NextFin News - In the spring of 2021, Anjney Midha, then a partner at a boutique venture firm, walked into the glass-walled offices of 22 of the world’s most prestigious venture capital institutions with a pitch for a startup called Anthropic. He walked out with 21 rejections. Today, as of March 30, 2026, those same institutions are reportedly scrambling for a piece of Anthropic’s latest $20 billion funding round, which has catapulted the company’s valuation to a staggering $350 billion.
The contrast is more than just a missed opportunity; it represents the most expensive collective miscalculation in the history of venture capital. In 2021, the 21 firms that turned Midha away cited "concentration risk" and "unclear commercial pathways" for a team of OpenAI defectors obsessed with "AI safety." At the time, Anthropic’s Series A was valued at roughly $1 billion. By avoiding what they deemed a "high-risk outlier," these firms missed out on a 350-fold return in just five years. The capital that did enter—led by Skype co-founder Jaan Tallinn’s $124 million Series A—was driven not by traditional ROI metrics, but by a "non-mainstream" belief that AI safety was the only moat that mattered.
Roelof Botha, the global decision-maker at Sequoia Capital, was among those who initially resisted leading early rounds. According to internal reports, Botha’s hesitation was rooted in a traditional asset allocation framework that viewed massive compute-spend as a "bottomless pit." This conservative dogma eventually led to a significant leadership reshuffle at Sequoia in late 2025, with Alfred Lin and Pat Grady taking more aggressive stances. Sequoia’s eventual participation in the January 2026 round at the $350 billion valuation serves as a "cognitive premium"—a multi-billion dollar penalty for five years of observation from the sidelines.
However, the current valuation surge is not without its skeptics. Siddharth, a senior analyst at a leading tech consultancy, suggests that the $350 billion figure may be more reflective of "survival instinct" than fundamental value. Siddharth, who has historically maintained a cautious "neutral" stance on generative AI valuations, argues that the current frenzy is driven by the fact that AI infrastructure investment now accounts for nearly 30% of U.S. capital expenditure. In his view, the entry of late-stage capital is less about foresight and more about the terrifying reality that, without AI integration, traditional business models face obsolescence. This perspective is not yet the consensus on Wall Street, where many still view the $350 billion valuation as a justified bet on the "sovereign-level" importance of the Claude R1 model series.
The shift in venture logic is also visible in the macro data. In the first half of 2025, U.S. GDP growth excluding AI-related software and equipment fell to a meager 0.7%, while AI infrastructure investment soared by 28%. This has forced a "dimensional reduction" strike by non-traditional capital. Figures like Eric Schmidt and Dustin Moskovitz, who backed Anthropic when VCs wouldn't, operated on "capital replacement" logic—using personal wealth to fund technology they believed was essential for human survival, effectively squeezing out the spreadsheet-driven models of traditional LPs.
The irony of 2026 is that the very "safety" focus that made Anthropic uninvestable to 21 firms in 2021 has become its primary commercial engine. Enterprise clients, wary of the "hallucination" risks and ethical lapses of less regulated models, have flocked to Claude’s "Constitutional AI" framework. This has allowed Anthropic to project an annualized revenue run rate that justifies its massive valuation to a new generation of investors. The 21 rejection letters of 2021 now stand as a historical monument to the limits of traditional risk management in an era of exponential technological shifts.
Explore more exclusive insights at nextfin.ai.
