Companies

Reflection AI

reflection.ai

Reflection AI builds open AI models and autonomous coding agents for enterprise, government, and sovereign users.

HQNew York, New York, United States
Employees51-200
Funding$2.13B
Valuation$8B
Revenue$4.6M
51 active roles
Profile 6mo agoJobs checked 16h ago
AI / MLFoundation Model ProviderB2B SaaSSeries B$1B+

About

Reflection AI builds open foundation models and autonomous coding agents for software development. It targets enterprises, governments, and sovereign entities, differentiating itself through open research, customizable models, and customer ownership and control of AI.

Market

Reflection AI competes across frontier foundation models, agentic AI, and developer tooling, positioning itself as an open-model alternative to closed labs and as a Western counterpart to open-model players such as DeepSeek. It differentiates through an integrated stack combining large-scale LLM pretraining, reinforcement learning, MoE architectures, and agentic applications with an emphasis on model ownership, self-hosting, and sovereign control. In software development, Asimov emphasizes deep code comprehension and persistent engineering knowledge rather than only code generation.

Target Customers

Reflection AI primarily targets large enterprises—especially engineering organizations—and governments or sovereign entities that want to own, run, and control capable AI. Its strongest fit is with enterprise, government, and regulated-industry buyers that require controlled deployments, data privacy, and the ability to operate AI within their own infrastructure.

At a Glance

Problem

Reflection AI is targeting the expensive, slow, and difficult work of making software understandable and executable at scale. Large, complex codebases bury context across files, GitHub threads, chat history, and other artifacts, while engineering knowledge work remains difficult to automate reliably. Its core thesis is that autonomous coding agents can turn this bottleneck into machine-assisted work, with codebase comprehension as the clearest immediate use case and autonomous coding as the broader ambition.

The economic logic is labor leverage: a system that can research, understand, and act on a codebase could reduce the time highly paid engineering teams spend locating context and performing repetitive computer-based knowledge work. The research does not quantify savings or disclose customer ROI, so this is a use-case and business rationale rather than a measured outcome.

Product / Service

Reflection combines open foundation models with agentic software for AI-assisted development. Its named product is Asimov, a code-research agent designed to help engineering teams understand large, complex codebases—effectively “deep research for code comprehension.” The company describes itself as both a research and product company building open superintelligence, and it trains mixture-of-experts agentic models, suggesting a model-plus-agent stack rather than a single chatbot.

Commercially, the evidence describes a B2B SaaS model aimed at enterprise engineering organizations, complemented by full-stack partnerships for enterprise and government customers and an open-model strategy. The intended benefit is controlled, scalable access to intelligence: customers can use the agent to navigate code and, over time, automate more autonomous coding, while open model weights and on-premises deployments extend the offering beyond a hosted product.

Market

Reflection sits at the intersection of frontier AI, open foundation models, AI-assisted software development, and autonomous coding agents. It is not only a code-tool vendor: its positioning spans open models, research agents, enterprise software, government use, and on-premises deployment. The available evidence points to competition with other open-model labs and coding-agent platforms, but it does not provide a definitive named competitor list; named rival comparisons therefore remain unresolved.

It has substantial financing and commercial and strategic traction despite no publicly verified revenue figure or customer-scale metric in the supplied research. Reflection has raised $2.0 billion in total funding and is reported to have an $8 billion private valuation; it launched Asimov in July 2025, listed 62 open positions in the latest directory evidence, and announced infrastructure and deployment partnerships including GMI Cloud and Dell. It should therefore be described as well-funded and actively commercializing, rather than as a company with publicly verified revenue.

Founders & Leadership

Misha LaskinFounder
Co-Founder & CEO
Ioannis AntonoglouFounder
Co-Founder, President & CTO
Joseph SpisakVP of Product & Head of Open Source
Drayton WadeVice President, Operations

Funding History

2025-03
Seed$25M

Sequoia Capital, CRV

2025-03
Series A$105M

Lightspeed Venture Partners, CRV

2025-10
Series B$2B

NVIDIA, Lightspeed Venture Partners

Recent News

2026-07-15partnership
AI Startup Reflection Inks $1 Billion Compute Deal To Challenge China’s Open Source Dominance

Reflection announced a deal to purchase $1 billion of computing capacity from Nebius to train its open-source models. The agreement follows Reflection’s major compute deal with SpaceX and comes ahead of its planned public model release.

2026-06-22partnership
SpaceX signs computing power deal with open-source AI startup Reflection worth up to $6.3 billion

SpaceX agreed to provide Reflection access to Nvidia GB300 chips at its Colossus 2 data center. Reflection will pay $150 million per month beginning July 2026, potentially totaling approximately $6.3 billion through 2029.

2026-05-26
AI & Tech Brief: Reflection AI arrives in Washington

The Washington Post covered Reflection’s push into Washington and its argument that open-source models are a national-security imperative.

2026-05-22partnership
Exclusive: Reflection AI to power Genesis Mission

Reflection partnered with the U.S. Department of Energy to support the Genesis Mission and serve as an AI model provider for the national laboratories. The company will provide customizable models and use DOE computing resources as the technology is deployed across research projects.

2026-05-18partnership
Dell Technologies Closes the Gap Between AI Ambition and AI Outcomes

Dell announced that Reflection’s open-source frontier models are coming on-premises through the Dell AI Factory. The offering is positioned for governments, sovereign entities, and regulated industries that need controlled AI deployments.

2026-05-01partnership
Pentagon inks deals with Nvidia, Microsoft, and AWS to deploy AI on classified networks

The U.S. Defense Department signed agreements with Nvidia, Microsoft, AWS, and Reflection AI to deploy their AI systems on classified networks.

2026-04-23funding
Reflection CEO confirms latest funding round closed at $25 billion pre-money valuation

Reflection CEO Misha Laskin confirmed that the company’s latest funding round closed at a $25 billion pre-money valuation and discussed the global competition in open AI models.

2026-03-02funding
Nvidia-backed ‘open’ AI start-up courts investors at $20bn-plus valuation

The Financial Times reported that Nvidia-backed Reflection was courting investors at a valuation exceeding $20 billion, highlighting continued investor interest in its open-model strategy.

2025-10-09funding
Nvidia-backed Reflection AI raises $2 billion in funding, boosts valuation to $8 billion

Reflection raised $2 billion in a funding round valuing the company at $8 billion. Investors included Nvidia, Eric Schmidt, Citi, 1789 Capital, Lightspeed, Sequoia, and other existing and new backers.

Active Roles

51
San Francisco, CA/Sales/6d ago
New York, NY/Data & Analytics/9d ago
San Francisco, CA/Data & Analytics/21d ago
Washington, D.C./Marketing/34d ago
San Francisco, CA/Data & Analytics/45d ago
New York, NY/Marketing/50d ago
New York, NY/Engineering/51d ago
New York, NY/Engineering/51d ago
San Francisco, CA/Engineering/76d ago
San Francisco, CA/Professional Services/83d ago
San Francisco, CA/Marketing/85d ago
Seoul/Forward-Deployed Engineer/86d ago
Seoul/Forward-Deployed Engineer/86d ago
San Francisco, CA/HR & Recruiting/91d ago

Business Model

Reflection AI operates a B2B SaaS model selling AI coding-agent and software-engineering tools to enterprise organizations. Reported enterprise contracts typically range from $15,000 to $25,000 per user annually, with deployments initially focused on engineering teams.

Products

Asimov, a code-comprehension and code-research agent for engineering teamsFrontier open foundation models and associated open model weightsEnterprise and sovereign AI deployment solutions, including private AI Factory infrastructure

Customers

U.S. Department of Energy (DOE) / 17 U.S. national laboratories — announced AI-model provider and foundational intelligence layer for the Genesis MissionU.S. Department of Defense (Pentagon) — publicly reported deal for use of Reflection AI in classified environmentsShinsegae Group — MOU to build a sovereign AI cloud serving Korean enterprises and government agencies

Tech Stack

Large language models (LLMs)Reinforcement learningMixture-of-Experts (MoE) architecturesLarge-scale pretrainingMulti-agent architectureLong-context retrieval and reasoning agentsFrontier-scale compute clusters

Competitors

OpenAI
Anthropic
DeepSeek
Qwen
Mistral AI
AI2/OLMo
NVIDIA
Cursor

Key Investors

1789 Capital, B Capital, Citi Ventures, CRV, Lightspeed Venture Partners, NVentures, Databricks, Sequoia Capital, SV Angel, Reid Hoffman, Lachy Groom, DST Global, GIC, Disruptive, Nvidia