Michael Anders of ICONIQ Capital said the firm invested in Anthropic after seeing improving unit economics, stronger margins, and durable customer adoption.

The investment highlights the shift among venture capital firms from speculative interest in generative AI to a requirement for sustainable financial metrics. As AI companies seek massive capital injections to scale compute, investors are prioritizing evidence of long-term profitability over raw growth.

Speaking at the Forbes Iconoclast Summit, Anders said these specific financial signals were necessary to move the firm from general excitement to the conviction required for a multi-billion-dollar investment. "Improving unit economics, stronger margins, and evidence of durable customer adoption were the signals that transformed excitement into conviction," Anders said.

ICONIQ Capital has committed billions of dollars to AI investments, including its stake in Anthropic [5]. This move comes amid reports of significant capital movement within the AI sector, though figures regarding Anthropic's current valuation vary widely across reports.

Some reports indicate that Anthropic is seeking to raise $5 billion in a new funding round at a proposed valuation of $170 billion [1, 2]. However, other reports state the company raised $65 billion in a Series H round, which valued the firm at $965 billion [3, 4].

These discrepancies in reported valuations reflect the opaque nature of private AI funding rounds. Despite the variance in numbers, the focus for firms like ICONIQ remains on the underlying health of the business model. Anders said the transition to a high-conviction investment depended on the company's ability to demonstrate that its customers were adopting the technology in a durable manner.

“"Improving unit economics, stronger margins, and evidence of durable customer adoption were the signals that transformed excitement into conviction."”

The transition from 'excitement' to 'conviction' described by ICONIQ Capital signals a maturing AI market. While early investments were driven by the novelty of Large Language Models, the current phase requires proof of a viable business model. The stark contradictions in reported valuations—ranging from $170 billion to $965 billion—suggest a highly volatile pricing environment where the perceived value of AI infrastructure is still being contested by the market.