Cathie Wood urges Federal Reserve to consider AI-driven deflation

Cathie Wood has a message for the Federal Reserve: the inflation gauges you trust may be missing a big part of the picture.

The ARK Invest CEO argues that the Consumer Price Index and Producer Price Index undercount the deflation coming from AI software. She wants the Fed to start paying attention.

What Wood is actually saying

Wood’s latest commentary was published around November 7 to 9, 2026. Her core claim is that CPI and PPI understate how fast AI is pushing prices down.

CPI tracks what consumers pay for a basket of goods and services. PPI tracks what producers receive for their output. Together, they are among the main signals the Fed watches when deciding whether the economy is running hot.

Wood’s view is that neither index fully captures what is happening in AI software and inference. Inference is the computing work an AI model does each time it answers a question or generates an output.

In remarks from late September and early October 2026, Wood pointed to a 99.99% annual decline in AI inference costs at constant performance levels.

Good deflation versus bad deflation

Wood draws a sharp line between AI cost decline and demand-driven deflation. She argues the AI cost decline comes from supply-side productivity gains, not weak demand.

Her evidence for strong demand is OpenAI. She cited the company’s annual revenue run rate climbing from $20 billion to a projected $70 billion.

Wood has been making this case publicly for a while. Around September 29 to October 3, 2026, she directly pushed back on inflation concerns raised by Bill Ackman. She labeled the AI cost collapse “good deflation.”

“Good deflation.”

The GDP forecast and the real-time data

Wood’s argument extends beyond software pricing into a broader macro forecast. ARK projects that real GDP growth could exceed 7% while inflation falls well below the Fed’s 2% target.

She also leans on alternative data. Real-time inflation trackers such as Truflation show core inflation readings in a range of 1% to 2.1%. Those figures sit below the official numbers, which supports Wood’s claim that government statistics lag what is happening on the ground.

What this means for markets and the Fed

For investors, the stakes come down to interest rates. If the Fed came around to Wood’s view, it could justify a more accommodative stance than current official inflation data would suggest.

That alignment is worth keeping in mind. Wood runs funds that would benefit from easier policy and a stronger AI narrative. Her thesis deserves scrutiny on its merits rather than acceptance on enthusiasm.

There is also a sector question. AI deflation is not evenly spread. Software and inference costs may be plunging, but housing, insurance, and many services have their own dynamics that cheaper chatbots do not obviously fix.

The things to watch are straightforward. First, whether the gap between real-time trackers like Truflation and official CPI readings keeps widening. Second, whether Fed officials begin referencing AI productivity in their public commentary. Third, whether AI demand, measured by revenue figures like OpenAI’s projected run rate, keeps pace with falling costs.