Pro) via OpenAI- and Anthropic-compatible APIs, paired with Sailboxes — persistent Linux VMs purpose-built for long-horizon autonomous AI agents that run for days or weeks with persistent state, auto-sleep, elastic scaling, and observed-usage billing. Completion Windows enable customers to trade latency for 30–80% token cost savings, and live VM migration infrastructure allows elastic compute use without agent awareness.

Product overview

Sail Research offers a unified AI infrastructure platform for long-horizon agents, combining a cost-optimized inference platform with purpose-built cloud VM environments. The core products are: (1) the Inference Platform, serving leading open-source models via OpenAI/Anthropic-compatible APIs with Completion Windows for cost-latency tradeoffs; and (2) Sailboxes, persistent cloud VMs for long-horizon agents with observed-usage billing, auto-sleep, and live migration. Supporting products include Voyages (telemetry and observability for agent runs), Enterprise tier (HIPAA-compliant, volume pricing, SLAs), the Agent Cost Calculator, and LoRA/RL fine-tuning support. The platform is designed so agents can run for days or weeks with persistent state and maximal cost efficiency.

Differentiator

Problem solved

Functional benefit

Brands

• Sailboxes: Cloud environments purpose-built for long-horizon AI agents, offering full VMs with persistent state, auto-sleep capabilities, and elastic resource scaling
• Sail Inference
• Voyages
• Voyages SDK

Products and services

• Inference Platform Core inference API platform serving leading open-source AI models (GLM-5.2 FP8, Kimi-K2.6, gpt-oss-120b, Nemotron 3 Super 120B, Gemma 4 31B IT, DeepSeek V4 Pro) via OpenAI- and Anthropic-compatible Responses, Chat Completions, and Messages endpoints, optimized for throughput over latency with Completion Windows for cost-latency tradeoffs and drop-in migration. Targets AI-native teams building long-horizon autonomous agents that need up to 10x lower cost per token than alternatives.
• Sailboxes Persistent Linux cloud VMs purpose-built for long-horizon AI agents, offering full-machine compute with persistent state, auto-sleep during idle periods, elastic resource scaling, checkpointing, forking, pause/resume, live migration across hosts, and observed-usage billing — customers pay only for active CPU, memory, and disk consumption rather than reserved capacity. Designed for agents that run for hours, days, or weeks.
• Voyages Telemetry and observability layer for long-running background agents on Sail, recording agent runs as dashboard traces with named agents, spans, events, model-call attribution, and Sailbox executions, with SDKs for Python, TypeScript, and Rust for instrumenting agent runs. Designed for debugging, auditing, and real-time visibility into agent trajectories.
• Agent Cost Calculator Standalone cost estimation tool that lets developers calculate the cost of running long-running agents on Sail versus alternative providers, using token volume, model selection, and completion window inputs to project spend and tune the speed-vs-cost tradeoff for a given workload.

Quantifiable outcome

• Up to 90% token cost savings vs alternatives
• +4 more outcomes

Named customers4 records

Segments3 records

Ideal customer profiles2 records

Technology focussed Yes

API detail

Has API

Yes

API docs

API detail

Core technology

AI maturity

App detail

Integration1 record

AI capability7 records

Feature5 records

Scale indicators7 records

Recent moves6 records

Expansion highlights6 records

Direct peers

• Replicate: Direct competitor offering a cloud API to run open-source ML models with usage-based pricing and developer-friendly deployment. Closely comparable business model to Sail's inference platform, including serving many OSS LLMs via a simple API.
• Fireworks AI: Direct peer providing fast and cost-efficient inference APIs for open-source LLMs with fine-tuning and deployment tooling. Targets the same AI-native developer customer with usage-based token pricing, making it a near-substitute for Sail's core inference offering.
• Modal Labs: Direct competitor offering developer-focused cloud compute for AI workloads, including serverless inference and sandboxed code execution for agents. Highly comparable product surface — both sell PLG API access for inference plus VM/sandbox environments for agent code execution.
• Anyscale: Direct peer offering AI compute platform built on Ray for distributed model serving, training, and inference at scale. Targets enterprise and AI-native teams with scalable LLM serving, overlapping materially with Sail's inference and RL rollout positioning.
• Together AI: Direct competitor offering an open-source model inference platform with usage-based pricing, fine-tuning, and serving for leading OSS LLMs — same product category, same customer base (AI-native startups building agents), and identical go-to-market. Notably, Together AI Chief Scientist Tri Dao is a Sail angel, signaling the companies view each other as direct rivals.

Broad incumbents

• Lambda Labs: Established GPU cloud provider that has expanded into managed inference services for open-source LLMs. Broader portfolio (GPUs, clusters, training) but overlapping inference offering makes it a relevant incumbent competitor for AI-native workloads.
• CoreWeave: Large-scale GPU cloud provider serving AI training and inference workloads at hyperscale. While not as developer-productized as Sail, CoreWeave is the underlying compute layer for much of the inference ecosystem and a key competitor for inference spend.
• Hugging Face: Open-source model hub that has expanded into serverless inference (Inference Endpoints, Inference API) for the same OSS models Sail serves. Comparable on model serving; broader as a community/platform layer.

Emerging players

• E2B: Emerging sandbox provider offering secure cloud sandboxes specifically for AI agents and code execution. Closest direct comparator to Sailboxes — same customer segment (AI agent developers) and overlapping use cases for persistent, fault-tolerant execution environments.
• Cerebrium: Emerging serverless AI infrastructure platform offering GPU-accelerated inference and model deployment. Smaller scale than Together/Fireworks but overlapping on serving OSS LLMs for AI-native developer customers.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat5 records

Key risks7 records

Key highlights7 records

Customer concentration

• Xx.com/sailresearchco
• GitHubgithub.com/sailresearchco
• LinkedInlinkedin.com/company/sailresearch

Compliance3 records

Revenue estimate

Valuation estimate

Number of profiles

Profiles2 records

Funding overview

Funding rounds2 records

Investors8 records

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Sail Research