The AI rally has an unusual problem for anyone calling it a bubble: earnings are growing much faster than stock prices. Technology has gained more than 20% this year, while earnings growth has reached roughly 56% and valuations have compressed, according to Andreessen Horowitz's September 2026 State of Markets report.

A market typically rises for two reasons: investors can pay increasingly higher prices for the same dollar of earnings, or companies can generate enough additional profit to justify higher stock prices.

According to a16z, the latter is happening. Data shows the tech sector up 22.1% year-to-date, while earnings per share have grown 55.7%. At the same time, the sector's price-to-earnings multiple has contracted 21.6%.

The firm's broader conclusion is blunt: the market keeps reaching new highs, but "it's profits, not prices [are] doing the work."

Valuation Data Makes the Contrast More Striking

Tech stocks rising 22% while earnings grow about 2.5 times as fast is a very different setup from prices racing ahead of fundamentals. And a16z says the sector's valuation multiple is now roughly 20% below its five-year average.

In other words, investors are paying less for each dollar of technology earnings than they have historically, even as those earnings are growing rapidly.

The report also argues that the market is no longer simply a Mag 7 story. While the biggest technology companies have lagged the broader technology universe this year, hyperscaler spending on AI infrastructure has pushed semiconductor earnings to unprecedented levels, with market returns following the money.

In other words, the AI trade is increasingly reaching companies supplying the chips, memory, power and other infrastructure required to build AI systems.

The Bubble Test Gets Harder

None of this proves that AI stocks cannot fall. The a16z data describes the market through the report's September 2026 snapshot, and earnings expectations can change quickly if AI spending slows or the returns on that investment disappoint.

But it does establish a more complicated picture than the usual bubble narrative. Technology stocks are up more than 20%, yet earnings have grown roughly 56% and valuation multiples have contracted.

For investors, the next test is whether earnings can continue growing fast enough to justify the capital being poured into AI infrastructure. If profits keep catching up with — or outpacing — stock prices, the debate around an AI bubble becomes increasingly a question of future earnings rather than simply how far stocks have climbed.