These two chip stocks could join the elite $2 trillion tech giant club, says top analyst

Melius analyst Ben Reitzes argues that Micron and AMD could each reach $2 trillion in market capitalization, driven by AI-led demand for memory and a renewed role for CPUs in agentic-AI systems. He says AMD would need to demonstrate more than $50 in EPS power versus its current 2030 target of more than $20, while Micron could benefit from substantial future buybacks. The outlook is contested: Michael Burry expects Chinese capacity additions to end the memory shortage and trigger a downcycle within two years. Micron reports earnings after the close, with FactSet consensus at $31.72 EPS and $51.33B revenue; options imply a 7% post-results move after shares gained 276% year to date.

Analysis

The key valuation question is not whether AI raises memory content, but whether MU can convert a capacity-constrained HBM/DRAM cycle into through-cycle free cash flow. A $2T equity value would require the market to underwrite both structurally higher memory margins and unusually disciplined capex; either breaks if Samsung and SK Hynix prioritize share recovery over pricing. The near-term setup is asymmetric after a 276% YTD move: an in-line print is unlikely to suffice, while any indication of HBM qualification delays, weaker contract-pricing cadence, or a capex step-up can trigger multiple compression within days.

AMD's CPU exposure is a legitimate second-order AI beneficiary, but it should be treated as an attach-rate thesis rather than a replacement thesis for NVDA. Higher host-CPU, memory, and networking requirements expand AMD's addressable market, yet hyperscalers can mitigate that demand with ARM-based internal silicon and tighter GPU utilization. Reaching $50+ EPS by 2030 implies sustained datacenter share gains, materially higher gross margins, and limited price competition from Intel and custom ASICs; that is a 6-18 month execution debate, not an earnings-quarter catalyst. The reported MU consensus figures appear internally inconsistent with historical company scale, so trading decisions should use verified sell-side estimates and the options-implied move rather than the cited EPS/revenue numbers.

Key Decisions for Investors

• Do not add directional MU exposure ahead of earnings unless verified consensus and HBM shipment assumptions support it. For existing longs, reduce or hedge into the event with a 1-3 month MU put spread sized to protect against a 10-15% drawdown; retain upside only if management raises HBM revenue, gross-margin, and capex outlook simultaneously.
• Preferred relative-value expression: long AMD / short NVDA in equal dollar amounts over 3-6 months only after AMD demonstrates sequential datacenter revenue acceleration and credible MI-series supply availability. The thesis is that incremental AI system spend broadens beyond GPUs; exit if AMD datacenter guidance disappoints or NVDA's next platform preserves GPU share and pricing more strongly than expected.
• Use MU as a cycle monitor rather than a $2T-momentum chase: initiate a tactical short or long MU put spread if management signals industry bit-supply growth above demand growth, customer inventory rebuilding has normalized, or capex rises materially. These indicators would validate the 12-24 month cyclical-downturn case before spot pricing visibly deteriorates.
• Monitor SK Hynix and Samsung HBM capacity announcements, plus China DRAM/NAND export and yield data, over the next 1-3 months. A faster-than-expected supply ramp is the most direct falsifier of a durable MU margin thesis and would likely pressure the broader memory complex before reported earnings weaken.