A Wall Street independent investment research firm, Aletheia Capital, has issued an exceptionally bullish report on Micron Technology (MU), dramatically raising its price target to $1,600 and predicting that AI memory components will account for over 70% of the value within AI hardware systems by 2027, fundamentally reshaping the market's traditional understanding of memory chips' importance.

Following the report's release, Micron's stock surged 10% overnight, rapidly heating up market sentiment. Aletheia forecasts that Micron's earnings per share for the 2027 fiscal year will grow 8.5 times from current levels, expanding another 1.8 times in FY2028, for a cumulative increase of 15 times. The firm also predicts the company will generate cumulative free cash flow between $350 billion and $400 billion over the three fiscal years from FY26 to FY28.

The core logic behind this forecast hinges on the sustained surge in prices for HBM and server DRAM, alongside a structural leap in the proportion of memory components within the bill-of-materials for AI hardware. If these predictions materialize, Micron stands to become one of the biggest beneficiaries of the current AI hardware cycle.

Price Target Jumps to $1,600 with a Complete Valuation Framework Shift

In this move, Aletheia has raised its price target for Micron from a previous $650 to $1,600, an increase of over 140%, implying roughly 47% upside from the current stock price. More notably, its valuation framework has undergone a fundamental transformation—shifting from a conservative pricing model based on historical peak price-to-book ratios to a profit-driven framework anchored on a 10x forward price-to-earnings multiple for fiscal year 2027.

The rationale behind this shift is that, driven by the structural pull from AI on memory demand, Micron's profitability is no longer appropriately measured by cyclical historical valuations but should instead be priced using the earnings multiples typical of growth-oriented technology companies.

HBM and Server DRAM Price Forecasts Both Revised Upward

In terms of price projections, Aletheia's upward revisions also exceed market expectations. The report indicates the firm now expects the average selling price for server DRAM to rise another 30% quarter-over-quarter in Q3 2026, far above its previous forecast of 10% to 15%, with an anticipated further increase of 10% to 15% in Q4 2026.

Regarding HBM, Aletheia predicts its ASP will double year-over-year by 2027. This judgment is based on the continued explosive demand for high-bandwidth memory from AI training and inference, while the high technical barriers on the supply side further support the upward shift in the price equilibrium.

Memory Component BOM Share to Exceed 70%; Vera CPU Rack Price Could Reach $26 Million

The most striking assertion in the report comes from its analysis of the bill-of-materials structure for AI hardware systems. Aletheia points out that the comprehensive value share of memory components in AI hardware systems is expected to jump from the mid-40% range in 2025 to over 70% by 2027, indicating memory is becoming the most critical core component in AI hardware systems.

Using Nvidia's Vera CPU as an example, the report notes that the SoCAAM component alone already accounts for over 70% of the BOM in the second half of 2026. Furthermore, a fully configured Vera CPU rack could have an average selling price as high as $26 million.