Untapped Ventures: 2025 Year in Review
- George Bandarian
- 1 day ago
- 5 min read
2025 was a year defined by clarity, conviction, and momentum.
At Untapped Ventures, our focus this year was on what we believe matters most at the earliest stages: understanding where AI is headed, partnering early with exceptional founders tackling hard problems, and building a community of LPs aligned around long-term outcomes rather than short-term noise. That work is the result of a collective effort across our team, founders, and LP community.
This Year in Review reflects how that work showed up across investing, research, community, and thought leadership, and how it’s shaping what comes next.
Investing with Conviction and Portfolio Momentum
Throughout the year, we continued deploying capital from Fund I with a disciplined, high-conviction approach to early-stage AI investing.
Fund I Overview:
Total fund size: $27 million
Limited partners: 99 LPs
Portfolio companies: 35 companies as of year-end 2025
New investments in 2025: 6 companies
Follow-on investments in 2025: 13 companies
Many of these investments came from founders building in the very frontiers we outlined through RFS. From early technical teams operating in stealth to companies reaching meaningful milestones, our capital followed conviction shaped by research, relationships, and real-time learning from the field.
We were fortunate to see several portfolio companies reach important milestones this year, including:
You.com closed its Series C, expanding its AI-powered search platform and deepening its enterprise offerings.
Lemurian Labs raised its Series A, scaling production of its edge AI processors designed for deployment in robotics, industrial automation, and critical infrastructure.
Extropic released its compute chip, marking a major technical milestone in physics-based computing.
Woz AI launched its Product Factory, a no-code platform that allows anyone to build real, scalable web and mobile products using AI.
These outcomes are not endpoints. They’re signals that the flywheel is working.
Mapping the AI Supercycle and Building Long-Term Trust
A core part of our work is education and long-term trust with our LP community.
In 2025, this meant consistently mapping where we are in the AI supercycle and helping LPs understand how quickly the landscape is shifting. One framework we shared throughout the year is that AI is not one technology, but a stack evolving at radically different speeds, with foundation models commoditizing rapidly while agentic systems, vertical AI applications, and AI-native infrastructure remain in their earliest innings.
We engaged directly with LPs to share how we think about timing, first principles, and conviction at the earliest stages. That work led to invitations to speak with institutional and private investor communities including Opal Group forums, YPO chapters across the United States, and Tiger 21 groups. These conversations were grounded in clarity and conviction, helping LPs understand where AI is moving from experimentation to real-world impact. The result is a deeply engaged LP community aligned around how we think, how we invest, and why we back founders early.
Content, Community, and Shared Learning
We view content, community, and gatherings as core tools for shared learning across both LPs and founders.
In 2025, we launched Season 1 of the Agentic Podcast, hosting deep conversations with founders and operators building at the frontier of AI. The season included 20 episodes and reached more than 350,000 listeners. Some of our standout conversations are shared below.
Dakota Rice & Tushar Nair (Co-founders of Harper) shared how they’re using agentic and voice AI to solve the insurance industry’s 400,000-agent shortage, delivering enterprise-grade service at 1/10th the cost. We discussed why becoming the broker unlocked software-like margins in the $85B E&S market and why Harper’s timing is strong as carriers exit high-risk states.
Richard Socher & Bryan McCann (CEO & CTO of You.com) discussed how they’re rethinking enterprise search and agentic workflows with Ari, their deep research assistant. We covered why consumer LLMs fall short for enterprise use, how You.com goes deeper than Google and faster than Perplexity, and where they believe AGI may emerge next, including simulation as a key frontier.
Season 2 is coming soon, kicking off with General Partner at a16z,
Andrew Chen, where we discuss the 2026 AI Founder Playbook and how AI is reshaping company building, growth, and early-stage conviction.
We also hosted a series of live webinars designed to give founders practical, operator-level insight into what it actually takes to build and scale AI companies. These sessions focused less on theory and more on real-world execution, featuring founders, operators, and investors actively working in the space.
One of our most successful sessions was Vibe Coaching with AI VC, featuring Will Manidis of ScienceIO, a live coaching-style conversation focused on helping founders think more clearly about building durable, defensible AI businesses.
Throughout the year, we spent time with AI founders and investors through keynotes, panels, and direct conversations at events such as Venture Summit West and the Early Stage Conference. We also hosted curated dinners and smaller gatherings in San Francisco and Los Angeles to bring founders, operators, and partners together in more intimate settings.
During SF and LA Tech Weeks, we hosted seven events that generated more than 2,500 registrations and brought together over 300 founders, investors, and operators in person. These included Day Zero, Stealth, and Pre-Seed + Seed AI Founder Dinners in San Francisco, as well as AI Founder, Investor, and Portfolio Dinners across both cities.
These moments weren’t transactional. They were about listening, sharing perspective, and building real relationships with people tackling hard, meaningful problems. That ongoing dialogue directly informs how we invest and how we support founders.
That shared learning directly shaped how we showed up for founders.
Rather than keeping our research internal, our team made it public and actionable. In 2025, we launched the Request for Startups (RFS) series, publishing our perspective across 12 AI frontiers we’re actively investing in.
Each theme reflects where we see real leverage emerging across the AI stack, from agentic systems and embodied intelligence to energy, life sciences, and critical infrastructure. The goal was simple: give founders clarity on where opportunity is emerging and where we’re excited to partner early.
RFS became a bridge between our LP and founder communities. For LPs, it offered transparency into how we form conviction. For founders, it provided a clear map of the problems and systems we believe will define the next decade.
Growing the Team
Supporting this work required growing the team behind it.
In 2025, Tripp Doll joined Untapped Ventures as a Venture Associate from the University of San Francisco. He has played a key role in deal sourcing, diligence, and portfolio support throughout the year.
We were also supported by an exceptional group of six interns across investing, operations, and marketing from institutions including UC Berkeley Haas, Chicago Booth, Yale, and Harvard. Their contributions greatly strengthened our research, diligence, and community efforts.
Looking Ahead to 2026
As we look ahead to 2026, the focus remains the same.
We’re preparing for Fund II as a natural continuation of the work we’ve been doing. The conviction, relationships, and pattern recognition developed through Fund I continue to guide how we invest, with a disciplined focus on backing founders early and staying close as they build.
I’m personally excited about where AI is heading next as systems move from experimentation into real-world deployment across critical industries.
We’re also looking forward to our 2026 Annual General Meeting, which will bring LPs, founders, and partners together to reflect, learn, and align around what comes next.
We enter 2026 with clarity, momentum, and conviction in the work ahead.
Comments