Theory Ventures

Theory Ventures — Research Report

Overview

Theory Ventures is a San Francisco-based, thesis-driven venture capital firm founded in 2023 by Tomasz Tunguz, a former Managing Director at Redpoint Ventures. The firm manages $688M in AUM across two funds — a $238M inaugural fund (April 2023) and a $450M second fund (November 2024) — and deploys concentrated capital into early-stage software companies that harness technology discontinuities to create durable market advantages. Theory Ventures invests at Seed and Series A, writing checks from $1M to $25M and expecting to take 12–15 core positions per fund.

Investment Thesis

Theory Ventures is explicitly thesis-driven: the team researches big ideas, develops informed perspectives, and backs founders from their earliest stages. The core thesis centers on The Decade of Data — the belief that organizations leveraging proprietary data to generate competitive advantages will dominate the next era of software. Theory invests in the platforms enabling this shift: data movement, transformation, analysis, observability, and AI-powered applications built on top of these data foundations.

The three primary investment theses are:

1. AI-Native Software — LLM-enabled enterprise software that replaces rule-based logic with dynamic, adaptable agents and workflows
2. Modern Data Stack Evolution — Tooling that enables organizations to store, query, and act on data at scale (vector databases, embedded analytics, data ops)
3. Web3 Data Infrastructure — Foundational blockchain data platforms that bring enterprise-grade data quality to decentralized networks

Theory explicitly targets companies turning technology discontinuities into go-to-market advantages — not incremental improvements, but software that is only possible because of a step-change in underlying technology (e.g., LLMs enabling new ERP architectures, or GPU abundance enabling vector search at petabyte scale).

Stage Focus and Check Size

• Primary stages: Seed and Series A
• Check size: $1M–$25M per initial investment
• Portfolio concentration: 12–15 core positions per fund, not a spray-and-pray approach
• Follow-on reserves: The firm reserves capital for follow-on rounds in top performers (evidenced by Series B participation in Doss, Dropzone, MotherDuck, Omni)
• Lead tendency: Theory Ventures consistently leads rounds — Tomasz Tunguz takes board seats and is operationally involved

Geographic Focus

Primarily United States (San Francisco Bay Area, New York, Seattle). The firm has made selective international investments (Initia — Singapore, Maze — UK) where thesis fit is strong. No explicit geographic exclusions stated.

Team

Tomasz Tunguz — General Partner & Founder

Tomasz is the founding GP and primary face of the firm. Before Theory, he spent 10+ years as a VC at Redpoint Ventures, rising to Managing Director, where he backed 8 unicorns including Looker (acquired by Google for $2.6B), Kustomer (acquired by Meta for ~$1B), Monte Carlo, Dremio, Hex, and MotherDuck. Prior to VC, he was a product manager at Google overseeing a billion-dollar business unit within AdSense. He writes a widely-read blog at tomtunguz.com covering AI, data, and software trends. Education: Dartmouth University. He focuses on data infrastructure, B2B SaaS, and AI applications, and covers portfolio companies including Context, Superlinked, Datable, Tobiko Data, Omni, and LanceDB.

Spencer Farrar — Partner

Spencer focuses on web3 infrastructure and early-stage data investments. He previously worked at a $4B asset management firm alongside Tomasz, where they co-invested in a decacorn. He also gained startup experience building an options trading platform. His investment focus includes Initia (web3 Layer 1, decacorn), Allium (blockchain data), and Superlinked (vector data). Education: UC Berkeley (dual-sport athlete; 4x PAC-12 All-Academic; NCAA Championship; Bronze Medal, World University Games representing Team USA).

Andy Triedman — Partner

Andy specializes in ML infrastructure and AI-native enterprise software. Before Theory, he was an investor at Innovation Endeavors (GV / Eric Schmidt's fund), backing ML-focused companies including Afresh, Third Wave Automation, ClearMetal, and Replica. He was also the first product manager at Replica (generative models / synthetic urban data), backed by Founders Fund. He began his career at Bain & Company. Education: Brown University (Computational Neuroscience — built brain-computer interfaces). His portfolio coverage includes Dropzone, Doss, Aampe, Maze, and LanceDB.

Lauren DeMeuse — Partner & COO

Lauren joined at Theory's founding to build an institutional-grade operations infrastructure. Previously she was CEO of Roam Analytics (NLP platform for healthcare, acquired by Parexel International), an EIR at 8VC, and held multiple senior roles at Palantir including Chief of Staff to the CIO and Head of Business Development for Healthcare. Education: Stanford GSB (MBA), Princeton (BA cum laude).

Bryan Bischof — Head of AI

Leads Theory's AI platform capabilities and internal AI tooling.

Kristin Beach — CFO

Manages fund finance and operations.

Portfolio Overview

Theory Ventures has built a concentrated, high-conviction portfolio of ~16 companies across four thematic areas:

AI-Native Enterprise Software:

• Aampe (aampe.com) — Agentic infrastructure for personalization; 100M+ AI agents deployed for consumer app personalization. $18M Series A (Dec 2024), led by Theory.
• BackOps (backops.ai) — AI operations platform automating supply chain workflows via integration with existing WMS/ERP/CRM systems.
• Doss (doss.com) — AI-powered ERP built with composable architecture optimized for both human and LLM interaction. $18M Series A (Apr 2025) led by Theory; subsequently raised $55M Series B.
• Koah (koah.ai) — Native advertising infrastructure for conversational AI ("DoubleClick for AI apps"). $20.5M Series A (Feb 2026) led by Theory.

AI Security:

• Dropzone AI (dropzone.ai) — AI SOC analyst; autonomous investigation and threat-hunting agents. $37M Series B (Jul 2025) led by Theory. 11x ARR growth in 2025.
• Maze (mazehq.com) — Agentic AI for vulnerability management; cloud security breach prevention. $21.8M Series A (2025).
• Artemis (artemissecurity.com) — AI-native detection engine for security operations; dynamic behavioral models across users, machines, and cloud workloads. $70M Series A (Apr 2026), Theory led.

Modern Data Stack:

• LanceDB (lancedb.com) — Developer-friendly embedded multimodal vector database built on the Lance columnar format. Series A.
• MotherDuck (motherduck.com) — Serverless cloud offering built on DuckDB; founded by Jordan Tigani (founding engineer of Google BigQuery). Series B.
• Omni (omni.co) — Business intelligence and embedded analytics platform with AI, SQL, and no-code interfaces. $120M Series B (Apr 2026).
• Superlinked (superlinked.com) — Vector computing infrastructure for AI applications; simplifies data into vectors for AI systems. Seed.
• Tobiko Data (tobikodata.com) — Data operations tooling (acquired Sep 2025).
• Datable — Modern data stack (acquired).

Web3 / Blockchain Data:

• Allium (allium.so) — Enterprise blockchain data infrastructure; real-time and historical data APIs across 40+ decentralized networks. Series A.
• Initia — Web3 Layer 1 blockchain; last valued at $350M (decacorn trajectory). IPO/public.

Exited:

• Context (acquired) — AI company.
• Tobiko Data (acquired Sep 2025) — Data operations.

Recent Activity

Theory Ventures is actively deploying from Fund II ($450M, closed Nov 2024). The pace is notably high for 2025–2026:

• Feb 2026: Led Koah $20.5M Series A
• Apr 2026: Led Artemis $70M Series A
• Apr 2026: Participated in Omni $120M Series B
• Jul 2025: Led Dropzone AI $37M Series B
• Sep 2025: Tobiko Data exit
• Apr 2025: Led Doss $18M Series A
• Dec 2024: Led Aampe $18M Series A

Decision Process and Founder Preferences

Theory operates as a partnership with three investment partners (Tomasz, Spencer, Andy). Tomasz as founder GP likely has final say, but there is genuine collaboration across the team. Andy Triedman describes a focus on founders with "extreme levels of determination, attention to detail, and unique customer insight." Spencer Farrar prizes responsive communication and strong founder relationships.

The firm publishes detailed investment thesis essays ("Theories") on their website and Tomasz maintains a prolific blog — founders are expected to engage with the intellectual framework. Theory does not require warm introductions but thesis alignment is essential. They prefer to be first institutional money and to lead rounds.

Anti-Thesis

Theory does not invest in:

• Consumer products without strong data or AI moats
• Hardware (except software controlling hardware)
• Non-software categories (biotech, climate hardware, etc.)
• Late-stage growth (Series C+) except in follow-on context

Notable Exits and Track Record

At Theory:

• Tobiko Data (acquired Sep 2025)
• Context (acquired)
• Datable (acquired)

At Redpoint (Tomasz's prior track record):

• Looker → Google ($2.6B, 2019)
• Kustomer → Meta (~$1B, 2020)
• Monte Carlo, Dremio, Hex, MotherDuck (all unicorns)

Frequent Co-Investors

Based on portfolio company announcements:

• Z47 (Aampe)
• Innovation Endeavors (likely)
• ICONIQ Capital (Omni Series B)
• Battery Ventures
• Andreessen Horowitz (a16z)