How AMD Moves on Fed Days: A Decade of Data

Semis Are Rate-Sensitive, and AMD Is No Exception

Semiconductors as a group tend to be among the most rate-sensitive sectors in the S&P. High multiples, long duration cash flows, and cyclical exposure to capital spending cycles all amplify the impact of Fed decisions. AMD, as one of the most watched semis, reflects this sensitivity clearly.

Across ~80 FOMC meetings from 2016 to 2026, AMD's average Fed-day return has been roughly +0.5% with a win rate near 54%. Standard deviation has been about 2.5% — higher than Apple (1.5%) and slightly higher than NVDA (2.3%).

Dovish vs Hawkish: Strong Asymmetry

AMD shows a clear rate-sensitivity asymmetry in Fed-day returns. On dovish days, AMD has averaged roughly +1.6% with a win rate near 67%. On hawkish days, AMD has averaged roughly -1.2% with a win rate near 38%. The magnitude is comparable to NVDA and slightly less extreme than Tesla.

This makes AMD a decent (if noisy) rate-proxy trade. When you expect a dovish Fed, buying AMD at the previous close has historically been a positive-expectancy bet. When you expect a hawkish Fed, the opposite.

• Dovish Fed days: ~67% win rate, ~+1.6% average return
• Hawkish Fed days: ~38% win rate, ~-1.2% average return
• Neutral Fed days: ~52% win rate, ~+0.3% average return
• Standard deviation of Fed-day returns: ~2.5%

AMD Moves Correlate Strongly With SMH on Fed Days

On Fed days specifically, AMD's daily return has a correlation of roughly 0.85 with SMH (the iShares Semiconductor ETF) — slightly higher than its long-run correlation. This makes sense: Fed decisions tend to produce sector-wide moves, and individual stock alpha dwindles on macro days.

For traders, this suggests that betting on AMD-specific news during Fed days is usually a losing proposition. The stock is dragged by the sector, and the sector is dragged by the macro signal.

Press Conference Matters More Than the Statement

Like other high-beta names, AMD's Fed-day price action tilts heavily toward the 2:30pm press conference. Roughly 60% of AMD's daily range on Fed days occurs after 2:30pm ET, suggesting that the Chair's commentary is more impactful than the initial statement release.

The initial 2pm reaction is often a head-fake. Traders who wait 30 minutes for the press conference to begin, then position based on the post-conference direction, have historically captured a cleaner signal.

Using the Data

A practical workflow for AMD on Fed days: wait for the press conference to finish (typically 3pm ET or later), then pull the current AMD chart and compare it to historical analogs. If analogs skew positive, the follow-through is more likely to be upward.

from chartlibrary import ChartLibrary cl = ChartLibrary(api_key="cl_...") result = cl.intelligence("AMD") print(result.forward_returns)

Related reading: our posts on NVDA Fed day reaction and market regime tracking cover adjacent data on rate-sensitive trading.