Loading prices…
STKR NewsSTKR News0 of 3 free this month
AI

What The Telecom Bust Taught Me About AI’s Debt Boom

Venture debt for AI infrastructure is exploding, but we have seen this movie before. If the cash flows do not arrive before the bills do, builders are in for a brutal wake-up call.

Originally on Crunchbase News →
AB

Adrian Boysel

Contributor

Nov 20, 2026

5 min read

Photo illustration / STKR News

I have been spending a lot of time looking at the balance sheets of AI infrastructure companies lately, and I am starting to get a familiar, uneasy feeling in my gut. It is the same feeling I had watching the telecom sector in the late nineties. We are seeing a massive disconnect between the money being borrowed and the actual revenue coming through the door.

Debt is currently the drug of choice for AI founders. It makes sense on the surface. Why dilute your equity when you can borrow a few hundred million dollars to buy the H100s you need to scale? But history shows that when infrastructure spending outpaces demand, the fallout is not just a correction; it is a wipeout. If you are building in this space, you need to understand the structural risks being baked into the industry right now.

The Ghost of Telecom Past

To understand where AI might be heading, you have to look at what happened during the telecom boom. Back then, billions were poured into laying thousands of miles of fiber-optic cable. The thesis was simple: the internet was coming, and the world would need infinite bandwidth. The thesis was actually correct. The internet did become the backbone of society. The problem was the timing and the cost of the debt used to build that backbone.

Companies borrowed money based on aggressive growth forecasts that never materialized in the timeframe they needed. When those companies could not meet their fixed repayment schedules, the entire sector collapsed. We are seeing the exact same pattern today. Founders are betting that AI demand will scale linearly forever, and they are leveraging their companies to the hilt to acquire hardware that might be obsolete in three years.

The Trap of Fixed Obligations

The danger of debt is that it is indifferent to your product-market fit. If you take venture capital, your investors might be unhappy if you pivot or slow down, but they generally cannot force you into bankruptcy. Debt is different. Debt has a schedule. It requires cash flow, not just "potential."

Right now, a lot of the debt being issued in the AI space is tied to the physical hardware—the chips and the servers. Lenders feel safe because they have collateral. But for a builder, this creates a situation where your monthly burn is dictated by a bank, not your actual operational needs. If your enterprise customers take six months longer to integrate your LLM than you expected, you do not just miss a quarterly goal; you potentially lose control of your company.

Why Hardware Is a Fickle Collateral

In the telecom days, fiber in the ground was supposed to be a safe asset. But when everyone defaulted at once, the market was flooded with cheap fiber, and the value of that collateral plummeted. AI builders need to ask themselves what happens to the resale value of their GPUs if the current hype cycle cools off. If the "collateral" drops in value by 50%, your lender is going to start asking very uncomfortable questions.

Furthermore, technology cycles are moving faster than ever. We are seeing new architectures and more efficient chips being announced almost every quarter. If you take out a five-year loan to buy today's top-tier hardware, but that hardware is effectively a paperweight in three years compared to the new competition, you are stuck paying off a legacy asset while your competitors lean into the next generation. That is a recipe for a slow death.

What Builders Should Watch For

If you are a founder, you need to stop looking at debt as "free money for chips." It is a high-stakes bet on your own execution speed. Here are the red flags I am watching for in the market right now:

  • Forecast Inflation: If your financial model assumes 20% month-over-month growth for the next three years just to cover debt service, you are essentially gambling, not building.
  • Lack of Revenue Diversity: If 80% of your revenue comes from other AI startups that are also fueled by debt, you are part of a circular economy that will collapse the moment the first domino falls.
  • Short-Term Thinking: Builders who are maximizing for compute capacity over sustainable unit economics will be the first ones captured by their creditors when the market tightens.

The Case for Sanity

I am not saying debt is inherently bad. For a mature company with predictable, recurring revenue, debt can be a powerful tool for scaling. But we are seeing seed and Series A companies taking on debt profiles that would make a mid-market manufacturing firm nervous. This isn't just about risk; it's about the loss of optionality.

When you are heavily leveraged, you lose the ability to experiment. You cannot afford to spend three months tinkering with a new model if that delay means missing a payment. You become a slave to your repayment schedule, which often forces you to make short-sighted decisions like taking on bad customers just to hit a monthly number.

The Reality Check for AI Founders

The AI revolution is real, just like the internet revolution was real. But being right about the technology does not mean you will win as a business. The companies that survived the telecom bust were the ones that managed their balance sheets with extreme discipline. They did not over-expand based on a hypothetical future.

If you are building an AI startup today, my advice is to treat debt like a surgical tool, not a blunt instrument. If you cannot explain exactly how you will pay back every dollar without raising another round of equity, you probably shouldn't be borrowing that dollar. The goal is to build a company that lasts, not just a company that has the most GPUs for six months before the bank takes the keys.

The biggest risk in AI right now isn't that the technology fails to deliver; it's that we build a financial structure that collapses under the weight of our own expectations before the technology has a chance to mature.

We are going to see a lot of talented teams get wiped out over the next 24 months, not because their code was bad, but because their finance strategy was built on a house of cards. Don't let that be you. Keep the ego in check, look at your cash flow with a skeptical eye, and remember that in a race, the person who runs out of breath first rarely wins, no matter how fast they started.


Read the original at Crunchbase News →

The Brief

Stay Updated on Cutting-Edge Tech

A six-minute morning dispatch on the markets and the technology shaping them.

Free. No spam. Unsubscribe anytime.

Write for STKR

Become a Contributor

Earn $STKR for published stories on markets, protocols, and culture.

  • Earn $STKR for every published piece
  • Editorial support from the STKR desk
  • Byline visibility across the network
  • First look at the upcoming creator program
Apply to Write

Keep reading

All stories

Comments

24 reader responses