Micron's AI Boom Lifts Earnings 1,370%, but Market Still Doubts It Will Last

Micron Technology's earnings have undergone a historic revaluation driven by its emergence as a critical supplier of high-bandwidth memory (HBM) for artificial intelligence systems. The company's earnings per share have climbed from a negative $5.34 in fiscal year 2023 to an estimated $73.23 in fiscal year 2026—a trajectory that has prompted Wall Street to fundamentally reassess the company's long-term earning power. Over the last year, average earnings projections for Micron have been revised upward by 794%, reflecting the magnitude of this repricing.

The operational catalyst behind this shift is clear: high-bandwidth memory, stacked directly onto powerful AI processors like NVIDIA's next-generation graphics processing units, has become the key system bottleneck for both training and deploying large language models. Micron has evolved from a commodity DRAM and NAND supplier into an integral component of the global AI hardware architecture. This shift was visible in the company's third-quarter fiscal 2026 results, when revenue surpassed consensus projections of $35.69 billion（約5.7兆円）.

Despite this momentum, the market's valuation suggests lingering doubt. Micron trades at a forward price-to-earnings multiple of just 6.2x, even though the company reports a return on equity of 66.6% and gross margins of 84%. This discount reflects concern that Micron will eventually revert to its historical boom-bust pattern. The company's prior downturn offers a cautionary tale: gross margins collapsed from 45.2% in 2022 to a mere 2.7% in 2023, demonstrating how quickly semiconductor cycles can deteriorate when end-market demand falls or global capacity overshoots.

However, Micron argues that this cycle is structurally different. The company has secured long-term supply agreements totaling more than $100 billion（約16兆円） in minimum contracted revenue, with contractual price floors designed to protect gross margins above 70%. Management has also indicated that Micron's entire HBM production allocation for fiscal year 2026 has already been sold out, and supply shortages are likely to persist beyond 2027. These multi-year commitments across the AI ecosystem suggest that Micron's current profitability may rest on firmer ground than prior cycles. Nevertheless, high-profile skeptics, including investor Michael Burry, have taken public short positions, betting that capacity expansion will eventually swamp the market and trigger another major margin contraction. The tension between Micron's structural demand hedges and the market's historical skepticism remains unresolved.

Micron's transformation from a cyclical commodity memory maker into an AI infrastructure cornerstone has been rapid and profound. The company's shift reflects a structural reality: high-bandwidth memory stacked directly on AI processors like NVIDIA's GPUs has become the key bottleneck for both training and deploying large language models. This transition is evidenced not only by the near-1,400% earnings upside revision but also by the company's Q3 fiscal 2026 revenue beat—exceeding consensus by an undisclosed margin against a $35.69 billion（約5.7兆円） baseline.

Yet the market's valuation—a mere 6.2x forward multiple despite 66.6% return on equity and 84% gross margins—reveals deep skepticism about the durability of this boom. That skepticism is rooted in Micron's brutal history: the prior cycle saw gross margins crater from 45.2% in 2022 to 2.7% in 2023. Investors worry that capacity expansion will eventually saturate the market and trigger another margin collapse, a concern amplified by public short positions such as Michael Burry's.

However, the structural case for durability rests on multi-year supply agreements totaling more than $100 billion（約16兆円） in contracted revenue, backed by contractual price floors. With Micron's HBM production already fully allocated through fiscal 2026 and supply shortages expected to persist beyond 2027, the company has arguably found a moat—at least in the near term. The question investors face is whether these contractual commitments can withstand a broader AI spending correction or whether they merely delay the inevitable cyclical downturn.

No comments yet. Be the first to share your thoughts!