AI's Reality Check: Altman Admits Error as Spending Data Tells Its Own Story
How this was made Verified AI
Every Intellegix briefing is generated from that day's broadcast and run through automated checks before it publishes — with a human paged on any flag. Here is the trail for this edition.
Sam Altman publicly acknowledged that he was wrong about the pace of AI adoption following GPT-4's 2023 launch, saying he expected far more economic disruption by now. His admission landed at the same moment that spending data from Ramp — which processes corporate card transactions, making it behavioral rather than survey-based — showed Anthropic's flagship model, labeled in the data as Fable 5, capturing only eleven percent of enterprise AI spending more than two months after launch.
The eleven percent figure requires context. Enterprise AI spending as a whole has grown dramatically, making eleven percent of a large pool still meaningful revenue. The issue is what the number implies about Anthropic's market position ahead of what appears to be an anticipated IPO: enterprises are choosing cheaper alternatives at roughly nine-to-one over Fable 5. The capability gap between frontier models and lower-cost alternatives has narrowed faster than the pricing gap has, undermining the core value proposition that frontier capability commands frontier pricing.
The competitive dynamics in the AI model market illustrate, in real time, how difficult monopoly power is to sustain in software markets with low switching costs and multiple well-funded competitors. Anthropic cannot simply raise Fable 5 prices because enterprises would shift to OpenAI, Google's Gemini Pro, or open-source alternatives like Llama. That competitive discipline is precisely what antitrust law is designed to preserve. Whatever regulatory concern exists about AI concentration has focused more on infrastructure — compute access, data, distribution — than on the model layer itself, which the Ramp data suggests is genuinely and effectively competitive.
On the political front, an Axios survey of likely 2028 Democratic presidential candidates — including figures such as Gavin Newsom, Gretchen Whitmer, and Josh Shapiro — found essentially none willing to either endorse or criticize Bernie Sanders' call to pause AI development. The calculation is evident: AI is a major employer and a significant source of campaign contributions in coalition-building states. The result is a policy debate of genuine consequence met, across an entire prospective presidential field, with studied silence.