Micron's recent earnings report and guidance have pushed the "memory cycle" narrative into a more aggressive phase: simultaneous increases in revenue, gross margin, and cash flow are forcing the market to recalibrate its valuation model and adjust its profit expectations upward. Barclays has raised its 2027 earnings per share forecast to a three-digit figure, lifting its price target from $450 to $675 while maintaining an Overweight rating. Based on the closing price of $461.73 on March 18, this implies approximately 46% upside potential.
Barclays analyst Tom O'Malley stated in a recent report that the company's revenue guidance is "42% above Wall Street expectations, coupled with a gross margin guidance of 81%, which fundamentally alters our model. We now project 2027 EPS will exceed $100." The consensus estimate, sourced from Bloomberg as of March 18, places Micron's FY2027 EPS expectation at $54.81, creating a stark contrast.
The core of this "super double" model is not a sudden, broad-based demand recovery, but rather supply-side constraints. Barclays believes that both DRAM and NAND will remain in an extremely tight supply-demand balance through 2026. Management indicated that key customers can only meet 50% to 67% of their own medium-term demand, making significant price reductions unlikely.
This optimism does not ignore headwinds: the company anticipates low double-digit percentage declines in PC and smartphone shipments for 2026. However, Barclays emphasizes the accelerating consumption of memory by data centers and AI—highlighting the HBM roadmap, capital expenditure expansion, and the signing of a first-ever 5-year strategic customer agreement as factors providing greater visibility and higher profit leverage.
The numbers within the guidance triggered this "model reset." For the February quarter, Micron's performance and pricing power exceeded expectations: revenue reached $23.9 billion (versus Wall Street's $19.7 billion), adjusted gross margin was 74.9% (versus 69.1%), and EPS was $12.20 (versus $9.00). Both DRAM revenue of $18.8 billion and NAND revenue of $5.0 billion surpassed consensus estimates.
What forced the market to recalculate was the May quarter guidance: revenue of $33.5 billion (versus Wall Street's $23.7 billion), gross margin of 81.0% (versus 72.4%), and operating expense guidance of $1.40 billion, below the expected $1.51 billion, leading Barclays to project an EPS of $19.15 (versus Wall Street's $11.29).
Prices are still rising. Barclays attributes this earnings revision almost entirely to a combination of "strong pricing amid tight supply, cost execution, and product mix." The February quarter data provides direct evidence: DRAM bit shipments grew by a mid-single-digit percentage quarter-over-quarter, while ASP increased by a mid-60s percentage. NAND bit shipments saw low-single-digit growth, with ASP rising by a high-70s percentage.
For the May quarter, Barclays continues to model ASP increases: DRAM ASP up 34% quarter-over-quarter and NAND ASP up 43%, predicated on continued industry supply tightness.
"Customers can only meet 50%–67% of demand." A particularly significant piece of management feedback cited in the report is that key customers can only fulfill 50% to 67% of their own medium-term demand. Barclays' conclusion is direct: the favorable pricing environment will persist and is unlikely to peak rapidly within a single quarter.
Supply-side constraints are also more specific: DRAM supply growth is limited by cleanroom space, construction cycles, a higher conversion rate to HBM, and a decline in bits per wafer. NAND faces similar cleanroom space limitations, with some suppliers reallocating cleanroom resources to DRAM.
The HBM roadmap is entering the delivery phase. Barclays views "increasing AI memory content" as the primary variable for navigating the cycle. The company disclosed it has begun shipping HBM4 12H 36GB (for Nvidia's Vera Rubin platform) in the first quarter of 2026, while also sampling HBM4 16H 48GB; HBM4E is under development, with volume ramp expected in 2027.
When asked about Groq's LPU and its use of SRAM during the earnings call, the company responded that the impact is "more complementary": the evolution of AI architectures will drive higher HBM/DRAM content per standard rack, rather than replacing the need for memory.
PC and smartphone markets are declining, while data center share is rising. Management expects low double-digit percentage declines in both PC and smartphone shipments for 2026. However, data center volumes are growing: the company anticipates that in 2026, the data center DRAM and NAND bit TAM will, for the first time, exceed 50% of the total industry TAM. Server shipments are projected to grow by low double digits, with new platform introductions also driving higher DRAM content per server.
Cash flow and capital expenditure are expanding simultaneously. Barclays emphasizes that this cycle is not about "squeezing profits through low capital investment." Management expects FQ3 cash flow to double sequentially (from FQ2's $5.5 billion), with Barclays estimating FQ3 cash flow of $13.7 billion against quarterly capex of approximately $7.0 billion.
The direction of capital spending is also clearer: the company guides for FY26 net capex to exceed $25.0 billion (previously around $20.0 billion), with the increase primarily from cleanroom investments. Total capex for FY27 is expected to rise significantly, with construction-related capex increasing by over $10.0 billion year-over-year to support HBM and DRAM investments.
Five-year SCA and Tongluo expansion. The company announced its first 5-year Strategic Customer Agreement (SCA), differing from the typical 1-year Long-Term Agreement (LTA), aiming to enhance business stability and visibility. Concurrently, it plans to commence construction of a second cleanroom of equal scale at the Tongluo site (acquired from Powerchip) by the end of FY26, dedicated to DRAM production.
Within Barclays' framework, these two developments extend the "favorable pricing" environment from a short-term upcycle into a longer-term certainty.
Price target raised, valuation multiple lowered. Barclays raised its price target from $450 to $675 but simultaneously lowered the valuation multiple: the new target is based on 6.3 times CY27 EPS of $106.77. The previous $450 target corresponded to 11.1 times CY27 EPS of $40.46. The rationale for the lower multiple is an expectation that the pace of sequential price increases will slow from here, but the base level of earnings has been substantially elevated.
The earnings gap itself is telling: Barclays' adjusted EPS forecast for FY2027 is $102.53, far exceeding the Bloomberg consensus of $54.81. Its FY2026 EPS forecast of $57.91 also surpasses the consensus of $37.54.
Upside $800, Downside $400. The report provides clear scenario boundaries: an upside scenario of $800, valued at 7 times CY27 EPS of $115, assumes faster HBM growth, stronger pricing, and better cost performance. A downside scenario of $400, valued at 4.7 times CY27 EPS of $85, corresponds to weaker AI demand, aggressive supply expansion by competitors, and a longer-than-expected price decline cycle.
Barclays' thesis is clear: if supply tightness persists into 2026 and HBM ramps according to plan, triple-digit EPS is not a peak fantasy but the central assumption they are willing to back with their price target.
Comments