The AI Bubble Question: How Serious Investors Are Positioning for Either Outcome

The most important debate in venture capital right now is not which AI company to back. It is whether the entire AI boom is a bubble.

It is a fair question, and the smartest investors are not answering it with a confident yes or no. They are doing something more useful. They are positioning their portfolios so that they are protected if it is a bubble and still exposed to the upside if it is not. That is a fundamentally different exercise from picking a side in an argument, and it is the exercise that separates disciplined capital from the crowd.

This post lays out the bear case honestly, gives the bull case a fair hearing, and then focuses on the part that actually matters: the single test that both sides quietly agree on, and what it means for how you should be deploying capital regardless of which way the debate resolves.

The Bear Case: Circular Financing and Sobering Economics

The most structurally concerning feature of the current AI boom is not high valuations. High valuations are a normal feature of every technology cycle. The concern is circular financing, and it is worth understanding precisely.

Circular financing is when a small group of chipmakers and cloud providers invest in AI companies that then spend that money buying the investors' own chips and cloud capacity. The cash loops among a handful of interconnected firms, which can make demand look organic and revenue look robust when much of it is the same dollars going around the circle. By 2026, analysts have identified more than $800 billion in these arrangements across the AI supply chain.

The web is dense. Nvidia invested in the leading AI developers and then became their largest supplier, committing enormous sums to companies that spend much of it on Nvidia hardware. Major cloud providers hold significant equity stakes in the frontier labs while also serving as their primary infrastructure vendors, generating cloud revenue that flows back into chip purchases. Chipmakers and AI labs have taken cross-stakes in one another while placing multibillion-dollar orders in both directions. Amazon is a major investor in Anthropic, which uses Amazon Web Services as its primary cloud provider. The same pattern repeats across nearly every major relationship in the AI economy.

GMO analysts have described this arrangement as reminiscent of the circular financing of the internet bubble. The comparison to the late-1990s telecom vendor-financing schemes, which collapsed when real-world usage failed to match artificially inflated revenue, is being made by serious people, not just perma-bears.

The underlying economics are sobering. OpenAI is reportedly on track to lose around $14 billion in 2026, nearly triple its 2025 losses, even as it projects $100 billion in revenue by 2029. By one analyst's framing, the AI industry currently burns roughly $400 billion per year while generating between $50 billion and $60 billion in revenue. And an MIT Media Lab report found that despite $30 to $40 billion in enterprise investment in generative AI, 95% of organizations reported zero measurable return.

Two features distinguish this from the dot-com era in ways that should concern investors. The first is debt. A meaningful portion of the AI infrastructure buildout is debt-financed, which creates systemic risk that pure equity speculation does not. The second is the power constraint. Grid limitations are delaying data center construction, which pushes revenue timelines further out while debt payments remain due on the original schedule.

Even Sam Altman has acknowledged that someone is going to lose a phenomenal amount of money.

The Bull Case Deserves a Fair Hearing

A disciplined investor does not stop at the bear case, because the bull case is also grounded in real evidence.

The most important counterpoint is that, unlike the dot-com bubble, the leading AI firms are generating actual revenue and driving measurable economic output. This is not 1999, when companies with no revenue and no path to it commanded enormous valuations. The frontier labs have real, large, fast-growing revenue, even if their costs are larger still. Jerome Powell has made exactly this point in distinguishing the current environment from prior bubbles.

The balance sheet backdrop is also far healthier than it was in 1999. Morgan Stanley has noted that corporate cash flow today is roughly triple its level during the dot-com peak, which gives the companies driving this cycle substantially more buffer to absorb a downturn without a cascade of failures.

And the underlying demand may genuinely justify the buildout. Data center demand is projected to grow more than 19% annually through 2030. The capital-intensive infrastructure being built, even if some of it is funded through circular arrangements, is producing real physical capacity that has real utility. Vendor financing has built genuine capital-intensive industries before, from railroads to telecom. The infrastructure outlives the financing arrangements that funded it.

The honest position is that both cases are strong. The boom has features of a bubble and features of a durable technological transformation at the same time, which is precisely why confident predictions in either direction are a warning sign rather than a source of insight.

The One Test Both Sides Agree On

Here is where the debate becomes actually useful for an investor, because underneath the disagreement, both the bulls and the bears converge on a single question.

The deciding factor for whether the AI boom is sustainable is how much revenue comes from outside the circle.

This is the point that cuts through everything. Circular financing is only dangerous to the extent that the revenue is circular. If AI companies are ultimately selling to real external customers, businesses and consumers outside the Big Tech circle, who pay real money for real value, then the infrastructure buildout is justified and the financing arrangements are just an efficient way to have pre-funded it. If the revenue never escapes the circle, then it is an accounting loop that flatters everyone's numbers until the moment the next investment round stalls.

As one analysis put it, for this ecosystem to be sustainable, AI companies must eventually generate revenue from external clients outside the Big Tech circle to pay off their massive infrastructure debts. The transition from capital-fueled growth to genuine, customer-driven utility is the ultimate test.

That single insight is the most valuable thing an investor can take from the entire bubble debate. It reframes the question from an unanswerable macro prediction into a concrete, evaluable company-level criterion. You cannot know whether the aggregate AI market is a bubble. You can absolutely know whether a specific company's revenue comes from real external customers who are not part of the circular financing web.

What This Means for Early-Stage Investors

The implication of the "revenue from outside the circle" test is direct and actionable. It tells you exactly what to prioritize in your own portfolio, regardless of how the macro debate resolves.

Underwrite external revenue above all else. The single most important protective quality in an AI investment right now is revenue from customers who are outside the circular financing dynamics of the frontier labs and their infrastructure partners. A company selling AI-powered software to real businesses that pay for it because it solves a real problem is insulated from the circular financing risk that threatens the core of the AI economy. That external, durable revenue is the hedge.

Prioritize businesses that would survive a funding winter. The circular financing risk is fundamentally a refinancing risk. Companies with enormous burn rates that must raise again soon are the ones exposed if sentiment turns and lenders demand revenue proof. Companies with reasonable burn, real revenue, and a path to profitability are the ones that survive a downturn and emerge stronger. In an environment with genuine bubble risk, capital efficiency is not just prudent. It is protective.

Value defensibility over participation in the theme. Being adjacent to the AI boom is not the same as being protected by it. A company whose only claim is proximity to a hot category carries the full risk of that category's potential correction. A company with a defensible position, proprietary data, deep customer relationships, real switching costs, holds its value even if the broader AI trade repriced tomorrow.

Diversify away from the concentrated core. The circular financing risk is concentrated in a specific place: the frontier labs, their infrastructure suppliers, and the web of deals connecting them. The further an investment sits from that concentrated core, the less exposed it is to a potential unwind of those specific arrangements. Companies solving real problems for real customers, in real markets outside the frothy center, offer exposure to the genuine utility of AI without the concentrated systemic risk.

Where the Protected Opportunities Live

Follow the logic to its conclusion and a clear picture emerges of where a bubble-aware investor should be looking.

The protected opportunities are the AI companies that look least like the circular financing core. They have real external customers rather than revenue that loops back to their investors. They are capital-efficient rather than dependent on the next enormous round. They solve specific, valuable problems rather than participating in the theme in the abstract. And they are frequently located outside the concentrated geographic and financial center where the circular dynamics are most intense.

This describes a specific and often overlooked segment of the market: companies applying AI to real problems, for real paying customers, in markets that the frontier lab frenzy has largely ignored. A company using AI to solve a genuine workflow problem for businesses in an emerging market is about as far from the circular financing risk as an AI investment can be. Its revenue comes from real customers. It is not dependent on the frontier lab funding cycle. And it is priced reasonably because it sits outside the concentrated core where valuations are most inflated.

In a market where the central risk is circular revenue and concentrated systemic exposure, real external revenue in an underpriced, overlooked market is the definition of a protected position.

How SeedScope Helps You Build the Protected Portfolio

The bubble-aware strategy comes down to sourcing companies with real external revenue, genuine defensibility, and reasonable valuations, frequently in markets outside the concentrated core. The challenge is access to those companies, which do not surface in the deal flow dominated by the frontier lab ecosystem.

This is exactly what SeedScope provides. With active founders across 30+ countries, filterable by stage, sector, and geography, SeedScope gives investors access to companies applying AI to real problems for real customers, in markets well outside the circular financing dynamics of the AI core. The AI-powered valuation benchmarking grounds every opportunity in real comparable data, so you can identify the companies whose fundamentals hold up regardless of what happens to the macro AI trade.

Whether the AI boom proves to be a bubble or a durable transformation, the companies with real external revenue and defensible positions are the ones that survive and compound. SeedScope is built to help you find them.

The Bottom Line

The question of whether the AI boom is a bubble cannot be answered with confidence, and any investor who claims certainty in either direction is telling you more about their bias than about the market.

What can be answered is how to position for either outcome. Both the optimists and the pessimists agree that the sustainability of the entire AI economy comes down to whether the revenue is real and external, rather than circular and self-referential. That single point of agreement is the most useful guidance in the entire debate.

Build a portfolio of companies with real external revenue, genuine defensibility, capital efficiency, and reasonable valuations, and you are protected if the bubble deflates and exposed to the upside if it does not. That is not a bet on the outcome of the debate. It is a strategy that works regardless of how the debate resolves. And in a moment of genuine uncertainty, that is exactly the kind of strategy worth having.

Build a portfolio grounded in real fundamentals, not circular hype. Explore active founders on SeedScope across 30+ countries. Start here →