Radical Ventures is a Toronto-based venture capital firm focused exclusively on artificial intelligence. Founded in 2017 by Jordan Jacobs — co-founder of AI hub Vector Institute and Layer 6 — and Tomi Poutanen, Radical backs AI-first companies from day-zero incubation through growth and pre-IPO scaling. It is among the world's largest AI-dedicated VCs: its US$800M Venture Growth Fund (2024) was reported as the largest AI-focused fund of its kind at launch, and it held a US$650M final close on a new early-stage fund in October 2025; total AUM is reported around US$2.4B (CPP Investments cites over US$2.5B; Bloomberg cited ~US$1.8B at the 2024 growth-fund close). Radical's ~59-company 'Vanguard' portfolio spans foundation models (Cohere, Writer), robotics and embodied AI (Generalist, Waabi, Promise Robotics), AI infrastructure (Crusoe, Prime Intellect, Etched), and AI for science, health and materials (Xanadu, Latent Labs, Nabla Bio, Orbital Industries). Turing Award winner Geoffrey Hinton is an LP and advisor; Fei-Fei Li is a Scientific Partner.
Radical describes AI as 'the most consequential technology in human history' and backs 'AI visionaries who will define the future of progress.' Its mandate spans day-zero incubations and early-stage seed investing through growth-stage and pre-IPO scaling, with the firm acting as sole investor at incorporation in some cases.
Radical's early-stage funds target nascent AI companies (latest early-stage fund: US$650M final close, October 2025), while its first Venture Growth Fund (US$800M, 2024) invests in AI-first and AI-enhanced growth companies. The US$800M fund was reported by the Financial Times as the largest of its kind for AI at the time.
The ~59-company portfolio disclosed on radical.vc is concentrated in applied and foundational AI: foundation models and enterprise AI, robotics and embodied AI, AI compute infrastructure and semiconductors, plus AI for health, biology, climate and space.
Radical Velocity provides portfolio-exclusive resources spanning talent (scaling AI-first teams), go-to-market, compute and technology, community, brand and public affairs, and finance and governance, plus a free virtual AI Founders Masterclass for AI entrepreneurs.
Radical's advisory bench includes Dr. Fei-Fei Li (Scientific Partner), Colin Murdoch (Senior Advisor), and — per public reporting — AI pioneer and Turing Award winner Geoffrey Hinton as LP and advisor.
Radical Ventures' disclosed portfolio as published on radical.vc (Companies Archive, 'The Vanguard'). Of the 59 listed companies, Radical marks CentML, Covariant, Prepared, Solver and Untether AI as Acquired, and Xanadu as IPO.
AI behavior engine for games — agentic NPCs with dynamic, unscripted, real-time decision-making; founded by former Google DeepMind researchers.
3D bioprinting combined with therapeutic cells and computational design to create implantable tissue therapeutics, starting with diabetes.
Human-first Voice AI agent for healthcare access — answers every call instantly in the caller's language and connects them to the right care.
Efficient AI compute optimization for ML workloads; acquired by Nvidia in 2025.
AI-designed next-generation battery chemistries for EVs and the energy transition.
Trust and safety operations platform helping organizations combat AI-powered abuse, fraud and harmful content.
Security-first enterprise AI platform that helps organizations own their intelligence layer, with foundation models and end-to-end solutions.
AI that gives robots the ability to see, reason and act — dexterous automation for warehouses and supply chains.
Vertically integrated, energy-first AI data centers at gigawatt scale — including the 'Stargate' facility for OpenAI.
AI-native data security platform that discovers, classifies and protects sensitive data across hybrid clouds.
Automates curation and selection of AI training data so teams train better models with less data and compute.
Real-time world models and ultra-optimized infrastructure — generative experiences that respond at the speed of reality.
LLM-powered copilot for data science that accelerates building, deploying and improving predictive AI models.
Public benefit company automating scientific discovery — AI systems that run experiments in parallel and iterate autonomously; founded by Jeff Dean, Sanjay Ghemawat, Quoc Le and Oriol Vinyals.
Software that turns energy-intensive AI data centers into flexible grid assets, cutting peak power draw without losing compute.
Frontier inference clusters, co-designed across the stack with Low Voltage Inference and Cluster Scale Memory for faster, cheaper AI inference.
Lets publishers and brands build and distribute their own AI agents to consumers while retaining control over data and content.
Insurance-native AI agents that handle the full brokerage workflow — conversations, underwriting and compliance.
Embodied AI models that let robots master complex physical tasks; founded by DeepMind and Boston Dynamics alumni.
Unifies AI and biotech — machine learning across the small-molecule discovery pipeline to advance new therapies.
Neural search over unstructured, private data — instant natural-language answers and automated diligence workflows.
London-based AI research lab building general-purpose inventive AI to accelerate scientific discovery.
University of Toronto spinout using an AI-driven robotic platform to autonomously design and test drug formulations.
Founded by a key architect of DeepMind's Nobel Prize-winning AlphaFold2; AI foundation models making biology programmable via de novo protein design.
AI-driven deal-modeling platform that generates financial models and LBO analyses for investment professionals.
Earth intelligence delivered from orbit — full-stack satellite systems for real-time planetary monitoring and climate resilience.
Generative AI plus human-relevant testing to design antibodies against the hardest targets in medicine.
AI-native investment bank for SMBs — matching sellers with buyers and automating the M&A process at scale.
Generative AI to discover and design new materials for the energy transition, replacing slow lab trial-and-error.
AI-moderated user research running hundreds of qualitative interviews simultaneously across languages and geographies.
Building 'Archie', an AI engineer for the physical world that automates design of complex systems across domains.
Scientific superintelligence — frontier models, physics-based simulation and autonomous robotic labs for physical-science discovery.
Hyperspectral satellite constellation with AI-powered geospatial analysis for agriculture, energy and the environment.
Patient-centric medical imaging — patients own and access their imaging while hospitals get modern sharing tools.
Assistive AI for 911 call centers — real-time transcription, instant translation and intelligent dispatch support.
Open superintelligence stack — decentralized infrastructure for training, evaluating and deploying frontier AI models.
AI and robotics industrial production system for sustainable, affordable housing.
Enterprise-grade multimodal AI models that understand text, images, video and audio, with privacy and security built in.
AI recruiter that conducts two-way voice interviews around the clock for high-volume hiring.
AI-driven chip design with reinforcement learning — compressing semiconductor development from years to weeks.
AI-native IT service management that automates help-desk and operational requests through natural language.
AI embedded in hospital workflows to address surgical backlogs, staffing shortages and clinician burden.
Building 'Ash', a foundation model trained for psychological support and specialized therapeutic care.
Agent that completes repository-level tasks from natural-language instructions — building its own models, datasets and platform.
Conversational AI agents across chat, email and voice that qualify leads and drive sales pipeline.
At-home blood testing via magnetic resonance — non-invasive, needle-free metabolite monitoring for metabolic health.
Multimodal AI for video understanding — making massive video libraries as searchable and analyzable as text.
AI-powered newborn cry analysis to detect conditions like birth asphyxia before they cause harm.
Contextual code intelligence connecting scattered dev-tool data to surface the 'why' behind code.
AI-generated digital twins of patients to make clinical trials smaller, faster and more successful.
Energy-efficient, at-memory AI inference silicon from the data center to the edge.
Automates data labeling and training-data workflows so ML teams build models dramatically faster.
Turns hand-drawn sketches into photorealistic renderings and 3D models for industrial designers.
Autonomous AI that hauls — generative AI-first autonomous trucking with the world's most advanced closed-loop simulator.
Founded by Fei-Fei Li; large world models that give AI spatial intelligence over the full 3D world.
Full-stack generative AI platform for the enterprise, powered by Palmyra models and graph-based retrieval.
Photonic, room-temperature quantum computers built to solve problems beyond the reach of classical machines.
AI search that answers complex questions and automates workflows across private data, live web and LLMs.
Agent-first personal AI assistants that autonomously carry out everyday digital tasks.
No rounds led by Radical Ventures are in the tracker yet. Deals will appear here as they are recorded.