2026 Industry Report

Global SSD and DRAM/NAND Supply Status and Outlook

Executive summary

Global memory and SSD supply in April 2026 is best characterized as structurally tight, driven by an AI-led reallocation of manufacturing capacity and constrained cleanroom expansion (especially on advanced nodes and in HBM-adjacent production flows). TrendForce’s latest pricing survey projects another step-change in contract prices in 2Q26—+58–63% QoQ for conventional DRAM and +70–75% QoQ for NAND Flash—explicitly linking the increase to DRAM capacity being redirected toward HBM and server applications and NAND capacity being increasingly allocated to enterprise SSDs, while consumer applications scale back under cost pressure. [1]

The most procurement-relevant segment calls, using the user’s requested labels, are:

• Enterprise SSDs: Shortage. Multiple TrendForce publications describe enterprise SSD demand as AI/data-center led, inventories low, and 2026 supply deficit likely, with “meaningful capacity expansion unlikely until late 2027 or 2028.” [2]

• Consumer/client SSDs (PC + retail SSD ecosystem): Tight supply (selective shortages). Suppliers are limiting shipments to client SSDs as capacity is pulled into higher-margin enterprise products; TrendForce forecast ≥40% QoQ client SSD contract increases for 1Q26 amid “capacity crowding out.” [3]

• Server DRAM (DDR5 RDIMM/LRDIMM and high-density modules): Shortage / very tight. TrendForce describes ongoing CSP expansion and LTAs, with “near-term supply remains tight,” and earlier projected server DRAM pricing rising >60% QoQ in 1Q26. [4]

• Desktop/laptop DRAM modules (PC DRAM): Tight to shortage. TrendForce describes reduced shipments to PC OEMs/module makers and “PC DRAM prices…expected to at least double” in 1Q26 (followed by further increases in 2Q26 under tight supply). [5]

• Mobile DRAM (LPDDR4X/LPDDR5X): Tight supply. TrendForce expects continued contract price increases in 2Q26; it does not expect mobile DRAM demand to contract significantly in 1H26, even as smartphone brands face rising memory cost pressure. [4]

Over the last 12 months (2Q25→2Q26), memory pricing moved from mild stabilization/gradual increases in mid‑2025 to an unusually steep escalation in late‑2025 and early‑2026. For example, TrendForce revised 1Q26 expectations to +90–95% QoQ for conventional DRAM and +55–60% QoQ for NAND Flash, and then projected another large rise in 2Q26. [5]

Inventory conditions are mixed by channel: company balance sheets (e.g., Micron inventories essentially flat at $8.27B in Feb 2026 vs. Nov 2025) can coexist with tight downstream availability where allocation, LTAs, and segment prioritization constrain accessible supply. [6]

Strategic takeaways for buyers and suppliers:

Buyers should assume continued tightness (not a quick snap-back) through the next 6–12 months, prioritize LTAs and allocation management, and avoid being overexposed to a single DRAM generation or SSD NAND type. This is consistent with IDC’s framing that the memory ecosystem is experiencing a shortage that could persist “well into 2027,” driven by AI-induced capacity reallocation to high-margin memory solutions. [7]

Suppliers will likely maintain pricing power in enterprise/server segments while using process upgrades (higher-layer NAND, DRAM node migrations) rather than large new cleanroom builds as the primary near-term “supply lever,” consistent with TrendForce’s view that 2026 capex increases are modest and have limited effect on 2026 bit supply growth. [8]

Current supply status by segment

Segment comparison table

The table below reflects global conditions as of April 2026, with emphasis on TrendForce/DRAMeXchange and primary company commentary. Where an exact numeric inventory measure is not public, inventory is described qualitatively and the assumption is stated.

Segment	Current status	Price trend (last 12 months)	Inventory & availability	Main causes (demand + supply)
Consumer SSDs (client/OEM + retail ecosystem)	Tight supply (select shortages by capacity/NAND type)[9]	Up sharply from mid‑2025;  TrendForce projected ≥40% QoQ client SSD contract  increase in 1Q26 and continued upward pressure in 2Q26 [9]	OEMs restocking at times, but suppliers limiting client SSD shipments to sustain pricing; end-demand pressured by higher BOM costs [10]	Capacity diverted toward enterprise SSD; cautious NAND expansion; notebook demand weakening into 2026 reduces unit pull but not enough to offset allocation squeeze[11]
Enterprise SSDs	Shortage  [12]	Up strongly: enterprise SSD market described shifting to undersupply (2025) and entering record-high contract pricing conditions into 2026 [13]	Finished-product inventories described as low; CSPs increasingly sign LTAs to secure supply [14]	AI inference + CSP buildouts; HDD shortages spilling into SSD demand; constrained NAND capacity and prioritization of high-margin products[15]
Desktop/laptop DRAM modules (PC DRAM)	Tight to shortage [5]	Up sharply: TrendForce expected PC DRAM pricing to “at least double” in 1Q26; 2Q26 still supported by tight supply [5]	OEMs with weaker allocation fulfillment procure at higher prices; downstream availability constrained by suppliers reducing shipments to PCs/module makers [4]	DRAM capacity reallocated toward server/HBM; limited cleanroom expansion; “catch-up pricing” across segments  [16]
Server DRAM (DDR5  RDIMM etc.)	Shortage / very tight [4]	Up strongly: server DRAM projected to climb  >60% QoQ in 1Q26; pricing remains supported in 2Q26 under tight supply [9]	Supply “remains tight”; high-capacity RDIMMs are a primary target; LTAs used to secure future capacity [4]	AI inference deployments + general-purpose server growth; suppliers prioritize server DRAM profitability and shift capacity accordingly  [4]
Mobile DRAM (LPDDR)	Tight supply [4]	Up: mobile DRAM contract prices expected to keep rising in 2Q26 even as brands face cost pressure [4]	Demand not expected to contract significantly in 1H26; but brands may adjust production plans starting 2Q26 under cost pressure [1]	AI features increase memory/storage content in flagships; supply constrained by allocation toward higher-margin lines and process transitions away from low-density parts [17]

What is “short” vs “tight” in this cycle

This market is unusual because “tightness” is not purely a cyclical demand spike; it is reinforced by (a) allocation discipline (suppliers preferring higher-margin segments) and (b) slow-to-add cleanroom capacity, which TrendForce explicitly calls out as limited in DRAM, with only modest line expansion possible for Samsung/SK hynix and major new U.S. DRAM capacity (Micron ID1) not expected to come online before 2027. [8]

IDC similarly describes the shortage as being driven by AI-centered capacity diversion toward HBM and high-capacity DDR5, restricting general-purpose memory module supply and pushing prices up broadly. [18]

Price and inventory dynamics over the last 12 months

Contract price trajectory

The most consistent public, segment-level pricing signals in the last 12 months come from TrendForce’s quarterly contract-price outlooks and updates:

• Conventional DRAM (overall): eased in 2Q25 (down 0–5% QoQ expected), then rose in 3Q25 (+10–15% projected), continued in 4Q25 (+8–13% projected), then surged in 1Q26 (+90–95% projected), and remains on a steep uptrend in 2Q26 (+58–63% projected). [19]

• NAND Flash (overall): stabilized in 2Q25 (while wafer and client SSD pricing strengthened), then projected to rise 5–10% in 3Q25 and 4Q25, followed by a step-change in 1Q26 (+55–60%) and 2Q26 (+70–75%). [20]

A normalized price index chart

The chart below converts those quarterly % ranges into a normalized index (2Q25 = 100). Assumption: each quarter uses the midpoint of TrendForce’s quoted QoQ range; for 2Q25 NAND, the index assumes “stabilize” implies ~0% net change in overall NAND contract pricing (TrendForce language), even though certain subsegments (e.g., wafers) rose. [21]

xychart-beta
    title “Memory Contract Price Trend (Index, 2Q25 = 100; midpoint ranges)”
    x-axis [“2Q25″,”3Q25″,”4Q25″,”1Q26″,”2Q26”]
    y-axis “Index” 0 –> 420
    line “Conventional DRAM” [100,112.5,124.3,239.3,384.1]
    line “NAND Flash” [100,107.5,115.6,182.0,314.0]

Interpretation: even with conservative midpoint handling, the implied index suggests a roughly 3.8× increase in conventional DRAM contract pricing and ~3.1× for NAND contract pricing from 2Q25 to 2Q26, consistent with the narrative that 2026 contract pricing is being “caught up” across segments under tight supply. [22]

Spot and street pricing signals

Spot and channel pricing can diverge from contract pricing in allocation-driven environments. TrendForce’s publicly visible price pages show that as of early April 2026:

• DRAM spot prices and module spot pricing remain elevated, but can swing week-to-week, reflecting channel inventory clearing and negotiations (TrendForce provides daily/weekly spot tables). [23]

• NAND Flash spot and wafer spot prices similarly show strong pricing on certain categories (notably mature-node items like SLC/MLC), while some wafer spot lines can show short-term declines even within a broader uptrend. [24]

A key practical point for buyers: contract markets are being driven by allocation and LTAs, while spot markets can temporarily soften if a region clears inventory—without meaningfully easing OEM availability when suppliers limit shipments. TrendForce explicitly notes contracting retail/module demand in some NAND segments under price pressure even as overall tightness persists. [25]

Inventory levels

Public “inventory” data exists at three different layers, and they often point in different directions:

Company balance sheets:

• Micron reported inventories of $8.267B as of Feb 26, 2026 (roughly flat vs Nov 27, 2025), even as management described strong demand and tight industry supply conditions. [26]

• Samsung reported inventory of KRW 52.6T at Dec 31, 2025 in its 4Q25 earnings presentation, while its memory slide emphasizes limited supply availability and a focus on high-value products (HBM, server DDR5, enterprise SSD). [27]

Downstream/OEM inventory and allocation:

• TrendForce repeatedly describes declining OEM inventories and allocation fulfillment issues (e.g., PC OEMs forced to procure at higher prices due to reduced shipments). [22]

IDC’s device-market view:

• IDC describes channels “building inventory” ahead of further price increases in late 2025, but frames the overall system as being supply constrained by AI-driven reallocation and limited cleanroom flexibility. [18]

Assumption (explicit): precise “weeks of inventory” by segment (e.g., PC DRAM weeks at OEMs) is generally not public without paid datasets; this report therefore treats inventory as qualitative unless a primary balance-sheet number is available.

Supply-side: manufacturing capacity, utilization, and vendor strategies

Industry capacity expansion is limited in the near term

A pivotal constraint is that 2026 capex increases do not automatically translate into 2026 bit supply relief because cleanroom capacity and ramp timing are limiting. TrendForce (via DRAMeXchange) quantifies this directly:

• 2026 capex is projected to rise modestly and be driven more by process upgrades, higher-layer stacking, hybrid bonding, and HBM than by large net new capacity, limiting 2026 bit supply impact. [8]

• TrendForce projects 2026 DRAM capex at $61.3B and NAND capex at $22.2B, with specific vendor estimates: Micron $13.5B, SK hynix $20.5B, Samsung $20B, and Kioxia/SanDisk $4.5B (with Kioxia/SanDisk NAND investment +41% YoY). [8]

• TrendForce highlights limited DRAM cleanroom capacity, with Micron’s new U.S. ID1 DRAM fab not expected operational before 2027, limiting near-term supply elasticity. [8]

Vendor-level production strategy signals

Samsung:

• Samsung’s 4Q25 earnings materials explicitly state it addressed strong conventional DRAM demand while expanding HBM sales “despite limited supply availability,” and highlights profitability improvement tied to higher sales of HBM, server DDR5, and enterprise SSD. [28]

• Samsung’s balance-sheet appendix shows inventories at KRW 52.6T at Dec 31, 2025. [29]

SK hynix:

• SK hynix states it has “secured full customer demand” for its entire DRAM and NAND production for next year and plans capacity expansion via M15X and advanced process migration (including 1c nm DRAM and 321-layer NAND). [30]

• SK hynix further emphasizes scaling production capacity (M15X, Yongin Semiconductor Cluster plans, and advanced packaging facilities in Korea and Indiana) to address supply-demand imbalance. [31]

Micron:

• Micron’s Dec 2025 presentation states its 1‑gamma DRAM node will be the primary driver of DRAM bit growth in calendar 2026, and that both DRAM and NAND bit shipment growth in 2026 is constrained by industry supply; Micron expects tightness to persist “through and beyond calendar 2026.” [32]

• Micron also states it plans to raise fiscal 2026 capex to approximately $20B (vs $18B prior), pulling in equipment orders and accelerating installation timelines; it targets first wafer output from its first Idaho fab in mid‑2027. [33]

• Micron’s fiscal Q2 2026 press release describes record results driven by “strong demand environment” and “tight industry supply” and reports capex net of $5.0B in the quarter; inventories were $8.267B at Feb 26, 2026. [34]

Kioxia / SanDisk (Western Digital flash successor):

• Western Digital completed its planned company separation in Feb 2025, separating its Flash business (now SanDisk) from Western Digital’s HDD-focused entity. [35]

• Kioxia and SanDisk extended their Yokkaichi joint venture agreements through 2034 (Jan 2026), reinforcing long-term commitment to co-development and manufacturing scale at a major NAND production base. [36]

• TrendForce projects Kioxia/SanDisk NAND capex to increase substantially in 2026 (investment +41% YoY to $4.5B) focused on BiCS8 ramp and BiCS9 R&D, in contrast to Samsung and SK hynix/Solidigm limiting NAND investment while shifting toward HBM/DRAM. [8]

Intel/others (market implication):

• Intel is no longer a primary NAND manufacturer; the practical “other” in NAND/enterprise SSD supply is Solidigm (SK hynix group), which SK hynix highlights as part of its AI data center storage approach (QLC eSSD). [37]

Capacity utilization chart

Most memory makers do not publish fab utilization rates as a standard KPI. The chart below therefore uses a proxy: “effective utilization/booking pressure,” derived from (a) explicit “sold out/fully secured demand” language and (b) repeated statements that supply is tight and demand cannot be met despite efforts. This should be read as directional, not audited.

xychart-beta
    title “Inferred Effective Utilization / Booking Pressure (April 2026; proxy)”
    x-axis [“Samsung”,”SK hynix”,”Micron”,”Kioxia/SanDisk”]
    y-axis “Proxy %” 0 –> 100
    bar “Estimated range midpoint” [95,98,94,97]

Evidence basis: SK hynix’s explicit “secured full customer demand” for next-year DRAM and NAND indicates extremely high booking pressure; TrendForce describes limited DRAM cleanroom expansion and tight supply, and Micron states tightness persists beyond 2026 even as it increases capex and cannot meet customer demand. [38]

Demand drivers shaping 2025–2026 balance

Data centers, AI servers, and accelerators

AI infrastructure is the dominant driver on both DRAM and NAND/SSD:

• TrendForce attributes 2Q26 price increases to CSPs accelerating AI inference deployments and signing LTAs; high-capacity RDIMMs and enterprise SSD demand are central. [1]

• Micron describes an “extraordinary, multiyear data center buildout” and highlights rising memory and storage content requirements generation-to-generation, with capacity storage SSDs (including very large QLC drives) entering hyperscaler qualification. [39]

• IDC frames the memory shortage as driven by AI data centers consuming disproportionate memory per system and by manufacturers reallocating capacity toward HBM and high-capacity DDR5, restricting general-purpose DRAM/NAND supply. [18]

PCs (consumer and commercial)

PC demand is not the primary driver of memory tightness in 2026, but it shapes the client SSD and PC DRAM allocation battle:

• Gartner reported a PC rebound in 2025 (+9.1% YoY; >270M units), supported by consumer demand and Windows 11 upgrades. [40]

• Entering 2026, multiple analysts (IDC and TrendForce) point to memory supply constraints and rising component costs as reasons for weaker PC shipment expectations. IDC’s PC forecast page states global PC shipments are expected to decline 11.3% in 2026, explicitly tied to memory shortage and supply chain disruption. [41]

• TrendForce’s notebook-specific view is even more bearish: 2026 notebook shipments revised down to ‑14.8% YoY, citing weak demand and rising supply chain costs, and explicitly calling out “tight memory supply” pushing up system costs. [42]

Smartphones and mobile memory content

Even with uncertain unit growth, memory content per device is rising at the high end:

• TrendForce forecasts average smartphone storage capacity rising 4.8% YoY in 2026, driven by AI upgrades and process transitions reducing supply of low-capacity NAND products. It states edge AI models can require 40–60 GB of system storage as cache for local AI processing, pushing base storage configs upward (e.g., iPhone base storage increases). [43]

• IDC’s smartphone forecast highlights downside risk from the memory shortage: its smartphone market insights page projects ‑12.9% YoY smartphone shipments to ~1.1B units in 2026, explicitly citing an “intensifying memory shortage crisis.” [44]

Crypto mining

Crypto mining is not identified by the primary sources above as a major incremental driver for DRAM or NAND in 2025–2026 relative to AI data center buildouts. In this cycle, the dominant “accelerator-linked” memory driver is AI/HBM and associated server memory/storage content rather than crypto.

Outlook and actionable implications for next 6–12 months

Base-case forecast: tight supply persists through late 2026 and into early 2027

A defensible base case—consistent across TrendForce, IDC, and leading manufacturer commentary—is:

• Enterprise SSD and server DRAM remain the tightest segments because AI inference and data-center expansion continue and because suppliers prioritize these segments (profitability + LTAs). TrendForce expects a “clear shortage” in 2026 and that meaningful NAND capacity expansion is unlikely until late 2027 or 2028. [1]

• Client SSD and PC DRAM stay tight, even if end-device shipments soften, because allocation is constrained and suppliers limit shipments to lower-margin segments; TrendForce explicitly describes reduced shipments to PC OEMs/module makers supporting price increases in 2Q26. [4]

• Mobile DRAM remains tight but may show the earliest signs of demand-side moderation if smartphone production plans are adjusted under cost pressure; TrendForce still expects continued contract price increases in 2Q26. [45]

• Large, structural supply relief is unlikely before 2027 because new cleanroom and fab timelines are long: TrendForce expects Micron’s major new DRAM fab not operational before 2027, while Micron itself targets first Idaho output mid‑2027. [46]

Key risks and swing factors

Demand-side downside risks (could ease tightness, especially consumer-facing segments):

• PC and smartphone unit contraction from “memory-cost shock.” IDC explicitly provides downside risk scenarios in which higher memory costs drive weaker smartphone and PC volumes. [47]

• Price elasticity and specification downgrades (e.g., reduced SSD capacities, lower DRAM configs) that reduce bit demand, partially offsetting supply tightness; TrendForce notes PC and smartphone vendors reducing capacities to curb NAND demand in 2Q26. [1]

Supply-side upside risks (could worsen tightness):

• HBM and server prioritization accelerates further, crowding out conventional DRAM/NAND for consumer devices. TrendForce repeatedly emphasizes ongoing reallocation toward HBM and server applications. [16]

• Node transition and yield risk: DRAM 1‑gamma ramps and NAND high-layer/QLC migrations improve bit output but can be constrained by yield learning curves. Micron and SK hynix both emphasize node migrations (1‑gamma; 1c nm; 321-layer NAND) as key supply levers. [48]

• Geopolitics and trade policy: export controls and tariff changes can disrupt equipment flows or alter demand timing; TrendForce explicitly noted tariffs causing front-loaded procurement in earlier quarters and widespread defensive inventory behavior. [49]

Actionable conclusions

For enterprise buyers (cloud, data center, OEM/ODM):

• Treat enterprise SSD and server DRAM as the highest-risk categories for availability and price through at least the next 2–3 quarters; prioritize LTAs, qualify second sources, and lock high-density RDIMM and key SSD form factors early (TrendForce explicitly describes LTAs and high-capacity RDIMMs as the procurement focus). [1]

• If cost control is critical, evaluate architecture options that reduce dependency on the scarcest parts (e.g., mix of SSD tiers, or density optimization), but assume that “waiting for prices to normalize” is unlikely to be a viable strategy within 6–12 months given projected 2Q26 increases and limited near-term expansion. [16]

For consumer OEMs (PC/smartphone makers) and channel buyers:

• Expect continued allocation pressure for client SSD and PC DRAM; reduce exposure by standardizing a narrower set of SKUs/configurations (better leverage in negotiation) and by planning for configuration flexibility (ability to swap SSD capacity tiers or DRAM densities), consistent with TrendForce and IDC warnings about higher system costs and shipment uncertainty. [50]

• Use spot-market dips cautiously: they may reflect regional inventory clearing rather than structural supply relief when suppliers are still limiting shipments into low-margin segments. [51]

For suppliers (memory and SSD ecosystem):

• Profit-maximizing allocation (HBM/server DRAM/enterprise SSD) will likely remain rational in 2026 given capex constraints and buyer willingness to sign LTAs; TrendForce explicitly frames this as the prevailing strategy and ties it to tight supply through 2026. [16]

• The biggest operational “unlock” for marginal supply is likely process migration + yield improvement rather than greenfield volume expansion in the next 12 months, consistent with TrendForce capex analysis and Micron’s stated approach of maximizing output from current footprint while ramping leading nodes. [46]

Market structure: why this cycle is behaving differently

IDC characterizes the current memory shortage as driven by a potentially semi-permanent “strategic reallocation” of wafer capacity toward AI-centric products (HBM, high-capacity DDR5) rather than a typical boom-bust inventory cycle. TrendForce’s 2026 pricing survey strengthens that conclusion by projecting continued sharp increases even as some consumer end-markets weaken, because supply remains constrained and capacity is being redirected. [7]

[1] [2] [4] [10] [11] [16] [17] [22] [25] [45] [51] https://www.trendforce.com/presscenter/news/20260331-12995.html

https://www.trendforce.com/presscenter/news/20260331-12995.html

[3] [9] https://www.trendforce.com/presscenter/news/20260105-12860.html

https://www.trendforce.com/presscenter/news/20260105-12860.html

[5] https://www.trendforce.com/presscenter/news/20260202-12911.html

https://www.trendforce.com/presscenter/news/20260202-12911.html

[6] [26] [34] https://investors.micron.com/news-releases/news-release-details/micron-technology-inc-reports-results-second-quarter-fiscal-2026

https://investors.micron.com/news-releases/news-release-details/micron-technology-inc-reports-results-second-quarter-fiscal-2026

[7] [18] [47] https://www.idc.com/resource-center/blog/global-memory-shortage-crisis-market-analysis-and-the-potential-impact-on-the-smartphone-and-pc-markets-in-2026/

https://www.idc.com/resource-center/blog/global-memory-shortage-crisis-market-analysis-and-the-potential-impact-on-the-smartphone-and-pc-markets-in-2026/

[8] [46] https://www.dramexchange.com/WeeklyResearch/Post/2/12522.html

https://www.dramexchange.com/WeeklyResearch/Post/2/12522.html

[12] [13] [14] https://www.trendforce.com/presscenter/news/20250526-12591.html

https://www.trendforce.com/presscenter/news/20250526-12591.html

[15] https://www.trendforce.com/presscenter/news/20250925-12736.html

https://www.trendforce.com/presscenter/news/20250925-12736.html

[19] https://www.trendforce.com/presscenter/news/20250325-12528.html

https://www.trendforce.com/presscenter/news/20250325-12528.html

[20] [21] https://www.trendforce.com/presscenter/news/20250326-12530.html

https://www.trendforce.com/presscenter/news/20250326-12530.html

[23] https://www.trendforce.com/price/dram/dram_spot

https://www.trendforce.com/price/dram/dram_spot

[24] https://www.trendforce.com/price/flash/flash_spot

https://www.trendforce.com/price/flash/flash_spot

[27] [28] [29] https://images.samsung.com/is/content/samsung/assets/global/ir/docs/2025_4Q_conference_eng.pdf

https://images.samsung.com/is/content/samsung/assets/global/ir/docs/2025_4Q_conference_eng.pdf

[30] [38] https://news.skhynix.com/sk-hynix-announces-3q25-financial-results/

https://news.skhynix.com/sk-hynix-announces-3q25-financial-results/

[31] [37] https://news.skhynix.com/sk-hynix-announces-fy25-financial-results/

https://news.skhynix.com/sk-hynix-announces-fy25-financial-results/

[32] [33] [39] [48] https://investors.micron.com/static-files/530bd7ed-a8c8-4687-af4a-8c129f740e09

https://investors.micron.com/static-files/530bd7ed-a8c8-4687-af4a-8c129f740e09

[35] https://www.westerndigital.com/company/newsroom/press-releases/2025/2025-02-24-western-digital-completes-planned-company-separation

https://www.westerndigital.com/company/newsroom/press-releases/2025/2025-02-24-western-digital-completes-planned-company-separation

[36] https://www.sandisk.com/company/newsroom/press-releases/2026/2026-01-29-kioxia-and-sandisk-extend-yokkaichi-joint-venture-agreement-through-2034

https://www.sandisk.com/company/newsroom/press-releases/2026/2026-01-29-kioxia-and-sandisk-extend-yokkaichi-joint-venture-agreement-through-2034

[40] https://www.gartner.com/en/newsroom/press-releases/2026-1-20-gartner-says-worldwide-pc-shipments-increased-9-point-3-percent-in-fourth-quarter-of-2025-and-9-point-1-percent-for-the-full-year

https://www.gartner.com/en/newsroom/press-releases/2026-1-20-gartner-says-worldwide-pc-shipments-increased-9-point-3-percent-in-fourth-quarter-of-2025-and-9-point-1-percent-for-the-full-year

[41] https://www.idc.com/promo/pcdforecast/

https://www.idc.com/promo/pcdforecast/

[42] [50] https://www.trendforce.com/presscenter/news/20260330-12992.html

https://www.trendforce.com/presscenter/news/20260330-12992.html

[43] https://www.trendforce.com/presscenter/news/20260323-12983.html

https://www.trendforce.com/presscenter/news/20260323-12983.html

[44] https://www.idc.com/promo/smartphone-market-share/market-share/

https://www.idc.com/promo/smartphone-market-share/market-share/

[49] https://www.trendforce.com/presscenter/news/20250417-12551.html

https://www.trendforce.com/presscenter/news/20250417-12551.html