Got $1,000? 2 Magnificent Artificial Intelligence (AI) Stocks Down Over 15% to Buy Hand Over Fist

Micron (MU) is highlighted as benefiting from a memory chip shortage driven by AI data-center buildouts, with management expecting supply-demand tightness to persist beyond 2027. For FY2027, analysts project 81% revenue growth and EPS near $150 (vs. $73.32 expected for FY2026), supporting a “buy on the dip” thesis. Broadcom (AVGO) is cited as a major AI catalyst, targeting AI semiconductor revenue of $100B+ by 2027 (vs. $10.8B in its most recent quarter) with Wall Street expecting revenue growth of 67% this year and 62% next year.

Analysis

MU is the cleaner near-term earnings revision story, but the market is likely underestimating how fast this can break if capex from the largest AI buyers slows or if memory producers add capacity faster than expected. The key variable is not just shortage duration; it is whether HBM and enterprise SSD mix keep the industry in a high-margin regime long enough for buy-side models to ratchet higher. If that mix slips, MU can move from “scarcity winner” to “cyclical multiple compression” in a single guidance cycle.

AVGO is better framed as an infrastructure tax collector on AI spend: if hyperscalers keep shifting dollars from general-purpose accelerators to workload-specific silicon, AVGO can gain share even if total AI capex growth moderates. That creates a second-order headwind for NVDA and, to a lesser extent, AMD, because some compute budgets that would have gone to merchant GPUs instead get reallocated into custom chips. The catch is that this model is customer-concentrated and lumpy; a delayed tape-out or design win pushout won’t show up immediately, but it can hit the 1-3 month estimate narrative hard.

Contrarian view: the consensus is treating “AI semis” as one trade, but MU and AVGO are different exposures. MU is more vulnerable to a supply response over 6-18 months, while AVGO is more vulnerable to a near-term disappointment in hyperscaler conversion rates. I’d treat the current weakness as an opportunity only selectively: MU has the tighter catalyst path, while AVGO needs continued proof that custom silicon can scale without cannibalizing margins or slipping timelines.

Key Decisions for Investors

• Buy MU on 5-10% pullbacks into the next earnings window; thesis fails if management signals DRAM/HBM pricing inflection or capex normalization sooner than expected.
• Own AVGO on dips for a 6-18 month structural AI share-gain trade, but prefer call spreads over common stock to cap multiple-compression risk if customer concentration pushes out revenue recognition.
• Pair trade: long AVGO / short NVDA for the next 3-6 months if you want exposure to AI capex reallocation toward custom silicon rather than merchant GPUs; cover if NVDA re-accelerates hyperscaler training demand or AVGO custom win cadence slows.
• Use MU as the higher-beta expression and AVGO as the higher-quality expression; if you want one basket, overweight MU inside SMH/SOXX rather than chasing the whole semiconductor complex.
• Set a watch item on hyperscaler capex updates and memory pricing data; a two-quarter roll-over in pricing would falsify the bullish MU thesis even if AI demand remains strong.