R
Reka
Core neolabMultimodal and omni models and infrastructure for physical AI
Raised, confirmed$170M3 confirmed rounds
Valuation, confirmed—none disclosed
Valuation, reported—no open report
Founded2022Sunnyvale, CA, US
Key facts
| Field | Value | Source · checked |
|---|---|---|
| Classification | Core neolabResearch lab founded in 2022; $60M (Jun 2023) and $110M (Jul 2025) announced rounds | |
| Founded | 2022 | |
| Headquarters | Sunnyvale, CA, US | |
| Founders | Dani Yogatama, Cyprien de Masson d'Autume, Mikel Artetxe, Qi Liu, Yi Tay | |
| Website | reka.ai | |
| Careers | reka.ai9 open roles on 9 Nov 2026 | reka.aichecked 9 Nov 2026 |
| Profiles and work | X @RekaAILabsLinkedInHugging Face, 5 modelsGitHub, 26 public reposYouTubeDocs | How links are checkedchecked 10 Nov 2026 |
Funding rounds
Reka has raised $170M in rounds confirmed by the company, a regulator or two independent outlets. Status rules: methodology.
| Date | Round | Amount | Valuation | Investors | Status | Source · checked |
|---|---|---|---|---|---|---|
| 22 Jul 2025 | Funding round | $110M | — | NVIDIA, Snowflake Ventures | Confirmed | Reka announced it has secured a $110 million investment. This funding is backed by new and existing investors including NVIDIA and Snowflake.checked 9 Nov 2026 |
| 27 Jun 2023 | Funding round | $60M | — | Lead: DST Global, Radical Ventures; Snowflake Ventures | Confirmed | Reka funding of $60 million dollars led by DST Global Partners and founding investor Radical Ventures along with participation from strategic partners (including Snowflake Ventures)+1 more sourcechecked 9 Nov 2026 |
| 29 Sep 2022 | Form D offering | $7.45M | — | not disclosed | Confirmed | SEC EDGAR Form D: total offering 7451233.0, total sold 7451233.0, first sale 2022-09-29checked 10 Nov 2026 |
Notes on these rounds
- 22 Jul 2025 Funding round: Valuation not disclosed in the announcement.
- 27 Jun 2023 Funding round: Announced on emerging from stealth.
- 29 Sep 2022 Form D offering: SEC Form D reports securities sold under the exemption, not the announced round size.
People
| Name | Role | Before | Now | Source · checked |
|---|---|---|---|---|
| Dani YogatamaXLinkedInScholarHugging FaceSite | Co-founder & CEO | — | Current | |
| Cyprien de Masson d'Autume | Co-founder | — | Current | |
| Mikel ArtetxeXScholarHugging FaceSite | Co-founder | — | Current | |
| Qi LiuXScholarSite | Co-founder | — | Current | |
| Yi Tay | Co-founder | — | Current | |
| Mateusz MalinowskiXHugging Face | Research leadership | Google DeepMind (Staff research scientist) | Current | reka.ai Malinowski, previously a staff research scientist, and Bińkowski, previously a senior research scientist at Google DeepMindchecked 9 Nov 2026 |
| Mikołaj BińkowskiLinkedIn | Research leadership | Google DeepMind (Senior research scientist) | Current | reka.ai Malinowski, previously a staff research scientist, and Bińkowski, previously a senior research scientist at Google DeepMindchecked 9 Nov 2026 |
Professional facts from public sources only. How people records are handled.
What it has shipped
| Name | Kind | Date | Open weights | Source · checked |
|---|---|---|---|---|
| Reka Inverse Dynamics Model (RIDM) | model | 2 Oct 2026 | Yes | reka.ai RIDM extracts low-level commands like moving forward, sideways, turning or tilting from real-world video, for interactive world models. It is trained entirely on video games.checked 10 Nov 2026 |
| Rho-1 | model | 5 Oct 2026 | No | reka.ai Rho-1 collapses that stack. Text, vision, and robotic actions are unified as tokens within a single context window.checked 10 Nov 2026 |
| Reka Edge | model | 10 Apr 2026 | Yes | reka.ai Meet Reka Edge, an efficient 7B vision-language model built for physical AI, edge deployment, visual reasoning, object detection, and video analysis.checked 9 Nov 2026 |
| Reka Flash | model | 12 Feb 2024 | — | |
| Reka Core | model | 15 Apr 2024 | No |
Latest posts on X
The most recent posts from @RekaAILabs, the X account of Reka, quoted as posted. Replies and reposts are left out. How posts are selected.
Today, we are releasing a research preview of Rho-1, our 19B omni model that understands and generates text, images, video and robot actions, all inside a single neural network.
View the post on X · With video
Standard simulated environments lack real-world physics and noise. Interactive World Models get closer, but training them needs video labelled with low-level actions, and real footage rarely has those labels. Today, we’re releasing the Reka Inverse Dynamics Model (RIDM), a compact model that extracts low-level camera/motor commands directly from raw video. 🕹️Trained entirely on video games 🌐Generalizes to real-world footage 🔐 Weights open-sourced under Apache 2.0 Read the research note and grab the weights on 🤗Hugging Face…
Read the full post on X · With video
Deploying physical AI on wearables and phones requires real-time perception inside extreme thermal and battery limits. Reka EdgeQ is an optimized on-device VLM running natively on @Qualcomm #Snapdragon 8 Elite’s Hexagon NPU. ⚡️ 0.73s Time to First Token (Image) 📹 +34 pts over Gemma 4 E4B on MLVU video benchmarks 🔋 6.9 mWh energy per inference (~3x lower heat overhead) By custom-tuning our ConvNeXt V2 vision encoder and decoder directly for the #NPU, EdgeQ keeps the GPU completely idle, minimizing thermal throttling under sustained continuous load. 🔗Our technical breakdown…
Read the full post on X · With video
checked 10 Nov 2026
Commonly cited, checked
Figures and statements about Reka that circulate in lists and summaries, and what the sources support.
| Claim | Finding | Source |
|---|---|---|
| $12M Series A-2 at a $1.3B post-money valuation (Feb 2026) | Not confirmed: the company's news archive lists no funding announcement in 2026; its latest announced round is $110M on 22 Jul 2025, with no valuation stated. | reka.ai announced it has secured a $110 million investment. |
| Unicorn valuation ($1B+) | No company or filing source opened states a valuation for Reka. | reka.ai Announcing our $60M funding to Build Generative Models and Advance AI Research |
Work at Reka, or see something wrong? Update this profile with a public source. How corrections work.