Anthropic’s impending IPO is Groupon’s 2011 story rebuilt at trillion-dollar scale: exponential revenue is being mistaken for permanent pricing power just as competition begins destroying the scarcity supporting it. Groupon’s numbers once looked invincible. Revenue exploded from $100,000 in 2008 to $30.5 million in 2009 and $713.4 million in 2010 - annual increases of 32,316% and 2,241%. Subscribers multiplied from 152,203 in June 2009 to 115.7 million two years later. Yet Groupon lost $101.2 million in the second quarter of 2011 even as quarterly revenue reached $392.6 million. Investors ignored the weakness because the growth curve was spectacular. Groupon sold 35 million shares at $20 in November 2011, raised $700 million and entered the market at a $12.7 billion valuation. Within three weeks, the stock was below its IPO price. Nine months later, it closed at $5.60—down 72% from the offering price and nearly 80% from its post-IPO high. The business did not suddenly become useless. Its product became reproducible. Competitors could copy daily deals, merchants could list elsewhere and customers could switch instantly. Groupon’s growth survived temporarily - 2012 revenue still rose 45% to $2.3 billion - but its valuation collapsed because markets recognized that growth without scarcity produces weak pricing power. Anthropic is approaching precisely this trap. Its valuation reportedly depends on revenue expanding from a $47 billion run rate to approximately $190–$200 billion in 2028, supporting discussion of an IPO valuation potentially reaching $2 trillion. That price does not merely assume that Claude remains useful. It assumes years of extraordinary growth, premium token pricing and durable margins. Open models are now attacking all three assumptions. Recent reporting places leading open-weight models only about four months behind frontier systems, with some delivering comparable benchmark performance at up to 80% lower cost. Seven of OpenRouter’s ten highest-volume models are Chinese open models. The direction is unmistakable: model capability is spreading while inference prices are compressing. The 100× collapse follows brutal arithmetic. When equivalent intelligence becomes available for one-hundredth the price, Anthropic must cut prices or surrender workloads. At unchanged volume, revenue falls 99%. Even a tenfold increase in token consumption leaves revenue 90% lower. Usage must rise one hundredfold merely to restore the original revenue; and that still does not recover the valuation multiple lost when intelligence is repriced from scarce proprietary technology to abundant commodity infrastructure. That creates a double collapse: revenue expectations fall while the market simultaneously slashes the multiple assigned to each remaining dollar of revenue. A company once valued like a monopoly is repriced like a utility. Anthropic may produce excellent models and still become a disastrous stock. Groupon proved that enormous adoption cannot protect a valuation once the market discovers there is no defensible tollbooth. Anthropic’s IPO will transfer that commoditization risk to public shareholders just before open models expose it. Groupon commoditized discounts. Open source is commoditizing intelligence. The ending will be the same; only the numbers will be larger.