iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Short answer: TSMC has the stronger quantified scale record in the available 2025 disclosures, while Samsung Foundry offers a competing 2nm gate-all-around roadmap and manufacturing footprint in Korea and the United States. TSMC says its N2 process entered volume production in the fourth quarter of 2025; Samsung lists SF2 as starting mass production in 2025. Those labels are not directly equivalent physical measurements, and neither company’s public material proves a universal performance or yield winner. The right choice depends on process availability, capacity assurance, geography, design ecosystem, packaging and the qualification needs of the chip.
TSMC and Samsung Foundry at a glance
Both companies are pursuing advanced logic production at the 2nm generation, but they disclose their businesses differently. TSMC publishes a large set of operating figures in its 2025 annual report, including annual wafer capacity, shipments, products and customers. Samsung’s official foundry pages emphasize process generations, application targets, manufacturing sites and design services, but the reviewed material does not provide a current total wafer-capacity number on the same basis.
| # | Preview | Product | Price | |
|---|---|---|---|---|
| 1 |
|
Introduction to Semiconductor Manufacturing Technology, Second Edition | $129.00 | Buy on Amazon |
| 2 |
|
Chip War: The Fight for the World's Most Critical Technology | $15.75 | Buy on Amazon |
| 3 |
|
Semiconductor Devices: Theory and Application | $17.52 | Buy on Amazon |
| 4 |
|
Semiconductor Manufacturing Technology | $146.64 | Buy on Amazon |
| Comparison point | TSMC | Samsung Foundry |
|---|---|---|
| 2nm-class process status | N2 entered volume production in 4Q 2025, according to TSMC; N2P is scheduled for volume production in the second half of 2026. | SF2 is listed as starting mass production in 2025; Samsung describes it as a second-generation MBCFET gate-all-around process. |
| Transistor architecture described in the cited material | First-generation nanosheet transistors for N2. | Second-generation MBCFET/GAA for SF2. |
| 2025 capacity disclosure | More than 17 million 12-inch-equivalent wafers of annual capacity. | No directly comparable current total is stated on the reviewed Samsung pages. |
| 2025 wafer shipments | 15.0 million 12-inch-equivalent wafers. | Not stated on a comparable basis. |
| 2025 customer and product disclosure | 12,682 products manufactured for 534 customers. | No equivalent total customer or product count stated. |
| Manufacturing footprint cited | Taiwan operations plus sites or subsidiaries in Nanjing, Arizona, Japan, Washington and China; additional facilities are under construction or planned. | Giheung, Hwaseong and Pyeongtaek in South Korea, plus Austin and Taylor in the United States. |
| Business model | Pure-play foundry; TSMC says it does not design or market semiconductor products under its own name. | Foundry services are part of Samsung Electronics’ Device Solutions business, alongside Samsung’s broader semiconductor operations. |
This table is a comparison of company disclosures, not an independent ranking. A process name, roadmap date or announced collaboration does not by itself establish yield, shipped volume, power efficiency or customer satisfaction.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Process technology and 2nm timing
TSMC N2 and the planned N2P variant
TSMC states that its N2 process began volume production in the fourth quarter of 2025, as planned. The company describes N2 as using first-generation nanosheet transistors, a gate-all-around approach intended to improve control of the channel as transistor dimensions continue to shrink. TSMC’s 2025 annual report also characterizes 2nm as entering high-volume manufacturing in that quarter and says the company expects a fast ramp during 2026.
TSMC says N2P, a performance-enhanced version of the family, is scheduled to enter volume production in the second half of 2026. That is a company schedule, not evidence that every customer design is already qualified or shipping. Customers evaluating N2 should therefore distinguish the base process’s stated production status from the later N2P availability window.
Samsung SF2 and related roadmap variants
Samsung’s current logic-node page lists SF2 as “Started mass production in 2025.” Samsung identifies SF2 as a second-generation MBCFET, its name for a gate-all-around transistor architecture, and lists mobile, high-performance computing, artificial intelligence and automotive applications for the node. The same page lists SF3 production as starting in 2022 and SF4 in 2021.
An earlier Samsung roadmap announcement specified a staged application plan: mobile SF2 production in 2025, expansion to HPC in 2026 and automotive in 2027. The current node page supports the 2025 start; the later application dates remain roadmap statements and should be treated as targets rather than proof of completed production for those markets.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Samsung’s company information material also describes SF2Z with backside power delivery for HPC and AI, SF2A for automotive, and future development toward a 1.4nm GAA process. Those descriptions identify roadmap variants. They do not establish that SF2Z, SF2A or a 1.4nm product is already in mass production.
Why “2nm” is not a like-for-like measurement
TSMC N2 and Samsung SF2 use the same nominal generation number, but the number is a company process label, not a shared physical standard. The names do not prove identical gate length, metal pitch, density, power characteristics or transistor dimensions. A procurement team should request the foundry’s process design kit, design rules, standard-cell libraries, SRAM options, reliability data and qualification status for the exact variant under consideration.
Production wording also matters. “Volume production,” “high-volume manufacturing” and “mass production” describe the companies’ own status terminology. They do not reveal a common utilization rate, defect density or yield percentage. Independent benchmarking would be needed to make those comparisons, and the cited company material does not provide it.
Capacity, shipments and manufacturing footprint
What TSMC’s numbers show
TSMC’s 2025 annual report states that its annual capacity exceeded 17 million 12-inch-equivalent wafers. The same report records total wafer shipments of 15.0 million 12-inch-equivalent wafers. Capacity is the potential production base for the year; shipments are wafers delivered during the year. They are different measures and should not be treated as utilization or as a market-share calculation.
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Scan for outdated or missing drivers - takes under a minute3Repair Windows errors before they cause bigger problemsTSMC also reports that advanced technologies—defined in the report as 7nm and beyond—generated 74% of total wafer revenue in 2025. This indicates that advanced nodes account for a large portion of the company’s revenue mix, but it does not disclose the yield or profitability of any individual node.
Where TSMC is building and operating
TSMC identifies major operations in Taiwan and sites or subsidiaries in Nanjing, Arizona, Japan, Washington and China. Its 2025 annual report says Arizona Fab 1 began 4nm volume production in the fourth quarter of 2024. The second Arizona facility was installing systems for 3nm and more advanced technologies, while construction of a third facility began in 2025.
The report says the second Japan Advanced Semiconductor Manufacturing (JASM) fab in Kumamoto is under construction, with plans that include 3nm. It also describes a planned Dresden site for specialty 28/22nm and 16/12nm processes. These facilities serve different technology and regional objectives; they should not be assumed to offer the same process menu as Taiwan’s leading-edge fabs.
What is and is not known about Samsung capacity
Samsung’s company information page lists foundry manufacturing in Giheung, Hwaseong and Pyeongtaek in South Korea, along with Austin and Taylor in the United States. Samsung says this network supports mass production of legacy and advanced semiconductors.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Clear out junk files and repair common Windows errorsFree Scan →The reviewed Samsung disclosures do not state a current total wafer capacity comparable to TSMC’s 12-inch-equivalent annual figure. Consequently, it is not valid to divide TSMC’s reported capacity by an unstated Samsung number, infer a precise capacity ratio, or declare which company is larger on wafer capacity from these pages alone. A fair comparison would require the same wafer diameter, period, node mix, facility scope and definition of capacity from both companies.
Customers, products and application markets
TSMC’s quantified customer disclosure
TSMC says it manufactured 12,682 products for 534 customers in 2025. The company identifies high-performance computing, smartphones, internet of things, automotive and digital consumer electronics as end markets. The product and customer counts demonstrate breadth, but they do not reveal how revenue is distributed, whether a few customers account for most leading-edge wafers, or the identity of every customer.
TSMC emphasizes its pure-play model: it manufactures customer products and says it does not design or market semiconductor products under its own name. For a chip designer, that separation can simplify concerns about competing with the foundry’s own branded logic products, although commercial and technical decisions still depend on contract terms, capacity allocation and qualification.
Rank #3
Samsung’s application coverage and ecosystem
Samsung Foundry materials list HPC, AI, mobile and 5G, automotive and IoT as supported applications. Samsung describes a design platform that includes standard-cell libraries, process design kits, intellectual-property blocks, design support, packaging and cloud services. These components matter because a nominally advanced node is usable only when a design team can complete physical implementation, verification, sign-off and manufacturing handoff.
Samsung reported in a 2024 forum announcement that close customer collaborations had produced an 80% increase in Samsung Foundry AI sales over the preceding year. This is Samsung’s own sales-growth figure and comparison period; it is not an independently audited market-share measure.
The 2026 Samsung–Broadcom announcement
On July 25, 2026, Samsung and Broadcom announced a memorandum of understanding to expand collaboration across memory and foundry technologies. The foundry portion focuses on Samsung 2nm-and-below processes for Broadcom products, including wireless broadband communications solutions, and may extend to advanced packaging.
The announcement also cites more than $200 billion of estimated collaboration across memory and foundry through 2030. That is an announced estimate covering both companies and both technology areas, not foundry revenue already realized. An MOU signals planned cooperation; it does not prove that a product has completed qualification, entered volume production or shipped to end customers.
Performance claims, architecture and packaging
How to read Samsung’s SF2 comparison
In a 2023 roadmap announcement, Samsung compared SF2 with its own SF3 process and claimed 12% better performance, 25% better power efficiency and 5% smaller area. Samsung presented those figures as an internal generation-to-generation comparison. They are not a cross-vendor benchmark against TSMC N2, and the announcement does not establish the workloads, library choices, voltage conditions, design size or measurement methodology needed to reproduce them.
Free tools Windows power users keep installed
One-click scans. No signup required.
Why vendor PPA claims cannot settle the debate
Performance, power and area depend on the complete implementation: architecture, memory hierarchy, cell libraries, clock targets, operating voltage, package, software workload and thermal limits. A foundry’s headline PPA claim can be useful for screening a process, but a customer should request workload-specific data and run its own implementation flows before selecting a node.
Packaging as part of the foundry decision
Advanced chips increasingly combine logic, memory and multiple chiplets. TSMC’s annual report discusses continued investment in Taiwan in 2nm and advanced packaging. Samsung’s foundry platform includes packaging, and the Samsung–Broadcom MOU says advanced packaging may be part of the collaboration. These statements show that both companies position packaging as part of the solution, but they do not provide a common test of package bandwidth, thermal resistance, cost or production yield.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Which foundry fits a particular chip?
AI and high-performance computing
For an AI or HPC design, compare the exact 2nm variant, available SRAM and high-density libraries, power-delivery options, reticle and package limits, high-bandwidth-memory integration plans and the foundry’s ability to reserve capacity. Samsung’s SF2 roadmap includes HPC and AI, while its SF2Z description adds backside power delivery for those workloads. TSMC reports N2 production and a planned N2P follow-on. Neither set of statements alone proves superior real-world accelerator performance.
Mobile and consumer processors
Mobile teams should prioritize energy efficiency at the target voltage, leakage across phone temperature ranges, radio and image-processing integration, package thickness and the schedule for the required qualification. Samsung’s stated SF2 plan begins with mobile, and TSMC’s N2 is already described by TSMC as in volume production. The relevant decision is the qualified process and available capacity when the product must launch, not the node number in isolation.
Automotive and industrial IoT
Automotive programs require long qualification cycles, reliability documentation, functional-safety support, temperature ranges and supply continuity. Samsung lists SF2 for automotive and describes SF2A as an automotive variant, with the application timing in its roadmap extending to 2027. TSMC’s footprint includes multiple regions and a planned Dresden site for specialty processes, but a geographic site does not automatically mean that every advanced node or automotive qualification is available there.
Products needing regional manufacturing
Government or customer contracts may require production in a particular country, dual sourcing or a documented business-continuity plan. TSMC’s Arizona, Japan and planned European activities and Samsung’s Korean and U.S. network create different qualification choices. Ask each supplier which fab, process variant and packaging line would actually manufacture the product, rather than evaluating a company solely by its global site list.
A practical foundry-selection checklist
- Define the product requirement. Record performance targets, active and standby power, die size, SRAM needs, package type, operating temperature, reliability class and first-ship date.
- Request the exact process collateral. Obtain the current PDK revision, design rules, standard-cell and memory options, IP catalog, design-service scope and sign-off requirements for the specific node and variant.
- Verify production status. Ask whether the process is in risk production, volume production or high-volume manufacturing for your application, and request qualification milestones rather than relying on a roadmap year.
- Check capacity on a common basis. Compare committed wafer starts, wafer diameter, fab location, node mix, ramp schedule and allocation policy. Do not combine TSMC’s 12-inch-equivalent figures with an unstated Samsung total.
- Model the complete package. Include interposer or bridge technology, substrate, memory integration, thermal design, test, known-good-die handling and package yield.
- Run an implementation comparison. Synthesize representative blocks with each foundry’s libraries, place and route them, close timing and power, and document the conditions behind every result.
- Review commercial and continuity terms. Evaluate mask costs, wafer pricing, engineering support, change-control rules, lead times, export restrictions, second-source options and the consequences of a delayed ramp.
- Confirm customer references appropriately. Separate a named collaboration or MOU from a shipping product, and ask for evidence that the relevant process has been qualified for a comparable workload.
Bottom line for 2026 decisions
TSMC currently provides the clearer public evidence of operating scale: its 2025 disclosures quantify capacity, shipments, products, customers and advanced-technology revenue, and the company says N2 entered volume production in late 2025. Samsung offers a credible competing path with SF2 mass production listed for 2025, a second-generation GAA architecture, Korean and U.S. manufacturing sites, and roadmap variants aimed at HPC, AI and automotive designs.
Those facts do not establish that one foundry wins every design. TSMC’s capacity disclosure is not directly comparable with Samsung’s unstated total, Samsung’s SF2-versus-SF3 PPA figures are not a cross-vendor test, and the Broadcom MOU is an announced collaboration rather than proof of shipments. Select the supplier only after matching the exact process variant, qualified fab, capacity commitment, design ecosystem, package and reliability requirements to the product schedule.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
