Participated · Seed · Aug 2026

Synthefy develops foundation models specifically for structured numerical data — data types such as tables, time series, transactions, sensor readings, and other numeric signals. Its flagship open-source model, Nori, is available under an Apache 2.0 license and the company says Nori ranks first across 130 public regression benchmarks; Synthefy reported almost 600K Nori model downloads and about 5,000 Python installations in the weeks after launch. The platform is designed for enterprise deployment with options including an open model, managed API, enterprise platform, and on-premise deployments integrated with AWS, Databricks, Snowflake, and Google Cloud. Synthefy positions its models as smaller, purpose-built architectures that require less compute and energy than large general-purpose language models. Targeted use cases cited include demand forecasting, pricing optimization, fraud and risk detection, infrastructure failure prediction, and operational improvements across retail, financial services, telecommunications, infrastructure, healthcare, and defense. The company was founded in 2023 and the team includes former engineers and researchers from Meta, Nvidia, Adobe, Uber, and top academic institutions.

Participated · Seed · Aug 2026

Xpander provides a universal, cloud-agnostic platform and runtime that lets enterprises build, run, and manage agentic AI applications within their own environments while centrally governing lifecycle and security. Its flagship agent, Omni, is positioned as an agentic Forward Deployed Engineer and scored 90.9% on the GAIA benchmark. The platform is available across major clouds and models and is already used by customers in retail, manufacturing, financial services, technology, and government. Xpander is SOC 2 Type II certified and GDPR compliant. Founded in 2024 by former AWS principal engineers David Twizer, Moriel Pahima, and Ran Sheinberg and headquartered in San Francisco, the company plans to use its recent Seed funding to accelerate market penetration.

Participated · Series A · Jun 2026

Gray Swan builds an AI security platform that combines real-time protection (Cygnal), automated adversarial testing/red-teaming (Shade), and a global red-teaming community (Arena) to detect and prevent dangerous model and agent behavior. Cygnal enforces custom enterprise policies in real time, while Shade runs automated pre-deployment adversarial tests as part of CI/CD, and Arena crowdsources attacks from more than 15,000 researchers to generate over one million attack trajectories. The company says its benchmarks are embedded in the safety evaluation processes of leading frontier labs and that it is cited in 11 recent frontier model system cards, including those from Anthropic, OpenAI, and Meta. Gray Swan works with over 20 customers and has partnerships with platforms such as Snowflake, including a native integration for runtime AI protection. The recent $40M Series A will fund go-to-market scaling, deeper lab and platform partnerships, and team growth to support enterprise deployments.

Participated · Seed · May 2026

Graphon AI builds a pre-model intelligence software layer that uses mathematical graphon functions to map hidden structural connections across video, audio, documents, and structured databases. Its continuous multimodal graph technology is designed to bypass token limits of legacy context windows and enable third-party foundation models and autonomous agents to reason over large, interconnected multimodal data arrays. The company has launched production deployments with GS Group to automate customer movement mapping in retail environments and provide real-time safety analytics across industrial construction sites. Management and advisors include former AI and robotics researchers from Meta, Amazon, Google, and NASA, and technical guidance from UC Berkeley professors Dr. Christian Borgs and Dr. Jennifer Chayes. Graphon plans to use its seed funding to further develop its relational representation platform, advance its core technology, and scale infrastructure to support enterprise-wide AI agent frameworks.

Participated · Series B · May 2026

DeepInfra operates a vertically integrated inference cloud that owns and manages GPU infrastructure across eight U.S. data centers and supports more than 190 open-source models via OpenAI-compatible APIs. The platform processes nearly five trillion tokens per week and offers enterprise features including zero data retention, SOC 2 and ISO 27001 certification. The company says its stack is optimized for high-throughput, agent-driven workloads and works closely with NVIDIA on performance optimizations including early deployment of Blackwell GPUs and NVIDIA Dynamo software. DeepInfra reported that revenue has tripled since the start of 2026 and is expanding global capacity to meet growing demand. Founded in 2022 and headquartered in Palo Alto, the company plans to use the Series B proceeds to grow compute capacity, improve developer tooling, and support next-generation models.