OpenAI’s ChatGPT now has 800 million active weekly users, a “boon for the optimists” who are funding a $3 trillion AI datacenter build-out. The question is whether this impressive user base can justify the $1.5tn “speculative” debt bubble that is building alongside it.
The “boom” narrative is clear: 800 million users show “exponential demand” for AI. This is driving tech’s “hyperscalers” (Google, Microsoft) to spend $750bn, fueling record valuations (Nvidia at $5tn) and “generational employment opportunities” in towns like Newport, Wales.
The “bubble” narrative is focused on the $1.5tn “funding gap” being filled by “private credit.” This “shadow banking” debt is financing “unproven” projects “without their own customers,” based on “lofty revenue expectations” that generative AI will be a $1tn market by 2028.
But an MIT study found 95% of organizations are getting “zero return” on generative AI pilots. This suggests that while 800 million people are using AI, businesses are not yet profiting from it. This is a critical disconnect.
The “speculative” $1.5tn in debt is being wagered on the hope that these 800 million users can be monetized before the “quickly depreciating assets” they run on lose their value. It’s a race between user growth and asset depreciation, with “structural risk to the global economy” at stake.