Micron Technology
Micron Technology, Inc. is an American semiconductor company that designs, manufactures, and sells memory and storage products worldwide, including dynamic random-access memory (DRAM), NAND flash memory, and solid-state drives.[1][2] Founded in 1978 as a small semiconductor design firm in the basement of a Boise, Idaho dental office, it is headquartered in Boise and has expanded to operate manufacturing facilities across multiple countries.[3][4] In fiscal year 2025, Micron reported record revenue of $37.4 billion, driven primarily by demand for its DRAM and NAND products used in computing, data centers, mobile devices, and automotive applications.[5][4] The company markets consumer-oriented products under the Crucial brand and gaming memory under Ballistix, positioning itself as a key innovator in memory solutions that support data-intensive technologies.[1][6]
Overview
Company Profile
Micron Technology, Inc. was founded on October 5, 1978, in Boise, Idaho, by Ward Parkinson, Joe Parkinson, Dennis Wilson, and Doug Pitman as a semiconductor design and manufacturing company initially focused on MOS memory chips, beginning operations as a four-person team in the basement of a local dental office.[7][1][8] The company is headquartered in Boise and operates as a publicly traded entity on the NASDAQ stock exchange under the ticker symbol MU.[2] As of fiscal year 2025, ending August 2025, Micron reported annual revenue of $37.38 billion, propelled by heightened demand for its memory products in artificial intelligence applications and data centers.[5] The firm employs approximately 53,000 people globally and maintains a market capitalization of around $246 billion as of October 2025.[9][10] Micron's operational scope encompasses serving key sectors such as data centers, automotive systems, and industrial equipment through its semiconductor memory solutions, following a December 2025 announcement to exit the Crucial consumer business by the end of fiscal Q2 2026 (February 2026) to prioritize enterprise, AI, and data center markets.[11][12][13]
Core Business and Market Focus
Micron Technology's core business revolves around the design, manufacturing, and marketing of dynamic random-access memory (DRAM) and NAND flash memory semiconductors, which constitute the foundational hierarchy for data processing and storage in computing architectures. DRAM provides volatile, high-speed access for active workloads, while NAND enables non-volatile retention for persistent data, directly supporting causal dependencies in systems ranging from mobile devices to hyperscale servers where memory bottlenecks limit overall throughput.[1][14] In fiscal year 2024, DRAM accounted for approximately 70-73% of revenue, with NAND contributing 25-26%, reflecting the dominance of these segments amid industry cycles characterized by pricing fluctuations from supply gluts and shortages tied to fabrication capacity expansions. This revenue structure underscores Micron's exposure to memory market dynamics, where empirical oversupply in prior years depressed margins, but targeted demand recovery in compute-intensive sectors has driven rebound. The company's strategic orientation prioritizes enterprise-grade solutions, including its December 2025 announcement to exit the Crucial consumer business by February 2026 in order to reallocate DRAM and NAND capacity toward enterprise and AI applications, as well as solid-state drives (SSDs) derived from NAND and DRAM variants for data centers, amplifying revenue from cyclical upswings without diversifying beyond core memory hierarchies.[13][15][16] Micron's market focus emphasizes applications demanding high memory density and bandwidth, such as AI accelerators and data center infrastructure, where advancements in proprietary process nodes—like the 1γ (1-gamma) DRAM technology employing extreme ultraviolet lithography—enable denser integration and power efficiency gains that empirically enhance computational performance per watt. These innovations refute commoditization narratives by establishing causal advantages in scaling memory alongside processor evolution, as evidenced by record data center revenue tied to AI-driven demand for elevated capacities.[17][12][14]History
Founding and Early Years (1978–1999)
Micron Technology was founded on October 5, 1978, in Boise, Idaho, by Ward Parkinson, Joe Parkinson, Dennis Wilson, and Doug Pitman as a four-person semiconductor design consulting firm operating from the basement of a local dental office.[3] The founders, leveraging their engineering expertise, secured an initial contract to design a 64K dynamic random-access memory (DRAM) chip for Mostek Corporation, marking the company's entry into memory semiconductor development amid a burgeoning U.S. industry dominated by integrated circuit innovation.[8] By 1980, Micron broke ground on its first fabrication facility in Boise to transition from design consulting to manufacturing, completing the plant in 1981 and shipping its inaugural 64K DRAM product that year for sale in the merchant market.[3] This vertical integration allowed Micron to produce and sell chips directly, positioning it as one of the few American firms pursuing independent DRAM production during an era of rapid technological scaling and intense global competition.[18] Early funding came from personal investments by the founders and subsequent backing from local investors, including potato magnate J.R. Simplot, enabling facility construction without heavy reliance on federal subsidies.[19] The mid-1980s brought severe challenges, including a sharp industry downturn in 1985 triggered by oversupply and pricing pressures from Japanese competitors, which led to Micron posting a $19 million loss in the second half of its fiscal year ending August 28, 1985.[20] To survive, the company implemented aggressive cost-cutting, including layoffs—its only such action under original management—and pivoted toward process improvements and higher-margin designs, refusing to exit the DRAM market unlike many U.S. rivals.[21] Micron also pursued legal recourse, filing antidumping petitions against Japanese firms for predatory pricing, which highlighted early patterns of resilience through litigation and operational efficiency rather than protectionist bailouts.[18] Throughout the 1990s, Micron concentrated on scaling to higher-density DRAM products, such as 256K and 1Mb generations, to counter Japanese dominance in production volume and cost advantages subsidized by government policies.[19] Despite market cycles and foreign competition capturing over 80% of global DRAM share by mid-decade, Micron's focus on proprietary process technologies and vertical control over design-to-fabrication sustained its viability as a U.S.-based merchant producer, going public in 1984 and achieving profitability through cycles via disciplined capital allocation.[18] This era established Micron's strategy of competing on technological merit amid asymmetric trade pressures, without exiting core memory markets.[22]Expansion and Acquisitions (2000–2010)
In the early 2000s, Micron Technology pursued strategic acquisitions to bolster its position in the consolidating memory industry, where overcapacity and cyclical pricing pressures necessitated consolidation of intellectual property and manufacturing assets for competitive survival. A key move was the acquisition of Lexar Media, Inc., announced on March 8, 2006, and completed on June 21, 2006, for approximately $850 million in stock. [23] This purchase expanded Micron's footprint in consumer-oriented flash memory products, such as digital media cards and USB drives, enabling direct retail distribution under the Crucial brand and reducing reliance on commoditized DRAM sales amid industry oversupply.[24] By 2010, Micron further strengthened its NAND flash capabilities through the acquisition of Numonyx B.V., a NOR and NAND flash joint venture spun off from Intel Corporation and STMicroelectronics N.V. The deal, valued at $1.27 billion in Micron stock, was announced on February 9, 2010, and closed on May 7, 2010.[25] [26] Numonyx's assets, including fabrication facilities and technology licenses, provided Micron with diversified non-DRAM revenue streams and helped mitigate risks from DRAM's boom-bust cycles, where excess capacity had driven prices down sharply.[27] These expansions occurred against a backdrop of severe industry challenges, particularly DRAM price collapses exacerbated by the 2008-2009 global recession, which led to widespread overproduction and demand contraction.[28] Micron's revenue, which had approached $16 billion in fiscal 2006, plummeted to $2.1 billion by fiscal 2009 due to falling average selling prices and inventory writedowns.[29] In response, the company initiated major restructurings, including a October 2008 plan to reduce its global workforce by 15% (approximately 2,850 positions from 19,000 employees) over two years, targeting high-cost operations and underutilized NAND production.[30] [31] This was followed by an additional 2,000 job cuts announced in February 2009, comprising 12% of the remaining staff, as pragmatic measures to control costs and align capacity with subdued demand rather than pursuing unchecked volume growth.[32]Strategic Realignments and Growth (2011–Present)
In the early 2010s, Micron advanced its NAND flash capabilities through the IM Flash Technologies joint venture with Intel, established in 2006, which facilitated collaborative development of 3D NAND technology to overcome planar scaling limitations and enhance density.[33] By 2015, the partnership commercialized 32-layer 3D NAND, achieving three times the capacity of prior generations with full production ramping up in Q4 that year, enabling higher bit densities essential for SSDs and mobile storage amid surging data demands.[34] Micron exercised its option to acquire Intel's remaining stake in IM Flash in 2018, completing the buyout by 2019 for approximately $1.5 billion, allowing full control over NAND production scaling while navigating cyclical memory markets.[35] Entering the 2020s, Micron pivoted toward high-bandwidth memory (HBM) solutions optimized for AI accelerators, shifting from earlier pursuits like Hybrid Memory Cube around 2018 to capitalize on GPU-integrated demands for ultra-high throughput in data centers.[36] This focus on HBM, adopted similarly by competitors such as Samsung, has exacerbated a global memory chip shortage driven by surging AI demand—for instance, Nvidia's GB200 GPU requires 192 GB of HBM3e memory per GPU—constraining supplies of commodity NAND and DRAM used in consumer products. The reallocation of production capacity to high-end AI chips has led to NVMe SSD sell-outs, NAND prices doubling within months, and purchase limits imposed in Asian consumer markets.[37][38][39][40] This realignment positioned HBM as a core offering, with production ramps supporting AI workloads requiring low-latency access to massive datasets, as evidenced by HBM revenue nearing $2 billion in Q4 FY2025 alone.[41] In April 2025, Micron reorganized into four market-segment-based units—data center, mobile, automotive/industrial, and client—to streamline AI-focused innovation and customer alignment, enhancing responsiveness to segment-specific growth drivers like AI training inference.[42] In December 2025, Micron announced its exit from the Crucial consumer business by the end of fiscal Q2 2026 (February 2026), discontinuing sales of consumer RAM and SSDs to redirect resources toward prioritizing supply for strategic AI and data center customers. This shift has further intensified shortages and price hikes for consumer PC components, including DDR4 and DDR5 RAM—with some 32GB DDR5 kits rising nearly fourfold from around $82 to $310—amid the broader industry reallocation to AI priorities.[13][39][43] Micron's FY2025 revenue reached a record $37.4 billion, up 49% year-over-year, propelled by data center demand surges from AI infrastructure where HBM and DRAM constituted over half of sales, underscoring causal links between compute scaling and memory bandwidth needs.[5] To address supply chain vulnerabilities exposed by geopolitical tensions and reliance on Asian fabrication, Micron announced expansions including two fabs in Idaho and two in New York, backed by up to $6.1 billion in CHIPS Act direct funding finalized in December 2024, part of a $125 billion private investment to onshore leading-edge memory production.[44] These subsidies empirically counter Chinese state-backed overproduction—evident in below-cost DRAM dumping that eroded Western market shares—but introduce risks of government dependency, potentially distorting price signals and long-term efficiency in a commoditized industry.[45]Products and Technologies
DRAM and Synchronous Memory Solutions
Micron Technology manufactures synchronous dynamic random-access memory (SDRAM) solutions, including DDR4 and DDR5 standards for enterprise servers and personal computers, as well as low-power DDR (LPDDR) variants such as LPDDR5X for mobile devices and automotive applications. DDR5 SDRAM provides up to twice the effective bandwidth of DDR4, supporting data rates up to 8,400 Mb/s and enabling higher performance in bandwidth-intensive workloads.[46] LPDDR5X achieves speeds of up to 8,533 Mb/s with 20% greater power efficiency compared to LPDDR4X, facilitating extended battery life in smartphones and laptops.[47] Micron's advancements in DRAM process technology include the 1β (1-beta) node introduced in November 2022, which yields 16 Gb dies with 35% higher bit density and 15% improved power efficiency over prior generations, adhering to scaling trends akin to Moore's law where densities approximately double every two years through transistor and cell shrinks.[48] Subsequent 1α and 1γ nodes extend these gains; the 1α process supports fast LPDDR5 speeds in low-power mobile DRAM, while 1γ enables modules like the 192 GB SOCAMM2 announced in October 2025, delivering over 20% power efficiency improvements for AI server applications by adapting mobile-optimized low-power DRAM to data center needs.[49][50] These nodes reduce operating voltages and enhance transistor performance via high-k metal gate (HKMG) structures, minimizing leakage in high-density arrays. As the core of system memory in computing platforms, Micron's DRAM underpins data processing in servers, PCs, and mobile systems, with bit shipments rising in low-teens percentages sequentially in fiscal 2025 amid AI-driven demand, outpacing supply and driving contract prices upward by 10-30% in tight inventory conditions.[51][52] Pricing cycles correlate inversely with inventory levels, as evidenced by recent surges tied to constrained supply rather than overproduction, contrasting historical booms and busts.[53] Micron maintains leadership in low-power variants, shipping LPDDR-derived modules for energy-efficient servers that reduce data center power draw while sustaining high throughput.[54] In December 2025, Micron announced its exit from the Crucial consumer RAM business by February 2026 to prioritize AI and data center demand, contributing to global consumer RAM shortages and price hikes of 20-500%.[13][38]NAND Flash and Solid-State Storage
Micron Technology has advanced NAND flash memory from planar architectures to stacked 3D NAND configurations, enabling higher densities and driving the adoption of solid-state drives (SSDs). The transition to 3D NAND involved vertically stacking memory cells, with Micron achieving milestones such as 176-layer technology for improved read and write latencies over prior 128-layer generations. By 2024, Micron introduced 232-layer NAND as its sixth generation, followed by 276-layer 3D NAND in products like the 2650 client SSD, representing ninth-generation (G9) advancements with transfer speeds up to 3.6 GB/s.[55][56][57] Cell technologies in Micron's NAND include triple-level cell (TLC) for balanced cost and performance, and quad-level cell (QLC) for greater density at the expense of write endurance, suitable for read-intensive applications. QLC NAND stores four bits per cell, offering up to 2Tb die sizes compared to TLC's three bits, but requires innovations like Micron's Adaptive Write Technology, which dynamically caches data in SLC and TLC tiers to enhance QLC write speeds to TLC-like levels in SSDs such as the 2600 series.[58][59][60] Micron offers SSDs under the Crucial consumer brand and its enterprise lineup, targeting client and data center markets with high endurance metrics. For instance, the Crucial P3 1TB SSD provides 220 TBW (terabytes written) endurance, while enterprise models like the 5400 SATA series deliver up to 50% greater endurance than competitors, supporting petabyte-scale deployments such as 2.5 PB per 2U rack with the 6550 ION SSD. In fiscal 2025, Micron's NAND revenues reached $8.5 billion, reflecting SSD demand growth amid AI and cloud applications.[61][62][63] The NAND market experiences cyclical pricing due to supply-demand imbalances, with oversupply in 2018-2019 causing sharp price declines after prior peaks, as bit growth outpaced demand. Micron and peers addressed this through capacity discipline, stabilizing supply and enabling recovery, which supported SSD proliferation by improving cost-density tradeoffs via 3D stacking and multi-bit cells.[64][65]High-Bandwidth Memory and Emerging Innovations
Micron Technology commenced volume production of its HBM3E memory in February 2024, featuring an 8-high stack configuration with 24 GB capacity and pin speeds of 9.2 Gbps, delivering over 1.2 TB/s bandwidth per stack.[66] This solution utilizes through-silicon vias (TSVs) for vertical stacking of DRAM dies, enabling high-density integration critical for AI and high-performance computing (HPC) applications, and has been qualified for integration into NVIDIA's H200 Tensor Core GPUs, with shipments beginning in the second quarter of 2024.[67] By September 2024, Micron advanced to 12-layer HBM3E stacks offering 36 GB capacity, further supporting NVIDIA platforms like HGX B300 for enhanced AI workloads.[68][69] Integration challenges in HBM production include maintaining yield rates amid complex TSV alignment and stacking processes, which demand precise thermal and mechanical controls to avoid defects in multi-die assemblies. Micron has addressed yield limitations through adoption of extreme ultraviolet (EUV) lithography in its 1γ DRAM node, achieving 30% higher bit density, 20% lower power consumption, and up to 15% performance gains compared to prior nodes, thereby improving manufacturability for HBM stacks.[70][71] Beyond HBM, Micron offers GDDR6X graphics memory with 16 Gb dies supporting data rates up to 24 Gbps, enabling system bandwidth exceeding 1 TB/s for discrete GPUs in gaming and professional visualization, distinct from HBM's stacked architecture but complementary for bandwidth-intensive graphics rendering.[72] In disaggregated computing, Micron provides CXL-based memory expansion modules that enhance capacity and bandwidth with low-latency cache coherency across heterogeneous systems, facilitating pooled memory resources for scalable AI training clusters.[73] Projections for HBM demand are tied to empirical growth in AI training requirements, with the market expanding from approximately $18 billion in 2024 to $35 billion in 2025, driven by hyperscaler capital expenditures on GPU-accelerated infrastructure rather than unsubstantiated hype.[74] Micron's HBM supply has been allocated through most of 2025, reflecting constrained availability amid verified increases in data center memory needs for large-scale model training.[75]Operations
Global Manufacturing Facilities
Micron Technology operates a network of semiconductor fabrication facilities (fabs) primarily focused on dynamic random-access memory (DRAM) and NAND flash production, with a strategic emphasis on 300mm wafer processing to achieve economies of scale. Key manufacturing sites include its headquarters and primary R&D-integrated fabs in Boise, Idaho, USA; back-end assembly and testing in Manassas, Virginia, USA; wafer fabrication in Singapore; DRAM production in Hiroshima, Japan (acquired via the 2013 Elpida Memory purchase); and operations in Taichung, Taiwan, alongside assembly sites in Muar and Penang, Malaysia, and emerging facilities in Gujarat, India.[4] These locations support front-end wafer processing and back-end packaging, with Singapore and Japan serving as high-volume DRAM hubs.[76] The company's global fab capacity exceeds 1 million wafers per month equivalent, predominantly from 300mm lines optimized for advanced nodes, though exact figures fluctuate with demand-driven utilization rates that historically range from 70-90% during peak cycles to lower levels in downturns.[77] Efficiency metrics, such as yield rates, are closely tied to process maturity and regional supply chain stability, with Asian fabs often achieving higher utilization due to proximity to raw materials and equipment suppliers, but U.S. sites benefiting from lower geopolitical exposure.[78] In response to escalating geopolitical tensions, particularly risks associated with Taiwan's dominance in semiconductor supply and U.S.-China trade restrictions, Micron has pursued diversification by expanding domestic U.S. capacity. Following the 2018 dissolution of its NAND joint venture with Intel, which repatriated some assets to the U.S., the company announced a $100 billion megafab cluster in Clay, New York, commencing construction in 2024 with plans for up to four 300mm fabs to produce leading-edge DRAM, supported by CHIPS and Science Act funding.[79] This initiative, alongside expansions in Idaho and Virginia totaling up to $200 billion in investments, aims to onshore critical production and mitigate disruptions from potential Taiwan Strait conflicts or Chinese export controls on Micron products.[80] Such moves address empirical vulnerabilities, as over half of U.S.-owned IC capacity historically resides overseas, exposing firms to tariffs, IP theft risks, and supply interruptions.[77][81]Research, Development, and Supply Chain
Micron Technology allocates substantial resources to research and development, with fiscal 2024 expenses reaching approximately $2.9 billion, reflecting a 10% increase from the prior year driven by higher personnel costs and prototyping activities.[82] This investment, centered at facilities in Boise, Idaho, supports advancements in memory scaling amid physical constraints like diminishing returns in transistor density and rising power leakage at sub-10nm nodes.[83] The company has amassed over 60,000 patents worldwide as of October 2025, underscoring a focus on proprietary innovations in materials engineering and process integration.[3] The innovation pipeline emphasizes node shrinks governed by semiconductor physics, including quantum tunneling limits and thermal dissipation challenges, rather than external mandates. Micron's 1γ (1-gamma) DRAM node, entering production in 2025, achieves over 30% higher bit density per wafer than the preceding 1β node through extreme ultraviolet (EUV) lithography, high-k metal gate structures, and backend-of-line optimizations, enabling DDR5 modules up to 9200 MT/s while curbing power draw.[17][84] Initial sampling of 1γ-based LPDDR5X devices occurred in mid-2025 for integration into 2026 mobile platforms, prioritizing bandwidth gains for AI inference workloads.[85] Micron's supply chain hinges on critical upstream providers, including ASML for EUV lithography scanners essential to sub-10nm patterning and Applied Materials for thin-film deposition and etching tools, alongside silicon wafer suppliers like Shin-Etsu Chemical.[86][87] These dependencies amplified vulnerabilities during 2020-2022 COVID-era disruptions, which caused wafer and equipment shortages exacerbating global memory imbalances. In response, Micron has pursued onshoring via expanded U.S. investments totaling over $15 billion by mid-2025 in Idaho and New York fabs, aiming to localize assembly and testing while retaining overseas specialization for cost-physics trade-offs.[88] Such measures address logistical chokepoints but stem fundamentally from yield imperatives in advanced nodes, where proximity reduces defect rates over long-haul shipping.[89]Financial Performance
Historical Revenue and Profitability
Micron Technology's revenue and profitability have historically fluctuated dramatically, reflecting the semiconductor memory industry's boom-bust cycles primarily caused by lagged capital expenditures relative to demand shifts, resulting in periodic overcapacity and inventory gluts.[90] During upcycles, strong demand from consumer electronics, PCs, and later data centers drives pricing power and high utilization rates, while downturns expose mismatches where prior investments in fabrication capacity flood the market amid weakening end-demand.[91] This volatility has led to gross margins swinging from negative territory to over 50%, with net losses in trough years offsetting prior peaks.[92] In the 2000s, Micron navigated multiple cycles tied to DRAM commoditization and global economic shocks; revenue grew from approximately $3 billion in fiscal year 2000 to a peak of $5.27 billion in 2006, before contracting to $2.58 billion in 2009 amid the financial crisis and oversupply.[93] Profitability mirrored these trends, with net income turning negative in bust phases due to pricing erosion and high fixed costs from underutilized plants. The 2010s saw recovery and further peaks, exemplified by fiscal 2018 revenue of $30.39 billion and net income exceeding $6 billion, fueled by NAND expansion and tight supply, though this was followed by softer conditions in 2019.[94][95] More recently, fiscal 2022 revenue reached about $30.76 billion with robust profits, but 2023 marked a trough at $15.54 billion in revenue and a $5.83 billion net loss, attributable to post-pandemic inventory corrections and delayed capex adjustments by competitors.[96] Recovery accelerated in fiscal 2024 to $25.11 billion revenue and $778 million net income, escalating to a record $37.38 billion in fiscal 2025 amid AI-driven demand for high-bandwidth memory.[96][97]| Fiscal Year | Revenue (USD billions) | Net Income (USD billions) | Notes |
|---|---|---|---|
| 2018 | 30.39 | 6.29 | Peak driven by NAND ramp-up[94][98] |
| 2022 | 30.76 | Positive (pre-trough high) | Strong DRAM/NAND pricing[96] |
| 2023 | 15.54 | -5.83 | Inventory glut and oversupply[96][98] |
| 2024 | 25.11 | 0.778 | Rebound with AI tailwinds[96][98] |
| 2025 | 37.38 | 8.54 | Record amid data center growth[97] |
Key Metrics and Investor Relations Trends
Micron Technology reported fiscal year 2025 revenue of $37.4 billion, a record high and nearly 50% increase from the prior year, propelled by AI-related demand for advanced memory solutions including high-bandwidth memory (HBM).[5] In the fourth quarter of fiscal 2025, ending August 29, 2025, revenue climbed to $11.32 billion, up 46% year-over-year from $7.75 billion, with non-GAAP earnings per share (EPS) of $3.03 exceeding analyst expectations of $2.77 and marking a substantial year-over-year EPS expansion of approximately 11 times on a diluted basis.[5] [102] Free cash flow for fiscal 2025 totaled $3.72 billion, representing about 10% of revenue and a sharp turnaround from prior cyclical downturns, enabling sustained capital investments in HBM and advanced DRAM nodes.[5] Key valuation metrics as of late 2025 included a forward price-to-earnings (P/E) ratio of approximately 12.2, reflecting market anticipation of continued earnings growth, while return on equity (ROE) improved to 15.76%, signaling enhanced profitability from operational efficiencies and premium product pricing.[103] [104] Micron's stock price surged over 100% year-to-date through October 2025, outperforming broader indices amid investor enthusiasm for HBM supply commitments fully allocated for the year and projected multi-year demand from AI data centers.[105] In November 2025, Morgan Stanley named Micron a top pick, raising its price target to a street-high $325 from $220 and citing a memory chip market supply shortage similar to 2018, positioning the company to benefit from the AI boom.[106] Investor relations communications in earnings releases and calls emphasized the shift from historical memory commoditization—characterized by price volatility and thin margins—to differentiated tiers like HBM3E, where Micron demonstrated pricing power through sold-out capacity and gross margins expanding to 41% in Q4 fiscal 2025, up 17 percentage points year-over-year.[107] [41] This trajectory underscores causal links between AI infrastructure buildout and Micron's ability to command premiums for high-performance memory, countering narratives of inescapable cyclical traps in the sector.[102]| Metric | Fiscal 2025 Value | Year-over-Year Change |
|---|---|---|
| Revenue | $37.4 billion | +50% |
| Q4 Revenue | $11.32 billion | +46% |
| Q4 Non-GAAP EPS | $3.03 | ~11x increase |
| Free Cash Flow | $3.72 billion | Positive turnaround |
| Forward P/E | ~12.2 | N/A |
| ROE | 15.76% | Improved from negative |
| YTD Stock Return (2025) | >100% | Driven by HBM demand |