Gradient Ventures is Google's dedicated AI investment fund, founded in 2017 and headquartered in San Francisco, California. It operates as a standalone entity within Alphabet, separate from Google Ventures (GV) and CapitalG. The fund targets seed and early Series A companies building AI-native products, typically writing checks in the $1–5 million range. Total AUM is not publicly disclosed.
Unlike traditional venture funds, Gradient pairs capital with direct access to Google's engineering networks, cloud infrastructure credits, and recruiting pipelines. This operational model positions it closer to a corporate accelerator than a classic VC, which attracts founders who want Google's distribution potential alongside funding. The fund operates across enterprise software, developer tooling, applied machine learning, and – to a limited degree – infrastructure plays that touch blockchain and digital assets.
Gradient's crypto and Web3 footprint is modest. With roughly five known portfolio positions in the space and a single lead investment, the fund treats distributed ledger technology as a secondary theme rather than a core thesis. Its primary lens remains AI-first companies, and blockchain deals typically enter through that angle – for example, projects that apply machine learning to on-chain data, fraud detection, or decentralized infrastructure management.
Notable Investments
- Algorithmia – ML model deployment platform (acquired by DataRobot in 2021). One of the fund's most cited early bets on AI infrastructure.
- Standard Cognition – Autonomous checkout using computer vision; raised over $50 million across rounds with Gradient participating early.
- Unbabel – AI-powered translation for customer support, bridging language barriers at scale.
- Cape Privacy – Federated and encrypted machine learning, relevant to privacy-preserving data networks including those with blockchain components.
- Dashbot – Conversational AI analytics platform used by chatbot and voice interface developers.
Specific crypto-native investments from the fund are not well-documented in public filings or press as of early 2026. Public information about individual deal sizes within the digital-assets sleeve is limited.
Team
Anna Patterson founded Gradient Ventures after a career at Google leading search infrastructure engineering. She served as managing partner from 2017 onward and shaped the fund's AI-first mandate. The investment team is small – typically four to six investors – supplemented by a larger network of Google engineers who serve as advisors to portfolio companies. Darian Shirazi joined as a partner, bringing operator experience from enterprise software. The lean team structure means Gradient focuses on concentrated positions rather than spray-and-pray deployment.
Recent Activity
Through 2024 and into 2025, Gradient sharpened its focus on generative AI infrastructure, backing companies building tools for large language model deployment, evaluation, and safety. The fund participated in several seed rounds for developer-facing AI platforms, consistent with the broader market rotation toward AI tooling following the generative AI wave. No major crypto-specific announcements were reported in this period, reinforcing the view that digital assets remain peripheral to the fund's strategy.
The fund's reported retail ROI in the crypto category sits at approximately 0.04 – a figure that reflects the limited scale and early stage of those positions rather than distressed outcomes. For context, Gradient's headline exits have come from AI infrastructure acquisitions, not token markets.
Gradient Ventures remains one of the more unusual vehicles in venture: a corporate fund with genuine early-stage conviction, backed by one of the world's largest technology companies. For crypto founders building at the intersection of AI and decentralized systems, it represents a potential strategic partner – though deal flow into Web3 has been selective. Founders seeking pure crypto-native capital will find more active investors elsewhere; those building AI products that touch on-chain data or trustless compute have a stronger case to make to Gradient's team.
Further reading: Gradient Ventures official site – Crunchbase profile – TechCrunch coverage