Venturing on the frontier
David Peterson

AI has both concentrated conviction and driven disorientation in venture capital. Every VC wants exposure to the latest AI winner, but the consensus plays are expensive. Some investors are paying up. Others are searching for the AI-native plays that are less crowded. 

Others still are rushing into novel territory: AI rollups of traditional businesses, deep tech and hard science, vertically-integrated infrastructure, AI-native services firms, and distressed tech assets.

There are real opportunities in all these categories. (We’ve invested in some!) But many venture investors are entering these markets with the wrong mental model.

For 15 years, software venture capital operated under a simple rule: "if the founder is great and the product works, there will always be capital available." That worked because software was capital-light, margins were high, and every round came from a fund that was just a larger version of the last one.

This approach breaks down on the new frontiers I’ve listed above for an unexpected reason: the next check isn't necessarily a VC check. It might come from an infrastructure lender, a sovereign wealth fund doing project finance, a government loan program, an industrial offtaker, or a specialty insurer. These financiers don't care about NRR or CAC or that sexy launch video you just posted on X. They care about offtake certainty, debt service coverage ratios, feedstock pricing, and regulatory approvals.

VCs may be some of the most well-networked financiers in the world, but only with each other. They don't know the infrastructure debt syndicate in Houston or the commodity traders who provide prepayment financing. This means many VCs are now underwriting companies where financing is the real risk, while having almost no ability to evaluate whether that financing will actually materialize. Even worse, they’ve grown accustomed to easy VC-driven fundraising cycles and are often unprepared when fundraising pathways are more fraught.

The New Model: Founders as Capital Architects

To win in these new frontier markets, founders will often need to devote as much attention to designing their capital stack as they do to product, engineering, and go-to-market. Their competitive advantage will be building products in forms that unlock cheaper, earlier, or more patient capital.

Consider Venture Global LNG: They didn't invent liquid natural gas technology. They redesigned the export terminal itself using modular components manufactured off-site. Construction time dropped from 5-7 years to 3 years, execution risk plummeted, and suddenly banks could underwrite it like infrastructure, not a moonshot. Result: $13+ billion in project financing unlocked, $5 billion in annual revenue, $37 billion market cap, on track to be the largest LNG player in the US.

Another great example is Solugen: They designed their "Bioforge" bio-manufacturing facilities specifically to qualify for DOE loan guarantees by meeting infrastructure standards (e.g. standardized components, predictable costs, contracted offtake). In 2024, they secured a $213.6 million DOE loan guarantee, the largest U.S. government investment in biomanufacturing in decades.

In both cases, novel technology was crucial in unlocking the opportunity, but it wasn’t enough. The insight that allowed both Venture Global LNG and Solugen to win was that the business could be architected to unlock non-venture financing. That was the true breakthrough. This is an inversion of the classic relationships between technology and finance. In traditional VC, finance enables technology. On the frontier? Sometimes technology enables financing.

Here’s another way to frame it. Venture capital is normally thought of as funding technical breakthroughs that unlock growth. Venture capital on the frontier is about funding capital coordination problems disguised as technology companies. 

For more detail on what doing venture on the frontier means, take a look at this deep dive with case studies, an evaluation framework, and more on why this moment is both dangerous and full of opportunity.

At Angular, we're looking for founders who see the world this way. Founders who don't just pitch the technology, but who can articulate the capital path. Who understand that innovation isn't just in what you build, but in how you make it financeable. If you're building in any of these frontier categories and think about financing this way, we want to talk to you.

The AI-Native Enterprise Playbook
Ten real-time observations on a rapidly evolving playing field

No More Painting by Numbers
It’s the end of the “SaaS playbook.

The Age of Artisanal Software May Finally be Over
Every wave of technological innovation has been catalyzed by the cost of something expensive trending to zero. Now that’s happening to software.

Founders as Experiment Designers
David on why founders should run everything as an experiment.

ENTERPRISE/TECH NEWS

Echos of a bubble. There are not a few learnings from previous bubbles in this piece from Eric Newcomber on why VCs are not jumping at the chance to invest in the AI neoclouds. "Taken together, this group has been a top beneficiary of the AI industry’s mad scramble for cloud computing and the electricity needed to power it. As one researcher explained to me: “The only thing that matters right now is turning electricity into intelligence.” And yet, the question remains whether these are good long-term bets. VCs largely think not. The list of top tier firms that have backed AI data center startups is relatively short. Bain and Founders Fund are in Crusoe. Accel is in Nebius. Bloomberg Beta was in Lambda. An investor at one of the world’s largest asset managers isn’t very bullish either, noting that beyond their relationships with chipmakers and access to land, power, and capital, there’s no technological differentiation among them. “The only difference is, can you raise money faster than the next guy,” he said. “At some point if you raise more money, then you get contracts. If you don’t raise enough money, then why do I care?”

AI after OpenAI. Michael Parekh analyzes how Microsoft is planning for its AI future after OpenAI. He draws a lot on a WSJ article on the same topic. “Though still close partners, Microsoft and OpenAI now compete in a number of ways. OpenAI is building its own data centers and striking partnerships with Microsoft rivals Amazon.com and Oracle. OpenAI also has an enterprise product that now accounts for roughly 40% of revenue, up from 30% at the beginning of the year, Chief Financial Officer Sarah Friar told the Journal Wednesday. Microsoft’s Copilot chatbot relies heavily on OpenAI, but Microsoft is building, testing and releasing its own voice, image and text models. Microsoft’s updated contract with OpenAI, which was announced last week and will give Microsoft a 27% stake in the startup’s new public-benefit corporation, allows the tech giant to pursue artificial general intelligence, or AI as capable as humans, on its own. Microsoft has access to OpenAI’s models until 2032, a schedule Suleyman said gives his team time to make its models into leading technology.”

HOW TO STARTUP

Make friends with fear. The inimitable Fred Destin argued in a linkedin post that founders should focus on getting to know their fears in order to manage and conquet them. “Fear of failure will diminish you as a founder. You will make worse decisions, you won't be able to act or work as effectively, it will physiologically impact you. If you have ever felt the fear move from your belly to your chest, you know what I am talking about. Most founders don't even want to envisage that their company could crash, as if simply thinking about it was already an acceptance of defeat. And yet failure is a real and often probable scenario. Not wanting to look at it makes the fear murky and diffuse. Instead, you want to bring it into the light. It is much better to actually visualise the outcome. If you are intimate with fear it loses its control over you.”

You’ll need to change. Kyle Poyar argues that founders typically need to change in order grow: "The winners didn’t necessarily start better. They became better. They changed their stripes, improving their revenue mix, monetization power, and retention. And they watched as the improvements kept compounding. Chili Piper co-founder Alina Vandenberghe had a similar realization. She told me they needed to let go of everything that propelled the startup to its first $1.5 million, reinventing the business for the next stage of growth."

HOW TO VENTURE

Revenue reset. Kudos to Pat McGovern at Bowery for telling it like it is. “It’s time for a reset on early-stage revenue expectations. Here’s the current state of play: VC’s publish select, unrepresentative examples of super high-growth companies, largely ones that are prosumer or selling to other startups. And then message that any startup not growing $0→$5MM in Year 1 isn’t interesting. Founders overcorrect in response. They inflate ARR in their decks, claim one-time pilots are recurring, fudge selling months, and position estimated expansions and unsigned revenue as ARR. VC’s get mad that founders are lying to them (and, in the process, committing borderline securities fraud).”

Blue Energy CEO and Co-Founder Jake Jurewicz named to the American Nuclear Society’s 2025 Nuclear News 40 Under 40 list.

Groundcover unveils industry’s first automated observability migration tool to unleash teams from legacy vendor lock-in.

CruxOCM
Software Engineer - Tools (Vancouver, Remote)

Groundcover
Growth Marketing Manager (San Francisco)

Sourcix
R&D Group Manager (Tel Aviv)