A $2 Trillion Test
Anthropic, the company behind Claude, has publicly filed its S-1 and is targeting an IPO as early as the week of November 16th. Anthropic is seeking a $2 trillion valuation, more than double its most recent funding round in April of this year.
There is little question that Anthropic’s growth has been extraordinary. Revenue increased from approximately $386 million in 2024 to $4.6 billion in 2025 to $16.3 billion in just the first half of 2026. Profitability tells a very different story. The company reported an $80.7 billion net loss during that same six-month period. Although much of that loss was related to noncash accounting adjustments regarding funding, the company still reported an operating loss of $693mm in the first half of 2026.
On one hand Anthropic Is largely viewed as the leader in the Artificial Intelligence race. But a $2 trillion dollar valuation paired with the financial metrics reported in the S1 leaves plenty of room for skepticism about how much an investor should pay for extraordinary growth in an intensely iterative environment where a company can go from leader to obsolete in no time flat. The company must ultimately demonstrate that this growth can produce durable profits after considering the extraordinary cost of computing, infrastructure, and continued innovation.
From “What can I do?” to “What can I earn?”
The IPO is coming precisely at a time when scrutiny around the AI trade is amping up. As investors begin to move beyond being impressed by what AI can do and toward asking what AI can earn. Recent reactions to earnings show that more capex is no longer automatically good news. Where a few years ago a conference call announcing additional spending of $20 billion on AI was a bullish signal, now investors are increasingly asking what kind of revenue does it generate, at what margin, and how is it being funded? Amazon, Alphabet, Meta, Microsoft and SpaceX are projected to spend more than $800 billion in aggregate capex during 2026.
Credit Markets are Asking Questions Too
Perhaps the clearest evidence of AI scrutiny can be found outside the stock market. Sophisticated credit investors, who tend to care less about technological promise and more about getting their money back, have begun to push back. While the big AI spenders show no sign of slowing down, increased discernment is showing up in corporate bond yields and credit default swaps, a way for investors to hedge risk in fixed income investments. Oracle’s credit deterioration is particularly interesting to investors since it’s far from being a speculative startup. The cost to insure its bonds for a 5-year time horizon has moved materially higher. While Oracle’s debt remains investment grade, at BBB-, it is one downgrade away from junk status.
AI Spending and S&P 500 Earnings
AI spending has been enormously supportive of corporate earnings and the ongoing, years-long rally in equity market. Year to date the S&P 500 is up more than 13%. At the same time, earnings expectations have advanced faster than stock prices, allowing the market’s valuation multiple to contract even as the index has remained near record levels. In other words, recent market gains have increasingly been supported by growing profits rather than investors simply agreeing to pay ever-higher multiples for those profits.
At the same time, as the AI buildout gains more scrutiny, estimates attributed roughly 40% of S&P 500 year-over-year EPS growth to AI-related spending, with the most direct benefit flowing to semiconductor manufacturers and the broader infrastructure supply chain supplying chips, networking equipment, data centers and related technology. That creates an interesting tension for investors. AI investment has helped generate real earnings growth, and those earnings have helped the market grow into what had previously looked like a demanding valuation. But the next phase may prove more difficult as tensions around spending and return on that investment rise.
Greg Barshied, CFA
Portfolio Manager