Reka

Commercial AI Research

Reka is a multimodal AI company headquartered in Sunnyvale, California, founded in 2022 by former researchers from DeepMind, Google Brain, and FAIR.

Reviewed by 7wData
Share Email

On this page

Profile

Reka builds multimodal AI models and infrastructure for processing text, images, video, and speech, with products for edge deployment, video analysis, and custom enterprise AI.

Reka is a multimodal AI company headquartered in Sunnyvale, California, founded in 2022 by former researchers from DeepMind, Google Brain, and FAIR. As of mid-2026, the company reports approximately $12.4 million in annual revenue and employs between 51 and 200 people. In July 2025, Reka raised $110 million in Series B funding led by Nvidia and Snowflake, reaching a valuation of over $1 billion and achieving unicorn status.

The company has evolved from a focus on custom enterprise language models to a broader portfolio of multimodal AI products. Its offerings include Reka Edge, a frontier-level edge intelligence model for physical AI; Reka Vision, a video processing infrastructure for tagging, searching, and clipping large volumes of video; Infer, an inference engine and API for multimodal AI; and Claru, a data infrastructure subsidiary that produces training data such as egocentric video and robotics trajectories. Reka has also developed Reka Speech, a high-throughput speech transcription and translation model, and a parallel thinking capability for its research tier.

The company has formed strategic partnerships with Alibaba Cloud to distribute its models through Alibaba Cloud's Model Studio, and with Oracle's defense ecosystem to advance multimodal AI for mission readiness. Reka Edge is available on OpenRouter, and the company has released an n8n community node for AI-powered video and image analysis. The company's market position is as a provider of multimodal AI infrastructure for enterprise and public sector customers, competing with larger foundation model providers while differentiating through edge deployment and video-specific capabilities.

Track Reka and 240+ vendors.

335k+ subscribers read the daily AI & data note. One email, both newsletters. Unsubscribe anytime.

Who buys this

  • Enterprises needing custom or private AI models for specific use cases
  • Broadcasters and media companies with large video archives
  • Public sector and defense organizations requiring secure, on-premises AI
  • Smart home and security companies needing video AI
  • Cloud platform customers (via Alibaba Cloud Model Studio)

Strengths and what to watch

Strengths

  • Multimodal model portfolio spanning from 2 billion to over 100 billion parameters, enabling flexible deployment on-device, on-premises, or in the cloud
  • Strategic partnerships with major infrastructure players including Nvidia, Snowflake, Alibaba Cloud, and Oracle, providing distribution and compute access
  • Clear product-market fit in video processing and edge AI, with dedicated products (Reka Vision, Reka Edge) that address specific enterprise pain points

Watch for

  • Revenue of $12.4 million against a $1 billion+ valuation implies a high multiple, raising questions about growth trajectory and path to profitability
  • The AI model market is crowded with well-funded competitors (OpenAI, Anthropic, Google, Meta) and open-source alternatives, putting pressure on differentiation
  • Dependence on partnerships with cloud providers (Snowflake, Alibaba Cloud, Oracle) for distribution creates concentration risk if those relationships shift

Recent moves

Key Information

Industry
Commercial AI Research
Founded
2022
Headquarters
Sunnyvale, California

Frequently Asked Questions

What is Reka AI and what does it do?

Reka is a multimodal AI company founded in 2022 by former DeepMind, Google Brain, and FAIR researchers. It builds AI models and infrastructure for processing text, images, video, and speech, with products for edge deployment, video analysis, and custom enterprise AI.

What products does Reka offer?

Reka offers Reka Edge for frontier-level edge intelligence, Reka Vision for video processing and search, Infer as an inference engine and API, Reka Speech for transcription and translation, and Claru for data infrastructure. It also provides custom enterprise AI models.

How much funding has Reka raised and what is its valuation?

In July 2025, Reka raised $110 million in Series B funding led by Nvidia and Snowflake, reaching a valuation of over $1 billion and achieving unicorn status. The company had previously raised $50 million in a 2023 round led by DST Global.

What is Reka Edge and how does it work?

Reka Edge is a frontier-level edge intelligence model for physical AI, designed to run on devices rather than in the cloud. It is available on OpenRouter and enables AI processing at the edge for applications like smart home and security systems.

What is Reka Vision used for?

Reka Vision is a video processing infrastructure for tagging, searching, and clipping large volumes of video. It targets broadcasters, media companies, smart home and security firms, and public sector organizations needing efficient video analysis without replacing existing VMS.

Who are Reka's main customers and partners?

Reka serves enterprises needing custom AI, broadcasters, public sector and defense organizations, and smart home companies. Key partners include Nvidia, Snowflake, Alibaba Cloud for model distribution, and Oracle's defense ecosystem for mission readiness AI.

Sources

  1. www.zoominfo.com — Revenue of $12.4 million, headcount 51-200, headquarters in Sunnyvale, CA
  2. reka.ai — Product announcements including Reka Edge, Reka Vision, Reka Speech, partnerships with Alibaba Cloud and Oracle, and availability on OpenRouter
  3. siliconangle.com — Series B funding of $110 million led by Nvidia and Snowflake, valuation of $1 billion+
  4. www.reuters.com — Initial $50 million funding round led by DST Global, Snowflake investment, founding team from DeepMind
  5. reka.ai — Partnership with Alibaba Cloud to distribute Reka models through Alibaba Cloud Model Studio