The recent chip selloff has raised an important question: Is the AI investment cycle actually slowing, or is the market simply reassessing expectations?
1️⃣ Chip weakness is noticeable — but not yet a trend
Monday saw a sharp pullback across semiconductors, with Nvidia down 3.4% and Micron around 5%. By Tuesday, Nvidia recovered about 0.6%, while AMD gained 2.19%.
That rebound matters because it suggests investors haven’t completely walked away from the AI trade.
2️⃣ The bigger signal is AI CAPEX 💰
This is where I think investors should look beyond the headlines.
A slowdown in frontier-model development doesn’t necessarily mean a slowdown in spending on GPUs, memory, networking, data centers and AI inference.
One recent Bank of America fund-manager survey found 79% of respondents did not expect AI hyperscalers to reduce 2026 capital expenditure.
📊 My key indicator:
AI demand → Hyperscaler CAPEX → Infrastructure orders → Semiconductor revenue
If that chain remains intact, the recent weakness could simply be a repricing of expectations.
3️⃣ What could prove the bears right? ⚠️
The biggest risk isn’t another one-day decline in Nvidia or AMD.
I’d be more concerned if major cloud companies start:
• Cutting AI infrastructure budgets
• Delaying data-center projects
• Reducing GPU orders
• Guiding to slower AI-related spending
That would provide much stronger evidence that the AI cycle is actually losing momentum.
📌 My takeaway
I’m not treating the current volatility as either a clear buying opportunity or the end of the AI trade.
I’m watching the money.
Headlines can change every day, but sustained changes in CAPEX, orders and earnings would tell us much more about where the AI cycle is heading.
👀 What do you think matters more right now: AI CAPEX, chip earnings, or the pace of AI development?
Comments