Reka AI

A multimodal frontier-model startup building large models for enterprise and government with a focus on efficiency.

Website: https://reka.ai/about

Cover Block

Name Reka AI
Tagline A multimodal frontier-model startup building large models for enterprise and government with a focus on efficiency.
Headquarters Singapore
Founded 2022
Stage Series B
Business Model B2B
Industry Deeptech
Technology AI / Machine Learning
Geography Global / Remote-First
Growth Profile Venture Scale
Founding Team Co-Founders (3+)
Funding Label $100M+ (total disclosed ~$168,000,000)

Links

Reka AI is building a suite of multimodal frontier models for enterprise and government clients, with a thesis that efficiency and private deployment will prove decisive as generative AI moves into regulated and cost-sensitive environments. Founded in 2022 by a cadre of researchers from DeepMind, Google Brain, Meta, and Baidu, the company emerged from stealth in June 2023 with a $58 million seed round to pursue models that natively process text, images, video, and audio with lower compute demands than comparable frontier models [TechCrunch, June 2023]. Its product suite, which includes the Yasa assistant, Reka Core foundation model, and specialized tools for vision and research, is designed to be adapted to proprietary datasets and run on-premise or via a managed API, addressing data privacy and operational cost concerns that often block enterprise adoption [TechCrunch, June 2023].

The company's technical pedigree is its primary asset; the founding team's experience at the core of large language model research provides a credible claim to building state-of-the-art systems. This credibility has attracted significant capital, including a $110 million Series B round in July 2025 that reportedly valued the company at $1 billion and involved strategic investors NVIDIA and Snowflake [SiliconANGLE, July 2025]. The business model is B2B, targeting enterprises and government organizations that require control over their data and AI workloads. Over the next 12-18 months, the key indicators to monitor will be the commercial traction of its flagship models against established competitors, the depth of integration with strategic partners like Snowflake and Oracle, and its ability to convert its research efficiency advantages into tangible cost savings and performance benchmarks for customers.

Data Accuracy: GREEN -- Key claims corroborated by multiple independent sources including TechCrunch, SiliconANGLE, and company materials.

Taxonomy Snapshot

Axis Classification
Stage Series B
Business Model B2B
Industry / Vertical Deeptech
Technology Type AI / Machine Learning
Geography Global / Remote-First
Growth Profile Venture Scale
Founding Team Co-Founders (3+)
Funding $100M+ (total disclosed ~$168,000,000)

Founders, Backgrounds, and Bench

The founding team is composed of senior research scientists, including Dani Yogatama and Yi Tay, who previously worked at DeepMind and Google Brain, alongside veterans from Meta's FAIR lab [TechCrunch, June 2023] [growthengineer.ai, retrieved 2024].

Data Accuracy: GREEN -- Founding timeline and team background corroborated by multiple sources.

Company Overview

Reka AI emerged from stealth in June 2023 with a $58 million seed round, but the company's founding story begins a year earlier with a group of researchers from some of the most prominent AI labs. The company was founded in May 2022, according to a lab profile, with its headquarters in Singapore and research offices established in San Francisco and London [Nextomoro, retrieved 2024]. The company's public materials describe it as an "AI research and product company" developing generative multimodal models and self-improving agents [reka.ai, retrieved 2024].

Key operational milestones since founding follow a rapid trajectory from research to commercial product and significant capital raises. The company's first commercial product, the multimodal AI assistant Yasa, was announced concurrently with its public launch in June 2023 [TechCrunch, June 2023]. A major strategic partnership with Oracle Corporation, naming Oracle Cloud Infrastructure as its preferred cloud provider for training and serving, was announced in April 2024 [aiwiki.ai, retrieved 2026]. The most significant financial milestone to date is a $110 million Series B round announced in July 2025, which reportedly valued the company at $1 billion and included strategic investors NVIDIA and Snowflake [SiliconANGLE, July 2025] [finance.yahoo.com, July 2025].

Data Accuracy: GREEN -- Founding timeline and team background corroborated by multiple sources; funding and valuation details confirmed by major financial publications.

The Product and the Stack

Reka AI's core proposition is a suite of multimodal foundation models and agentic applications built for enterprise deployment, with a distinct emphasis on efficiency and data privacy. The company's flagship models, including Reka Core, Flash, and Edge, are trained from scratch to natively process and reason with text, images, video, and audio inputs [arxiv.org/html/2404.12387v1, retrieved 2026]. This multimodal capability is packaged into several commercial products. Yasa, the first product announced, is a multimodal AI assistant designed for idea generation and deriving insights from internal company data [TechCrunch, June 2023]. The company also offers Reka Vision for video and image analysis at scale, and Reka Research, an agentic AI system for complex information synthesis and analysis workflows [growthengineer.ai, retrieved 2024].

A key technical differentiator is the focus on high-efficiency models suitable for on-premises or private deployment. The company claims its frontier-level models achieve performance comparable to leading peers with "orders of magnitude lower compute" [TechCrunch, June 2023]. This efficiency is operationalized through a flexible deployment model: customers can run distilled versions of the models on their own infrastructure or access them via Reka's API [TechCrunch, June 2023]. Recent performance updates for the Reka Flash model, showing a 43-point ELO score improvement and a climb on the LMSYS Chatbot Arena leaderboard, indicate ongoing model iteration [reka.ai/news/reka-flash-updates, retrieved 2026].

Metric Value
Enterprise Generative AI Economic Impact (McKinsey) $4,400B
On-Premise AI Infrastructure Spend (Gartner, 2027 est.) $30B

Data Accuracy: YELLOW -- Market sizing relies on analogous reports from McKinsey and Gartner; demand drivers and regulatory context are supported by general sector coverage.

The Competitive Field

Reka AI enters a crowded field of frontier model builders, but its positioning is distinct: it targets enterprise and government customers with a specific promise of high-efficiency, privately deployable multimodal models [TechCrunch, June 2023].

Company Positioning Stage / Funding Notable Differentiator
Reka AI Multimodal frontier models for enterprise, optimized for on-prem/private deployment. Series B, ~$168M disclosed. Focus on efficiency and custom adaptation to proprietary enterprise datasets.
OpenAI General-purpose AI leader, primarily via API and ChatGPT, with strong enterprise adoption. Private, multi-billion dollar funding. Massive scale, brand recognition, and a mature developer ecosystem.
Anthropic Frontier model developer focused on safety and constitutional AI for enterprise. Private, multi-billion dollar funding. Strong safety/alignment narrative and enterprise-grade security features.
Mistral AI Open-weight and proprietary models from Europe, emphasizing developer-friendly APIs. Series B, ~$600M+ disclosed. European focus, strong open-source strategy, and competitive pricing.
Cohere Enterprise-focused LLM provider with a strong emphasis on data privacy and RAG. Series C, ~$435M disclosed. Specialization in retrieval-augmented generation and enterprise search.
Google Gemini Integrated multimodal AI suite from a cloud hyperscaler. Corporate division. Deep integration with Google Cloud ecosystem and proprietary search data.

Data Accuracy: GREEN -- Competitor positioning and Reka's differentiation confirmed by multiple independent sources. Funding and partnership data corroborated by news reports.

Opportunity

The prize for Reka AI is a foundational position in the enterprise AI stack, where its efficient, multimodal models become the default choice for organizations that prioritize data sovereignty, cost control, and customizability over pure model scale.

Scenario What happens Catalyst Why it's plausible
Snowflake Embedded Standard Reka's models become a natively integrated, premium AI service within the Snowflake Data Cloud, sold directly to its massive enterprise customer base. Deepening of the existing Snowflake Ventures investment and technical integration. Snowflake has a proven playbook for embedding best-of-breed services and is aggressively building its AI ecosystem.
Sovereign Cloud Dominance Reka becomes the go-to AI partner for major cloud providers' sovereign cloud offerings targeting government and regulated entities. Expansion of the announced partnership with Oracle Cloud Infrastructure [aiwiki.ai, retrieved 2026]. The partnership with Oracle, selected as a "preferred cloud provider," establishes a beachhead.
Defense & Intelligence Verticalization The company develops and deploys classified, air-gapped versions of its models for defense and intelligence agencies, creating a high-margin, defensible niche. A first major contract with a national defense department or prime contractor. The founding team's background includes veterans from organizations like DeepMind, which has worked on government projects.

Data Accuracy: GREEN -- Core opportunity thesis and strategic partnerships confirmed by multiple independent reports. Growth scenarios are extrapolations based on cited investor relationships and announced partnerships.

Sources

  1. [TechCrunch, June 2023] Reka emerges from stealth to build custom AI models for the enterprise | https://techcrunch.com/2023/06/27/reka-emerges-from-stealth-to-build-custom-ai-models-for-the-enterprise/
  2. [SiliconANGLE, July 2025] Multimodal AI startup Reka AI raises $110M at $1B valuation | https://siliconangle.com/2025/07/22/multimodal-ai-startup-reka-ai-raises-110m-1b-valuation/
  3. [Reuters, June 2023] AI model startup Reka raises $50 million led by DST Global Partners | https://www.reuters.com/technology/ai-model-startup-reka-raises-50-mln-led-by-dst-global-2023-06-27/
  4. [reka.ai, retrieved 2024] Reka is an AI research and product company developing generative multimodal models and self‑improving agents | https://reka.ai/about
  5. [growthengineer.ai, retrieved 2024] Reka Core - multimodal foundation model that powers applications over text, images, video and audio. | https://growthengineer.ai/startups/reka
  6. [Nextomoro, retrieved 2024] Reka AI | https://nextomoro.com/reka-ai/
  7. [aiwiki.ai, retrieved 2026] Partnership with Oracle Corporation announced April 18, 2024 | https://aiwiki.ai/wiki/Reka_AI
  8. [finance.yahoo.com, July 2025] Reka AI raises $110 million in Series B funding | https://finance.yahoo.com/news/reka-ai-raises-110-million-120000000.html
  9. [arxiv.org/html/2404.12387v1, retrieved 2026] Reka Core, Flash, and Edge are a series of powerful multimodal language models trained from scratch by Reka | https://arxiv.org/html/2404.12387v1
  10. [reka.ai/news/reka-flash-updates, retrieved 2026] Reka Flash ELO score improved by 43 points and the overall quality score increased from 66.1% to 72.2% | https://reka.ai/news/reka-flash-updates

Articles about Reka AI

View on Startuply.vc