Poolside
- Company typePrivate
- Founded2023
- HeadquartersSan Francisco, United States
- Headcount251–500
- GTM typeB2B
- OfferingSoftware
Poolside firmographics
Firmographics- Name
- Poolside
- Legal name
- Poolside
- Website
- http://www.poolside.ai
- Company type
- Private
- Founded year
- 2023
- Operating status
- Operating
- Headcount range
- 251–500 employees
- Ownership category
- akta.pro rank
Poolside industry classification
Industry- Product category
- AI Foundation Models for Software Engineering
- NAICS
- Custom Computer Programming Services (541511), Computer Systems Design Services (541512), Computer Systems Design and Related Services (54151)
- SIC
- Services-Prepackaged Software (7372), Services-Computer Programming Services (7371)
- akta.pro primary industry
- Foundation Model Developers (LLM/Multimodal Model Labs) (HDAAACAA)
- akta.pro secondary industries
- AI Application Enablement Platforms (Copilot/Agent Frameworks, SDKs) (HDAEANAJ), Model Development & Training Platforms (AutoML, Notebooks, Feature Stores) (HDAEANAB)
Keywords
Where Poolside is headquartered
LocationHeadquarters
- HQ city
- San Francisco
- HQ country
- United States
- HQ region
- North America
Offices5 records
Markets served
Poolside business model
Business model- GTM type
- B2B
- Offering type
- Software
- Cost components
- Technology or R&D, Infrastructure, Personnel, Marketing or Sales, Operations
Revenue model
- Enterprise Foundation Model Licenses: Multi-year enterprise contracts licensing Laguna models for on-prem, VPC, or air-gapped deployment, typically scoped per-customer with custom training and fine-tuning. Priced based on model size, deployment footprint, and customization depth.
- API / Usage-Based Model Access: Consumption-based revenue from platform.poolside.ai, OpenRouter, Hugging Face, and AWS Bedrock endpoints. Used as a self-serve funnel and as overage pricing on top of enterprise contracts.
- Forward Deployed Research Engineer (FDRE) Professional Services: Poolside embeds Forward Deployed Research Engineers alongside customer engineering teams to integrate models, build custom RL pipelines, and operationalize the agent harness. Billed as part of enterprise deals and as standalone services.
- Custom Model Training Engagements: Fee-for-service engagements where Poolside trains customer-specific foundation models on private codebases using its Model Factory and RLCE pipeline. Land-and-expand into multi-year licensing.
Pricing tiers
| Model | Billing | Price |
|---|---|---|
| Freemium | Pay-as-you-go | Laguna XS.2 — open-weights preview, free under Apache 2.0 |
| Other | Multi-year contract | Laguna M.1 — proprietary model, weights available to researchers |
| Subscription | Multi-year contract | Enterprise — on-prem / VPC foundation model deployment |
| Usage-based | Pay-as-you-go | API access via platform.poolside.ai / AWS Bedrock / OpenRouter |
Go-to-market motion4 records
Distribution channels6 records
Marketing channels7 records
Poolside product offering
Product offeringCore offering
Poolside builds proprietary foundation models (the Laguna family, including Laguna M.1 and Laguna XS.2) trained specifically for software engineering tasks such as code generation, refactoring, code review, and test generation. It packages these models behind an enterprise AI platform (Poolside Platform) with an agent harness, IDE/CLI surfaces, MCP-based tool integrations, and the option to deploy fully on-premises or inside a customer's VPC for regulated buyers. The company also embeds Forward Deployed Research Engineers with customer teams to deliver outcomes, not just model access.
Product overview
Poolside operates a platform-plus-modules architecture for AI-assisted software engineering. The core foundation models Laguna M.1 (proprietary, 225B params / 23B activated) and Laguna XS.2 (open-weight, 33B params / 3B activated, 256K context) are accessed through the Poolside Platform, an enterprise AI platform introduced in May 2026 that bundles the pool agent harness (terminal/CLI coding agent), the Poolside Console admin dashboard, and the Shimmer cloud development experience layer. Agents connect to enterprise tools via MCP (Model Context Protocol) servers and communicate with IDE clients (VS Code, Visual Studio, Zed, JetBrains, CLI) via ACP (Agent Client Protocol), with Bridge (acquired from Fern Labs, November 2025) providing multi-agent orchestration. Underpinning all of this is The Model Factory, Poolside's internal training platform used to build the Laguna models.
Differentiator
Problem solved
Functional benefit
Brands
- Laguna: Poolside's AI model family for software development, including the Laguna M.1 and Laguna-XS.2 models.
- Shimmer
- pool
Products and services
- Laguna M.1
- Laguna XS.2
- Poolside Platform
- pool
- Shimmer
- Poolside Console
- Bridge
- The Model Factory
Quantifiable outcome
- ~30% engineering productivity uplift (stated by co-founder messaging)
- +4 more outcomes
Companies that use Poolside
Customer profileNamed customers2 records
Segments4 records
Ideal customer profiles4 records
Poolside technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- Yes
- API docs
- API detail
Core technology
AI maturity
App detail
Integration18 records
AI capability10 records
Feature10 records
Poolside partnerships and signals
Strategic signalPartnerships
Twelve partnerships are on record, tiered flagship, core and minor.
- Amazon Web Services (AWS)flagshipStrategic infrastructure and go-to-market partner. Poolside models are first-party available on AWS Bedrock and EC2, with deployment options running on AWS Trainium chips. Co-marketing at AWS re:Invent and joint customer engagements for regulated enterprise and government workloads.
- Dell TechnologiesflagshipPoolside is an AI-provider partner in the Dell AI Factory with NVIDIA, enabling Dell customers to procure Poolside models pre-integrated on Dell on-prem hardware for secure, on-premises AI deployments.
- NVIDIAflagshipNVIDIA is both a hardware partner (Poolside trains on 6,144 NVIDIA Hopper GPUs and is part of the NVIDIA Inception program) and a financial investor in Poolside's Series B. Co-markets the Dell AI Factory with NVIDIA.
- AtoscoreAtos and Poolside are joining forces to power Atos' Agentic Studios offering, integrating Poolside foundation models into Atos's enterprise agent platform.
- RedpandacoreStrategic partnership to provide an agentic data plane for Poolside's agentic AI workloads, including the Redpanda Connect integration.
- OpenRoutercoreOpenRouter serves as a distribution partner for Poolside's foundation models, providing a developer self-serve API channel.
- Hugging FacecoreHugging Face hosts Poolside's open-weights model releases (Laguna XS.2 and the research-tier M.1) and serves as a top-of-funnel distribution channel.
- EleutherAIcoreResearch collaboration with EleutherAI to co-develop and release the Common Pile v0.1 open-source training dataset.
- Fern LabscoreAcquired by Poolside in November 2025, bringing additional engineering capacity and establishing Poolside's UK presence in London.
- CoreWeave (former)minorPoolside and CoreWeave jointly announced Project Horizon, a ~2 GW West Texas data-center build, which was subsequently reported as having fallen through.
- GoogleminorReported discussions regarding reviving the Texas data-center project, in the context of a broader Google AI-infrastructure push.
- Atlassian (board seat)minorPoolside co-founder Jason Warner joined Atlassian's board, providing adjacency to the developer-tools ecosystem.
Scale indicators18 records
Recent moves7 records
Expansion highlights8 records
Poolside competitors and assessment
Company assessmentDirect peers
- Anysphere (Cursor): Anysphere builds Cursor, a leading AI code editor using frontier LLMs. It is the most direct competitor to Poolside's agentic coding workflow but is positioned as a developer-first product rather than a foundation-model-and-deployment platform.
- Cognition AI (Devin): Cognition builds Devin, an autonomous software engineering agent that plans and executes multi-step coding tasks. It is a direct competitor in agentic software engineering with similar enterprise and self-serve GTM.
- Codeium (Windsurf): Codeium offers AI code completion, chat, and the Windsurf agentic IDE. It is a direct competitor for AI-assisted coding workflows targeting both individual developers and enterprise buyers.
- Mistral AI: Mistral builds open-weight and proprietary foundation models including Codestral, a code-focused LLM. It is a direct competitor for code-generation models with comparable open-weight distribution and self-serve API access.
- Sourcegraph (Amp / Cody): Sourcegraph offers Cody and Amp, AI coding assistants that work across enterprise codebases. It targets the same Global 2000 enterprise software engineering buyer with an emphasis on large, proprietary codebases.
- Tabnine: Tabnine provides AI coding assistance with on-prem and VPC deployment options. It is a direct competitor for enterprises that need private deployments of code-generating AI.
- Magic AI: Magic is building foundation models and agents aimed at automating software engineering, with a long-context focus. It is a direct competitor in the long-horizon coding-agent category with a similar foundation-model-first approach.
- Augment Code: Augment Code builds AI coding agents targeting professional software engineering teams with context-aware code generation. It is a direct competitor in agentic coding for enterprise engineering organizations.
Broad incumbents
- GitHub Copilot: GitHub Copilot (Microsoft) is the incumbent in AI-assisted coding with the largest enterprise install base. It is a broad incumbent offering overlapping code-completion and chat capabilities rather than a deployable foundation-model stack.
- Anthropic (Claude Code): Anthropic's Claude (including Claude Code) is a frontier general-purpose LLM widely used for software engineering tasks. It is a broad incumbent offering overlapping code-generation and reasoning capabilities, though not specializing in code-only foundation models.
Market position
Strengths5 records
Weaknesses5 records
Competitive moat7 records
Key risks6 records
Key highlights7 records
Customer concentration
Poolside social profiles
Digital presencePoolside compliance and trust
Trust signalCompliance7 records
Poolside financial estimates
Financial estimateRevenue estimate
Valuation estimate
Poolside leadership team
Management profileNumber of profiles
Profiles15 records
Poolside subsidiaries and ownership
Company hierarchySubsidiaries1 record
Poolside funding detail
Funding detailFunding overview
Funding rounds9 records
Investors30 records
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
Poolside M&A and investment
M&A and investmentM&A
Investments
M&A and investment is available on the Subscription and Enterprise plan.Contact sales →
Frequently asked questions about Poolside
What does Poolside do?
Poolside builds proprietary foundation models (the Laguna family, including Laguna M.1 and Laguna XS.2) trained specifically for software engineering tasks such as code generation, refactoring, code review, and test generation. It packages these models behind an enterprise AI platform (Poolside Platform) with an agent harness, IDE/CLI surfaces, MCP-based tool integrations, and the option to deploy fully on-premises or inside a customer's VPC for regulated buyers. The company also embeds Forward Deployed Research Engineers with customer teams to deliver outcomes, not just model access.
Is Poolside a public or private company?
Poolside is a private company. It is classified as venture growth investor backed and is currently operating.
When was Poolside founded?
Poolside was founded in 2023. It employs 251 to 500 people.
Where is Poolside based?
Poolside is headquartered in San Francisco, United States, in the North America region.
How does Poolside make money?
Four revenue lines are on record. Enterprise Foundation Model Licenses are the primary driver. The others are API / Usage-Based Model Access, forward Deployed Research Engineer (FDRE) Professional Services and custom Model Training Engagements.
Who are Poolside's main competitors?
Direct peers on record are Anysphere (Cursor), Cognition AI (Devin), Codeium (Windsurf), Mistral AI, Sourcegraph (Amp / Cody), Tabnine, Magic AI and Augment Code. Broad incumbents are GitHub Copilot and Anthropic (Claude Code).
Does Poolside have an API?
Yes. Poolside offers an OpenAI Chat-compatible API for its foundation models (Laguna M.1 and Laguna XS.2), enabling developers to build code generation and software engineering automation. The API is available via the Poolside Platform at platform.poolside.ai. Models are also distributed through OpenRouter and can be run locally via Ollama. The platform supports the Model Context Protocol (MCP) and Agent Client Protocol (ACP), allowing integration with MCP servers (e.g., Linear, GitHub, Slack, Jira, Notion) and ACP-compatible IDE clients. Free tier is available in Preview. Developer documentation is at platform.poolside.ai.
What industry is Poolside in?
Poolside's product category is AI Foundation Models for Software Engineering. Its primary akta.pro industry code is HDAAACAA, Foundation Model Developers (LLM/Multimodal Model Labs), with a secondary code of HDAEANAJ, AI Application Enablement Platforms (Copilot/Agent Frameworks, SDKs). Its NAICS code is 541511 and its SIC code is 7372.