A man faces the realistic artist” robot “Ai-Da” using artificial intelligence at a stand during the ITU AI for Good Global Summit in Geneva – © AFP/File Alain JOCARD
What will be the defining stories in the venture capital world from a top investor’s perspective in 2026? To help answer this question, Digital Journal spoke with Mike Collins, CEO of Alumni Ventures,
Alumni Ventures is one of the most active venture capital firms in the world. The firm has over $1.5 billion in assets, and the firm is ranked #20 on Time Magazine’s Top VC firms for 2025.
Digital Journal: What are the most investable opportunities within the AI value chain right now?
Mike Collins: The short answer: not where the headlines are.
Most people obsess over foundation models and leaderboard scores. That’s interesting, but it’s not where most of the returns will come from. Our team at Alumni Ventures sees these three parts of the AI value chain that are especially attractive right now.
Picks, shovels, and plumbing (AI infrastructure and tooling)
The world is underestimating how much raw compute, cooling, and infrastructure this AI wave will need. Companies like Rigetti (quantum computing) and Lambda (GPU cloud) are tackling the hard infrastructure problems every serious AI company will eventually run into. We also love enabling tech like Frore Systems with solid-state cooling — the boring-sounding stuff that quietly becomes essential.
Vertical AI apps with clear ROI, not science projects
We’re especially excited about AI that is deeply embedded in real industries and workflows. Think Insitro, using physics-based algorithms and huge biological datasets to rewire drug discovery timelines. Or Lincode, using computer vision for precision quality inspections in manufacturing with customers like BMW and Boeing. These companies have three things we care a lot about: real customers, tangible ROI, and proprietary data and domain expertise that make them hard to copy.
“AI Plus” – the human and hardware layer
There’s a massive gap between what AI can theoretically do and what organizations can actually implement. That gap is where a lot of value will be created. We’re drawn to businesses that help enterprises adopt AI in practical ways — combining hardware and AI, or humans and AI, into new workflows and systems. We call this “AI plus”: AI plus people, AI plus data, AI plus real-world operations.
In this phase of the AI cycle, we’re less interested in “wow” demos and much more interested in who is quietly becoming indispensable infrastructure. You can see more of our AI investment portfolio at AV.VC.
DJ: How should investors think about defensibility in an era where AI capabilities are rapidly commoditizing?
Collins: This is the $64,000 question — and a lot of people are asking it the wrong way.
Everyone wants to know: Who has the best model? Who’s ahead? Who’s behind? That’s the wrong scoreboard. Models will keep getting better and cheaper across the board. We’ve seen this movie before with cloud computing and mobile.
Here’s how we think about defensibility at Alumni Ventures.
Data moats beat model moats
Models are converging; differentiated datasets are not. We want companies that own or generate proprietary data that compounds over time. Take one of our portfolio companies (that is confidential), for example. It didn’t win in blockchain analytics because its AI is magically better than everyone else’s. It won by aggregating and enriching a huge, hard-to-replicate dataset that financial institutions, governments, and crypto businesses — from PayPal and Visa to agencies like the FBI — now depend on for risk detection and compliance.
Distribution, integration, and workflow lock-in
In a world where “good enough AI” is everywhere, the winners will be the companies that deeply understand how customers actually work. That’s why we like businesses like Cents, which layers AI on top of a sticky vertical SaaS platform. When you become part of the workflow — not just another tool — the switching costs get very high.
Trust, regulation, and long-term permission to operate
In regulated industries like healthcare, finance, and defense, navigating the regulatory maze becomes a serious moat. A company like Synchron — developer of brain-computer interfaces — earns a multiyear head start simply by doing the hard, slow work of building clinical evidence and working through approvals.
Big, tough, “you almost couldn’t do this pre-AI” problems
We love founders who tackle problems that felt basically impossible before AI. One of the most common investor mistakes is a lack of imagination about what’s now possible. If a team is using AI to unlock a market or solve a problem that simply didn’t pencil before, that gets our attention.
For accredited investors interested in seeing how we regularly assess defensibility, I’d encourage them to plug into our deal flow by joining our syndicate or attending a webinar. Both free to join, no obligation to invest.
DJ: What differentiates a fundable AI company in 2026 compared to one in the past?
Collins: The bar has gone way up.
Things that used to take 50 people, three years, and $10 million can now be done by three people in six months with $1 million. That’s great for founders and investors — but it also means expectations are higher.
Here’s what we look for now.
A business that rides the wave, not one that gets crushed by it
We want companies that benefit as AI and compute get better and cheaper, not ones that are easily replaced by the next model release. If your product depends on being the only one with access to a capability that’s clearly going to be widely available soon, that’s fragile.
Unit economics that work at scale
This hasn’t changed. Can you make money per customer in a way that improves, not worsens, as you grow? We want to see a clear path from “cool tech” to “durable business.”
Evidence of real product–market fit, not just hype
We’re looking past the early-enthusiast crowd. Are customers coming back? Are they expanding usage? Is there pain if they try to churn? AI companies now need to show traction earlier because the cost and time to build the first version are lower.
And ultimately: it’s the team, stupid
This part never changes. At Alumni Ventures, we’ve backed thousands of founders, and the throughline is always the same: resilient, curious, execution-focused teams win. We care about how fast you learn, how quickly you ship, and how you respond when the world changes around you — because it will.
DJ: Beyond AI, which emerging sectors are most underpriced by the market heading into 2026?
Collins: Three big areas stand out for us: Energy, Space, and Health. If folks are interested in these or others, we have a free-to-join Syndicate where you can see the deals we do and invest if they are accredited.
Energy: the next great infrastructure buildout
The energy load from data centers and AI compute is colliding with decarbonization goals. This is not a small adjustment — it’s a fundamental rebuild of the grid. That’s why we’re excited about companies like X-energy with its advanced modular reactors and SHINE Technologies that produces fusion-based isotopes with a roadmap to fusion power. As hyperscalers like Amazon and Microsoft sign long-term deals tied to nuclear and clean energy, the market is still catching up to how big this buildout will be.
Space: from getting to orbit to living and working there
Launch has gotten dramatically cheaper. Now the bottleneck is what you can do once you’re in orbit— satellite deployment, servicing, manufacturing, and more. Companies like Impulse Space that offers orbital transfer and transportation and Axiom Space producer of commercial space stations are building the infrastructure for an in-space economy. We see this as a multi-decade platform, not a niche.
Health: compressing the timeline between insight and impact.
The combination of AI, biology, and demographics is incredibly powerful. We mentioned Insitro earlier. It’s a good example of how large biological datasets and physics-based AI models can compress drug development timelines from decades into years. As societies age, the economic value of extending healthspan, not just lifespan, is enormous.
What ties these sectors together? They’re capital intensive, long duration, and temporarily out of fashion during the “software eats the world” phase. Now they’re front and center for solving real, existential challenges — and that’s where we like to be invested.
DJ: How are global macro and geopolitical tensions shaping venture flows in 2026?
Collins: Geopolitics has moved from background noise to a major driver of where capital flows. We see three big patterns.
Onshoring and “friendshoring” are turning into real opportunity
With the CHIPS Act and similar policies around the world, countries are pushing to rebuild or “friendshore” critical manufacturing and supply chains. This is showing up in semiconductors, advanced manufacturing, defense tech, and logistics. What was once considered unsexy “industrial” is now strategic infrastructure. Defense tech, in particular, has gone from taboo to mainstream in just a few years. We’ve noted that ourselves, with our Strategic Tech Fund now one of our most popular.
Capital is fragmenting into parallel ecosystems
The easy, open flow of capital between the U.S. and China is largely over for anything touching critical tech. That’s forcing separate innovation ecosystems to mature in parallel. We’re also seeing secondary hubs — Japan, The Middle East, and emerging European centers — attract more attention and talent. For investors who can navigate multiple regions, this fragmentation creates more ways to win.
The rise of “tech sovereignty”
Every major economy wants its own AI stack, semiconductor capabilities, and clean energy infrastructure. That’s creating local champions and local regulatory complexity at the same time. The flip side: more countries are willing to partner with private companies to build this stuff.
The big shift: we’re moving from a mostly U.S.-centric innovation map to a multipolar one. For a global investor like Alumni Ventures, that means more edges, more nuance — and more opportunity if you know how to underwrite across different regimes.
DJ: How do you expect the fundraising environment for venture funds to evolve in 2026?
Collins: The fundraising reset we’ve been living through is painful for some, healthy for the system. Here’s how we see it.
Many institutional LPs ended up overallocated to venture after the 2020–2021 boom. They’re rebalancing. But this is less “we’re done with venture” and more “we’re done with undifferentiated venture.” The best-performing managers are still raising and, in many cases, oversubscribed. The middle is getting squeezed.
We’re seeing a split between large, established funds raising $1B+ and smaller, specialized funds (often sub-$200M) with a clear edge in a sector, stage, or network.
The “tweener” funds in the $300–700M range without a crisp story of why they exist are having the hardest time.
We’re not competing for the same institutional capital pool. Our diversified base of 11,000+ individual investors gives us a very different funding engine. That’s something many traditional managers envy in this environment: a broad, engaged LP base instead of a handful of gatekeepers.
GPs are increasingly using continuation funds and structured secondaries to provide partial liquidity while holding on to their best assets longer. That blurs the line between primary and secondary — and opens up more creative ways to serve LPs.
My view: 2026 will clearly separate managers who built durable businesses from those who simply raised funds in easy times. We fully intend to be in the first camp.
DJ: What role will secondary markets play in providing liquidity to venture investors in the new year?
Collins: Secondary markets are finally moving from side show to core infrastructure in venture.
For years, the traditional venture model relied on IPOs and big M&A to solve the liquidity problem. With the IPO window narrow and M&A under heavier antitrust scrutiny, secondaries are stepping into the spotlight.
Three shifts we’re watching.
Institutional secondary funds are scaling up
Specialized secondary funds are raising meaningful pools of capital to buy LP stakes and provide liquidity earlier in the fund lifecycle. That gives investors a pressure-release valve without forcing GPs to rush exits.
Company-level secondaries are becoming standard
As companies stay private longer, employees and early investors reasonably want liquidity before a full exit. Structured secondary programs — often managed via dedicated platforms — are becoming a normal part of the growth journey rather than a sign of distress.
Tokenization and new rails might be just over the horizon
The ability to fractionalize, package, and trade private exposure on blockchain rails could eventually lower friction for secondary transactions and expand who can participate. It’s not all the way there yet, but the direction is clear.
For our investors and syndicate members at Alumni Ventures, we view secondary markets as both an opportunity and a tool. Opportunity, because we can sometimes buy into great companies with more information and at attractive prices. Tool, because we can potentially offer some liquidity earlier in the journey — while still aiming for long-term outcomes.
Big picture: as venture matures into a true asset class, robust secondary markets are not optional — they’re required. I think 2026 is the year this infrastructure really scales.
Overall, our goal at Alumni Ventures remains building the largest venture network in the world. While we have our eyes on 2026, we’re playing a much longer game.
