AI founders seem to have a never-ending list of reasons — and hyperventilated pitch decks — explaining why their financial losses don’t matter. Some are hopeful, some are delusional, and some are just echoes of arguments that would-be billionaires floated in the dot-com era—updated with better graphic design.

A new article at LLRX.com, entitled The Imminent AI Bubble Crash (and Why It Won’t Matter in the Long Run), explains some of the most common excuses, but space limitations there prohibited a complete list. Here are some of the other most common attempts to justify the bubble, along with brief rebuttals and a bit of good-natured skepticism:

1. “We’re prioritizing growth over profits.”

Rebuttal: Growth is great, but not when it’s the financial equivalent of gaining weight by eating subsidized ice cream. Users acquired through free money tend to vanish once the free money does.

2. “This is a land-grab moment.”

Rebuttal: That assumes there’s valuable land—and that someone will eventually pay rent. Plenty of dot-com veterans can point to the “land” they grabbed; it now serves as a digital ghost town with excellent parking.

3. “Monetization will come once we turn on premium features.”

Rebuttal: This is the startup version of “I’ll start my diet on Monday.” It sounds good, but the conversion rate from free users to paying customers often ends up in the single digits, as in one digit, and not a high one.

4. “Compute costs are high today, but they’re falling fast.”

Rebuttal: True, but usage is rising even faster. Sam Altman reports that ChatGPT loses more money on $ 200-a-month premium subscriptions than on $20 subscriptions.

5. “Our unit economics are improving.”

Rebuttal: Losing less money per user is not the win that founders think it is. It’s like a restaurant bragging that it now loses only $9 on a $10 burger.

6. “Every user interaction creates proprietary data.”

Rebuttal: Proprietary data is valuable—if your competitors don’t have nearly identical data and access to the same base models. Many so-called “data moats” turn out to be kiddie pools.

7. “We’re building defensibility through R&D.”

Rebuttal: R&D is important, but it’s not a moat if everyone else is also spending aggressively on R&D—especially when the competitors are named Google, Meta, or “OpenAI’s Entire Microsoft-Funded Budget.”

8. “We’re becoming the indispensable platform layer.”

Rebuttal: True platform companies get adopted because others depend on them—not because the founders wish really hard. With 50 nearly interchangeable AI layers, the market looks less like a platform race and more like speed dating.

9. “Our burn rate is intentional, and we have plenty of runway.”

Rebuttal: Runway only tells you the plane hasn’t crashed yet. If the business model doesn’t change, all that “runway” guarantees is a longer, more scenic descent.

10. “Regulation will wipe out weaker competitors.”

Rebuttal: Possibly. But regulation has a long history of hitting everyone equally—and sometimes hitting the big players harder. Banking on regulators to save your business is a strategy that has rarely survived contact with regulators.

11. “All important AI companies went through this phase.”

Rebuttal: Survivorship bias is strong. For every success story, there’s a small graveyard of companies that burned cash with equal enthusiasm but did not leave memoirs.

12. “The total addressable market is enormous.”

Rebuttal: A huge TAM is comforting, but it doesn’t guarantee anyone’s survival. The ocean is enormous, too—and full of shipwrecks.