Editorials

Micron's fiscal 2026 results represent a dramatic inflection from historical cyclical patterns in the semiconductor industry. The company generated $133.2 billion in revenue with net income of $85.0 billion—a net margin of 63.8%. Operating cash flow reached $89.7 billion. These are not the results of a "normal" year in memory semiconductors; they reflect extraordinary demand and exceptional pricing power.

To put this in context, the company's operating income for the year reached $99.3 billion on that same revenue base—a 74.5% operating margin. Such margins are rarely sustained outside of software monopolies and are almost unheard of in capital-intensive manufacturing.

The acceleration within the fiscal year tells an important story. Quarterly net income progression reveals the surge: Q1 contributed $5.2 billion, Q2 reached $13.8 billion, Q3 jumped to $28.2 billion, and Q4 delivered $37.7 billion. This quartile acceleration—with Q4 alone generating 44% of annual net income—signals both the strength of late-cycle demand and the potential fragility inherent in such concentration.

The company converted $85.0 billion of net income into $89.7 billion of operating cash flow—a conversion ratio of 105%, indicating exceptional cash realization and minimal working capital drain. This efficiency is noteworthy, though it merits scrutiny.

Against this cash generation, Micron deployed $30.7 billion in capital expenditures. The company thus generated $59.0 billion in free cash flow after capex—a figure that dwarfs historical norms and demands interpretation. Management's capital intensity (capex as a percentage of revenue) stood at 23%, up from historical levels, reflecting aggressive capacity expansion to capture AI-driven memory demand.

The cash position at fiscal year-end reached $38.4 billion, while total debt stood at $3.3 billion (current: $0.5 billion; long-term: $2.8 billion). Micron thus enters a potentially softer cycle with fortress balance sheet liquidity, though debt servicing costs had largely compressed to negligible levels ($106 million in interest expense for the full year).

Gross profit reached $107.5 billion on revenue of $133.2 billion, implying a gross margin of 80.7%—extraordinary by any historical standard for commodity memory products. The Q4 results show the magnitude of the demand surge: Q4 revenue of $54.2 billion generated gross profit of $47.0 billion (86.7% gross margin). This compares to Q1's $13.6 billion in revenue yielding $7.6 billion in gross profit (56.1% margin).

Selling, general, and administrative expenses totaled just $1.9 billion—a SG&A ratio of 1.4% of revenue—while research and development consumed $5.7 billion (4.3% of revenue). These ratios reflect Micron's scale advantage, yet they also highlight the leverage embedded in the current earnings structure. A material decline in AI-driven memory demand would compress these margin percentages sharply.

The critical question is whether the AI-driven capacity build represents genuine structural demand or a cyclical surge that will eventually encounter supply-demand rebalancing. Memory historically operates in boom-bust cycles; the current results may represent an exceptional peak rather than a new sustainable baseline.

Management's capital allocation revealed conservative restraint despite extraordinary cash generation. The company repurchased just $650 million in common stock during the fiscal year—a trivial deployment relative to available cash. Dividend payments totaled $610 million, maintaining a minimal payout ratio of 0.7% of net income.

The strategic priority remained capacity expansion. The $30.7 billion capex budget—23% of revenue—signals management confidence in sustained AI memory demand. However, this capex intensity merits scrutiny against the historical memory cycle. If demand moderates faster than supply adjusts, excess capacity and margin compression become material risks.

During the year, Micron repaid $9.3 billion in debt, reducing financial risk even as the company expanded production capacity. This defensive posture—paying down debt while maintaining high capex—suggests management is at least partially hedging against cyclical risk, even if public statements emphasize structural AI tailwinds.

Based on computed metrics, Micron trades at a P/E multiple of approximately 24x on trailing earnings, with an EV/EBITDA of roughly 17.3x. The enterprise value stands near $1.15 trillion on a market cap of approximately $1.17 trillion. The company's price-to-tangible book value exceeds 11.8x, while the P/S (price-to-sales) ratio reaches 13.3x.

These valuations embody an optimistic assumption: that the AI memory cycle will sustain profitability near current levels, or that secular AI demand will prevent the typical memory industry cyclicality. History suggests skepticism is warranted.

If we assume more conservative profitability—say, a normalized operating margin of 30-35% rather than the current 74.5%—normalized earnings would compress to $40-47 billion annually. Against current market value, that would imply a P/E multiple of 25-30x on normalized (not peak) earnings, which is elevated for a cyclical industrial company, no matter the AI narrative.

Owner earnings (operating cash flow minus capex) reached $59.0 billion in FY2026. At current valuation, the company trades at 19.5x owner earnings. Sustainable owner earnings, normalized for a less exceptional cycle, might settle into the $20-30 billion range, implying valuations of 38-58x normalized owner earnings—again, expensive for a cyclical manufacturer.

Micron's business is structurally cyclical. Memory supply and demand have historically followed boom-bust patterns, with extended periods of margin compression following capacity overbuilds. The current AI-driven surge has generated exceptional demand, but several structural risks deserve attention:

Capacity Supply: The capital intensity of memory manufacturing creates a coordination problem. Manufacturers including Micron, Samsung, SK hynix, and others are all expanding capacity in response to AI demand signals. If aggregate capacity additions exceed realized AI memory demand growth, a classic oversupply cycle will emerge within 12-24 months.

Customer Concentration: Approximately half of Micron's revenue flows from its top ten customers—likely dominated by a handful of hyperscale AI companies (OpenAI suppliers, major cloud platforms, etc.). This concentration creates earnings volatility if any major customer moderates capex, as they inevitably will if training-era capex spending gives way to inference-era efficiency.

Pricing Power Transience: The current gross margins of 80%+ reflect acute scarcity pricing for AI-grade memory (especially HBM and leading-edge DRAM). As supply normalizes, pricing power will evaporate rapidly. Memory commoditizes; it always has.

Geopolitical and Trade Risk: The company operates globally, with facilities across Taiwan, Singapore, Japan, the U.S., Malaysia, China, and India. Heightened U.S.-China tension, Taiwan supply-chain risk, and export controls on advanced semiconductor equipment create downside tail risks that earnings models rarely price in appropriately.

Warren Buffett and Charlie Munger have long been skeptical of semiconductor commodity producers, noting that the industry's capital intensity and competitive dynamics destroy shareholder value over time. Memory is perhaps the purest commodity within semiconductors.

Micron's FY2026 results are genuine. The company did generate $85 billion in net income. But this reflects a moment in time—an exceptional convergence of AI infrastructure demand, supply constraints, and pricing power. The question for investors is whether this represents:

1. A new structural baseline (unlikely): AI training demand could remain robust, sustaining HBM and DRAM pricing. This would require AI capex to maintain elevated levels indefinitely, which is improbable given historical IT investment cycles.

2. A peak that moderates gradually (more probable): Gradual supply-demand rebalancing over 2-3 years, with margins declining to 40-50% operating levels as capacity catches up. This is the historical pattern.

3. A peak followed by cyclical collapse (also possible): A sharp correction if AI training capex moderates faster than expected, or if customer concentration creates a demand cliff. Historical memory cycles support this as plausible.

Each scenario has materially different valuation implications. Scenario 1 might support current valuations; Scenario 2 implies 30-50% downside from current levels; Scenario 3 implies 50-70% downside.

Over FY2026, Micron generated $59 billion in free cash flow but deployed almost none to shareholders through buybacks (just $650M) or dividends ($610M). Instead, the company retained this capital and deployed it into expanded manufacturing capacity.

This is rational if management genuinely believes AI memory demand will sustain elevated levels. It is imprudent if management harbors doubts but feels pressured by competition to build. The decision to aggressively capex during peak profitability is exactly when a prudent operator should consider returning cash, since peak profitability typically precedes cyclical downturns.

The company's strong balance sheet—$38.4 billion in cash, minimal debt—provides a cushion for the downside scenario. However, it also represents a potential opportunity cost if near-peak earnings are capitalized into $30.7 billion of capex that must be justified against normalized future returns.

Micron Technology delivered FY2026 results that stand among the best in its history: $133 billion in revenue, $85 billion in net income, and $89.7 billion in operating cash flow. Gross margins reached 80.7% and operating margins 74.5%. These are real, audited results reflecting genuine demand and pricing power.

But extraordinary earnings in commodity memory semiconductors are rarely permanent. The company trades at a P/E of 24x and an EV/EBITDA of 17x, implying that investors have largely priced in the assumption that these exceptional results persist or that the company can redeploy its $30.7 billion annual capex budget into sustained high-return manufacturing.

The memory cycle has not been broken. Rather, AI has temporarily suppressed it through acute demand for training infrastructure and specialized memory (HBM, DDR5). As this capex wave matures and customers optimize for inference efficiency, pricing power will moderate. Normalized operating margins of 30-40% are more defensible than the current 74.5%.

For investors, the critical questions are whether current valuations reflect peak profitability risk, whether the company's aggressive capital deployment will prove justified, and whether Micron's historical cyclical vulnerability has been structurally altered by AI demand. The financial results are stellar. The valuation risk is material.