Andrew Peng - Collide Capital
"Great founders are everywhere. It’s just that venture access is not."
In 2021, Andrew Peng was spending his evenings in Discord servers, attending NFT parties, and testing blockchain platforms. He was a first-year MBA student interning at Alpaca VC, and he had decided to go deep on the metaverse because, as he put it, it was early enough that nobody was the expert quite yet. The result was a 60-slide thesis that mapped the emerging space and identified which major tech incumbents had the most at stake. One slide argued that Facebook had “the most to gain and the most to lose” from the metaverse and was therefore the most likely to reposition itself around it. Two months later, Meta launched.
The thesis spread across Twitter, Reddit, and Discord. That moment of virality produced three concrete outcomes: Andrew sourced his first self-originated venture deal, Alpaca’s participation in The Sandbox’s Series B, after the founder reached out having apparently seen the thesis; he landed a full-time role at Alpaca; and he brought in his first LP relationship, a Korean conglomerate looking for managers with a developed view on metaverse and blockchain real estate. He is no longer bullish on the metaverse. But the lesson from that experience became the operating system for everything that followed.
“The lesson wasn’t about the metaverse. It was about how to research a market, how to find founders, how to articulate a point of view, how to earn credibility.”
Andrew believes the most durable edge in venture comes from doing the work others skip before the market prices the opportunity. This conviction informs his thesis on industrial resilience, his sourcing strategy, and his standard for the founders he backs.
From Coal Towns to Cap Tables
The son of immigrants from Shanghai, Andrew grew up across Ohio, Indiana, the mountains of West Virginia, and eventually the DC suburbs. That early geography, he says, informed a lot of his early career choices. He began his career at a boutique consulting firm before leaving at 24 to build out the consulting practice at Smart Growth America, a national nonprofit. There, he advised local governments across more than 40 states on innovation and growth—not just in what he calls “sexy downtowns,” but in places like the ex-coal towns of West Virginia still searching for a 21st-century identity.
He eventually moved into real estate private equity, running acquisitions at a Black-owned firm in DC and closing over $500 million in deals across office, multifamily, and mixed-use properties. That volume was clarifying. “After enough deal volume, you start to see the patterns. I loved deploying capital to shape markets, but I wanted to be closer to the founders and technologies changing how those markets actually work.” He felt farther from innovation. He enrolled in a dual-degree program at Wharton and Harvard Kennedy School and joined Alpaca VC as an intern. He spent four years there as principal and head of research before joining Collide Capital roughly six months ago.
Collide launched its first fund, a $66 million vehicle, in 2022. Fund 2, announced on the homepage of TechCrunch, came in at $95 million, a 50 percent increase, with most original LPs recommitting. The firm is led by two general partners: Aaron, founder of Blavity and AfroTech, the largest Black tech conference globally, who brings the operator lens, and Brian, who came from growth equity at Goldman Sachs and brings institutional investor discipline. The firm invests pre-seed through Series A, focused on fintech, supply chain infrastructure, and the future of work as Gen Z reshapes workforce expectations around communication, flexibility, and trust.
Research Upstream of Conviction
The metaverse thesis matters not because of what it predicted but because of how it was built. For Andrew, research sits upstream of conviction, not downstream of hype. The typical motion looks like this: a theme gets hot, a partner flags it for a house view, someone produces a landscape mapping who is building what, and the work stops before reaching the “so what.“
“By then, maybe the best founders are already funded and prices have moved,” he says. “A lot of research I see is post-rationalization dressed up as thesis formation.”
His version starts before the market has formed a consensus. “If the research doesn’t change who I meet, or what I believe, or what I invest in, then it’s just desk-based, it’s just content, and it’s not connected to actual proprietary sourcing.” As an illustration of how that works in practice, he points to the work he is currently doing on the software development lifecycle. Before forming a thesis, he is already talking directly to engineers and product managers about specific gaps, whether in penetration testing, security workflows, or other underserved parts of the development process. By the time he is sitting across from a founder, he often has relevant customer introductions already in hand. Research and sourcing become the same motion.
This approach extends to how he thinks about the metaverse story in hindsight. He draws a parallel to the dot-com era: “Most of the predictions from the dot-com bubble have come true. People were right. They just weren’t right on timing.” The durability of the lesson matters more than whether the specific call aged well.
“You don’t need to be right forever. You need to show how rigorously you got to conviction at the time.”
The Physical Constraints Software Cannot Fix
Andrew’s most developed market view centers on what he calls industrial resilience, a category built on the premise that the next decade’s bottlenecks will not be solved by writing better code. His years in real estate private equity gave him a ground-level view of an industry that resisted software for a century. He maps that experience onto a broader class of markets: outdated, offline, regulated, or structurally fragmented industries where new enabling technologies are finally creating a wedge.
His specific focus within supply chain and industrial infrastructure is the physical layer that underpins every AI application. Power availability, grid interconnection capacity, rare earth magnets, copper, load forecasting, domestic manufacturing supply, and permitting are the inputs that determine whether AI ambitions scale in practice.
“Anthropic can make software more powerful, but it cannot magically create more substations or faster permitting or make more rare earth magnets or increase our domestic manufacturing supply.”
What makes these opportunities venture-backable, in Andrew’s framework, is a wedge that avoids pure project finance. He looks for repeatable models with some form of nonlinear scaling: proprietary IP, marketplace dynamics with supply-demand liquidity, hardware-software plays, or distributed network models. He also looks closely at founder background. Through pattern recognition across construction tech deals he tracked at Alpaca, he identified that the strongest signal for reaching a Series B in that category was whether the founder came from the construction industry itself. Not pedigree, not cap table composition. Domain origin. Founders who come from the industry bring the relationships, the regulatory expertise, and the technical knowledge that outside competitors cannot easily replicate.
Access Is the Alpha
Collide’s sourcing thesis starts from a straightforward premise: if great products can be built from anywhere, great founders can come from anywhere. If VC firms keep sourcing from the same demo days, accelerators, and networks, they are underwriting consensus. The firm’s Collide Campus Program annually deploys roughly two dozen undergraduate scouts a year at schools like Johns Hopkins University, Vanderbilt, Columbia, and the University of Michigan. These are schools with unicorn outcomes that Andrew believes are dramatically undercovered by institutional venture.
“By the time they have a good idea, I want them to think of us. I want Collide to be the number one name on campus.”
The diversity outcome follows naturally from the access strategy. Collide’s Fund 1 portfolio ended up more than 80 percent led by women, Latin, and Black founders. Andrew is clear that this was not a stated goal. “It’s an output of our strategy,” he says.
“If everyone is sourcing from the same networks and the same schools, you’re going to get the same founder archetype. Great founders are everywhere. It’s just that venture access is not.”
The standard he holds founders to reflects the same principle he applies to himself. He is not looking for polished pitch structures or optimized narrative flows. He is looking for founders who have suffered with the problem and learned something not available from a Google search.
“Don’t just perform venture. Show me the urgency. Show me the pain. Show me what you’ve learned that other people haven’t.”
Embracing the Existential Crisis of VC
When the conversation turned to what he does outside of work, Andrew’s answer was not what most might expect. He goes skydiving and bungee jumping despite being afraid of heights. He scuba dives despite being afraid of open water.
“You have to remember that you can die. That’s what forces you to prioritize why you’re living.”
He did not connect that directly to his work in venture, but the parallel is hard to miss. The same willingness to go somewhere uncomfortable and stay there runs through everything he has described about how he invests.
Andrew describes venture as going through an existential crisis. LPs who feel overcommitted are pulling back, and the result is zombie funds and firms raising smaller consecutive vehicles than they did before. Collide is not in that position. Fund 2 is 50% larger than Fund 1, and that capital, in Andrew’s view, arrives at exactly the right moment. The market is bifurcated. AI is absorbing a disproportionate share of attention and dollars, and consensus opportunities are expensive. What remains, in the places others are not looking, is where he has always done his best work.
“It’s not your job to win every obvious mega round,” he says. “The job of a small firm is to find overlooked markets and founders where the next repricing hasn’t happened yet.”
More from Uncovered Media
 | A guest post by
|