# Reflection AI

**Type:** AI research lab
**Founded:** March 2024
**Founders:** Misha Laskin and Ioannis Antonoglou (former Google DeepMind researchers)
**Backing:** NVIDIA
**Valuation:** $25B (as of May 2026 negotiations)
**Status:** Active, no publicly released models

## Overview

Reflection AI is a frontier AI lab committed to open-weight model development. Despite having released zero AI models publicly, the company received Pentagon IL7 clearance in May 2026 for deployment on classified military networks.

## Timeline

- **2024-03** — Founded by Misha Laskin and Ioannis Antonoglou, former Google DeepMind researchers
- **2026-05-01** — Received Pentagon IL7 clearance for classified network AI deployment alongside AWS, Google, Microsoft, NVIDIA, OpenAI, SpaceX, and Oracle, despite having released no models
- **2026-05** — Negotiating at $25B valuation with zero deployed products

## Significance

Reflection AI represents a case study in governance architecture preference over capability demonstration. The DoD's IL7 pre-commitment to a zero-model company reveals that procurement decisions are selecting governance architecture (open-weight commitment) rather than assessed capabilities or security track record. The $25B valuation is entirely based on future open-weight commitment plus founding team pedigree, with DoD agreement implicitly endorsing this valuation before any product exists.

## Sources

- Breaking Defense, Defense One, Winbuzzer, TechCrunch, Nextgov/FCW (May 2026)