Micron’s 239.2% Surge Leaves TSMC’s Capacity Crunch Less Rewarded

Micron shares appreciated 239.2% over the year through Nov. 4, 2026, versus 29.50% for TSMC, a 209.70-percentage-point gap in how the market has rewarded scarce memory over constrained foundry capacity.

The divergence runs through the semiconductor industry’s AI investment cycle. Micron, SK Hynix and Samsung Electronics have benefited from rising prices for DRAM and high-bandwidth memory as AI infrastructure absorbs more of both products. TSMC faces a different bottleneck: insufficient leading-edge manufacturing and advanced packaging capacity.

Dr. Robert Castellano frames the split directly: “The Market Is Rewarding Memory Scarcity, but Overlooking Foundry Capacity.”

What has the market rewarded?

Micron’s performance reflects investor confidence that shortages can become higher selling prices and stronger earnings. The analysis notes that Micron’s one-year return has declined substantially since Oct. 16, yet investors have continued to reward that pricing power.

TSMC has received a much smaller market reward for its own capacity constraints. Its 29.50% appreciation over the same period trails Micron’s by more than 200 percentage points. The investment question is whether that disparity reflects genuinely different earnings prospects or an underappreciation of TSMC’s ability to convert constrained capacity into sustained revenue growth.

This is a question about earnings conversion, not whether both sides face shortages. Memory suppliers can monetize constrained DRAM and HBM supply through price. TSMC must monetize added leading-edge wafer fabrication and advanced packaging into sustained revenue growth, while the selected expansion milestones run from 2026 to 2029.

How does capital spending change the picture?

The analysis turns to a 2023-2027 capital-expenditure comparison covering TSMC and major memory manufacturers, including Micron, SK Hynix and Samsung. It treats the spending responses of all four companies as evidence of how aggressively each side is addressing insufficient capacity.

The 2023-2027 horizon is a spending roadmap, not proof of output. The analysis discloses no company-by-company capex totals, wafer-start additions, process-node targets, advanced-package capacity or revenue forecasts, and it offers no confirmed 2027 volume or analyst revenue estimate. Investors therefore have stock returns and shortage claims, but no quantified bridge from spending to future shipments.

Could memory scarcity persist through 2027?

The analysis explicitly asks why memory shortages could persist through 2027, with rising AI demand for DRAM and HBM as the commercial mechanism behind that view. Its counterweight is equally important: the current shortage could eventually give way to another cyclical oversupply.

That two-sided case explains why memory equities can outperform even as their own expansion raises the risk of later price pressure. The 2023-2027 spending horizon frames the period in which new supply can arrive while investors still price scarcity. The analysis offers no contract-price forecast or inventory estimate with which to judge that timing.

Why does TSMC’s expansion differ?

TSMC’s expansion addresses a different kind of scarcity. Its constraint concerns leading-edge manufacturing and advanced packaging capacity, so the commercial payoff depends on bringing planned fabrication and packaging capacity online and converting it into sustained revenue. The analysis does not name a process node, wafer size or package-throughput target.

A table of selected semiconductor manufacturing capacity expansion milestones covers 2026-2029. Those dates define the roadmap horizon; they do not establish confirmed wafer capacity or future revenue. The analysis also asks how quickly new facilities can begin production once spending turns into usable capacity.

That timing can determine whether the current shortage stays a shortage. A long interval before output begins supports TSMC’s revenue opportunity, while a rapid ramp across the industry could ease constraints sooner. The analysis offers no production-start dates, however, so timing remains an unresolved variable rather than a forecast.

What is the investor takeaway?

The market has rewarded memory scarcity more visibly than foundry scarcity, but stock performance alone cannot settle the earnings question. Micron, SK Hynix and Samsung have direct price exposure to DRAM and HBM; TSMC has exposure to leading-edge manufacturing and advanced packaging throughput, with the analysis claiming insufficient capacity in both.

For TSMC, proof will have to come from sustained revenue growth as new capacity enters production, not from the announcement of spending alone. For memory makers, the equivalent proof is continued pricing strength as the 2023-2027 investments mature, without letting new supply trigger the cyclical oversupply the analysis flags.

Watch the 2026-2029 expansion milestones and the speed of production ramps: if TSMC converts planned capacity into revenue while memory supply catches up, the market’s 239.2%-versus-29.50% reward gap may narrow, but persistent strength in memory pricing would keep the advantage with Micron, SK Hynix and Samsung.