TSMC logo

TSMC

PublicFounded 1987🇹🇼Hsinchu, Taiwan
CEO

C.C. Wei (Chairman & CEO)

Data verified as of Sep 1, 2026

Summary

Taiwan Semiconductor Manufacturing Company, founded in Hsinchu in 1987 by Morris Chang, invented the dedicated-foundry business model and now dominates it utterly: 69.9% of the global foundry market in 2025, holding at roughly 73% through the first two quarters of 2026 (Counterpoint Research), with Samsung a distant second at 7%. It is effectively the sole manufacturer of leading-edge AI silicon - Nvidia, Apple, AMD, and Broadcom all fab their flagship chips there - which made it the primary beneficiary of the AI buildout: 2025 revenue hit a record $122.4B (+35.9%), and Q1 2026 revenue grew another 35.1% to $35.9B at a 66.2% gross margin, with high-performance computing now 61% of sales. Its N2 (2nm) node, the company's first gate-all-around process, entered volume production on schedule in Q4 2025 at Fab 22 in Kaohsiung, with the enhanced N2P variant due in H2 2026 and the A16 node (backside power delivery) now slated for 2027 after slipping from late 2026. TSMC raised its 2026 capex guidance to a record $60–64B in July 2026 - the largest capital budget in semiconductor history - while globalizing its footprint: Arizona's Fab 21 has produced 4nm chips at Taiwan-par yields since Q4 2024, and on its July 16 earnings call TSMC pledged roughly $100B more to expand the Arizona campus with at least four additional 2nm-and-below fabs, lifting its total US commitment to about $265B (the largest foreign direct investment in US history); Japan's JASM is building a second Kumamoto fab upgraded to 3nm, now targeted for 2028 after the upgrade pushed back its original late-2027 goal, and the Dresden ESMC fab targets production in late 2027. TSMC is also diversifying beyond logic: on August 10, 2026, it announced a ~$6.3B joint venture with Sony (Sony 60%/TSMC 40%) to build a Kumamoto image-sensor fab targeting 2029 mass production. A January 2026 US–Taiwan trade deal exempts TSMC from new 25% Section 232 semiconductor tariffs in proportion to its US buildout. The overhang remains geopolitical: the vast majority of its leading-edge capacity sits on Taiwan, at the center of US–China tension.

Company OverviewRead the long-form profile of TSMC

/ Recent News

View all →
  • View all news

    Browse the daily archive

Main Products

N2 Process Node

TSMC's 2nm-class node and its first with gate-all-around nanosheet transistors, delivering 10–15% more speed or 25–30% less power than N3E. Volume production began on schedule in Q4 2025 at Fab 22 (Kaohsiung) and Fab 20 (Hsinchu), with the enhanced N2P variant following in H2 2026 - the process powering the next generation of flagship smartphone and AI silicon.

In volume production since Q4 2025 with reported yields around 70%; ramping through 2026 on smartphone and HPC demand, N2P due H2 2026

Volume production startQ4 2025 (on schedule)
Performance vs N3E10–15% speed gain or 25–30% power reduction
Reported yield~70% (early 2026)

TSMC's 2.5D and 3D packaging technologies that integrate GPU dies with stacks of high-bandwidth memory on silicon interposers - the technology every leading AI accelerator depends on, and for two years the binding constraint on global AI chip supply. Capacity is being expanded roughly fourfold from late-2024 levels, including the new AP7 plant in Chiayi and two planned Arizona packaging facilities.

Aggressive expansion toward ~130-140k CoWoS wafers/month by end of 2026, up from ~35k in late 2024 (10–20% of the 2026 capital budget); as of August 2026, TSMC is also outsourcing the CoW front-end step to OSAT partners (ASE and others) to ease capacity pressure driven largely by Nvidia's reservation of an estimated 800,000-850,000 CoWoS wafers for 2026

CoWoS capacity target~130-140k wafers/month by late 2026 (~4x late-2024)
Share of 2026 capex10–20% of $60–64B
5.5-reticle CoWoS-L yield98-99% in volume production across multiple AI customer products (Aug 2026); memory and ABF substrates flagged as the next AI-supply bottleneck, not CoWoS itself
2027 allocation shiftReported (Aug 26, 2026) to reallocate share toward AMD (agentic-AI processor demand, MI355/MI400) and trim Nvidia's share as TSMC's expanding CoWoS capacity outgrows any single customer's 2026 lock-in; Nvidia still holds an estimated ~60% of 2026 CoWoS capacity (~595k wafers)
Arizona Gigafab Cluster (Fab 21)

TSMC's US manufacturing site in Phoenix, now backed by a ~$265B commitment - the largest foreign direct investment in US history - after TSMC pledged another ~$100B on its July 16, 2026 earnings call to add at least four more 2nm-and-below fabs. Phase 1 has produced 4nm chips at Taiwan-par yields since Q4 2024 for customers including Apple and Nvidia; Phase 2 brings 3nm production, with equipment move-in from Q3 2026, pilot production in early 2027, and volume production targeted for Q3 2027 (pulled in about a year from the original plan); Phase 3, targeting N2 and A16, has broken ground for ~2029. TSMC expects ~30% of its 2nm-and-below capacity to eventually sit in the US.

Phase 1 in N4 volume production since Q4 2024; Phase 2 (N3) equipment move-in from Q3 2026, pilot production early 2027, volume production Q3 2027; Phase 3 (N2/A16) targeting ~2029; four more fabs pledged Jul 2026

Total US commitment~$265B (raised another ~$100B on Jul 16, 2026)
Phase 1 productionN4 (4nm) since Q4 2024, yields on par with Taiwan
Phase 2 production targetN3 pilot production early 2027, volume production Q3 2027 (pulled in ~1 year)

A16 Process Node

In Development

TSMC's 1.6nm-class node combining gate-all-around nanosheets with Super Power Rail backside power delivery, targeting AI and HPC chips with 8–10% more speed or 15–20% less power than N2P. Volume production is now slated for 2027 after slipping from the original H2 2026 target; the follow-on A14, A12, and A13 nodes extend the roadmap to 2029 - all planned without High-NA EUV.

Development and verification completed August 2026 with initial Q4 2026 production on track; broad-based volume production remains 2027 (slipped from H2 2026 at the April 2026 Technology Symposium)

Performance vs N2P8–10% speed gain or 15–20% power reduction
Anchor customerReported to be Nvidia, for its next-generation 'Feynman' GPU (chiplets via SoIC, backside power delivery); Apple is expected to skip A16 and move directly to the follow-on A14 node - a break from Apple's decade-long role as the anchor tenant for each new TSMC node
Volume production target2027

What's Next

N2P volume production

The enhanced 2nm node enters volume production, offering a performance/power step over N2 for the 2027 flagship chip generation while N2 itself ramps through Taiwan's Fab 22 and Fab 20. On track as of early August 2026, with 2nm capacity headed toward 100,000 wafers/month by year end.

H2 2026

Arizona Fab 21 Phase 2 begins 3nm production

The most advanced node yet manufactured in the US: construction finished months ahead of schedule, equipment move-in starts Q3 2026, pilot 3nm production is targeted for early 2027, and volume N3 production for Q3 2027 - pulled in roughly a year from the original plan.

Q3 2027

A16 volume production

TSMC's first node with Super Power Rail backside power delivery enters volume production, now targeted for 2027 after the April 2026 roadmap update moved it from H2 2026. TSMC confirmed in August 2026 that development and verification are complete and that initial Q4 2026 production is on track, but broad-based volume production tied to customer qualification remains a 2027 event. Watch whether AI/HPC lead customers (the node's explicit target market) land on schedule.

2027

Dresden ESMC fab starts production

TSMC's €10B+ European joint venture with Bosch, Infineon, and NXP (TSMC 70%) finished structural construction in late 2025; equipment move-in runs through H2 2026 with production of 28/22nm and 16/12nm automotive-class chips targeted for late 2027 at 40k wafers/month.

Late 2027

2027 wafer price increases take effect

TSMC's newly finalized base wafer-price increases of 5-10% across advanced nodes (plus a 10-15% surcharge on incremental HPC orders) and its first mature-node price hike in three-plus years take effect, testing how much pricing power TSMC's near-monopoly on leading-edge capacity gives it against Apple, Nvidia, AMD, and Qualcomm.

January 2027

JASM Kumamoto Fab 2 operations - upgraded to 3nm

Japan's second TSMC fab, under construction since October 2025, was upgraded in February 2026 from the originally planned 6/7nm to 3nm, confirmed by CEO C.C. Wei and approved by Taiwan's government in March 2026 - Japan's first leading-edge node. The upgrade raised the investment from ~$13.9B to ~$17B and set planned capacity at 15,000 12-inch wafers/month; equipment installation and mass production are now targeted for 2028, later than the pre-upgrade late-2027 goal.

2028

Sony-TSMC image sensor joint venture begins production

Sony Semiconductor Solutions and TSMC signed a definitive agreement on August 11, 2026 establishing Advanced Vision Semiconductor Manufacturing Corporation in Koshi City, Kumamoto - Sony contributing about JPY465B and TSMC about JPY282B in phases, consistent with the 60%/40% split first announced August 10. The JV will co-develop and manufacture next-generation image sensors, initially targeting high-performance camera sensors for smartphones (including Apple's iPhone) with longer-term ambitions in robotics and vehicle AI object recognition. Mass production is targeted for 2029 - TSMC's first move into image sensors.

2029

Q3 2026 earnings report

TSMC's next quarterly report, expected around mid-October 2026 per its investor-relations financial calendar, following guidance (issued Jul 16, 2026) for revenue of $44.6-45.8B (+37% YoY at midpoint, +12% QoQ), gross margin of 65-67%, and operating margin of 56-58%, as the N2 ramp and overseas-fab dilution weigh on margins. As of the August 2026 monthly-revenue cadence the company was tracking ahead of that guidance. Watch whether 2nm capacity reaches the ~100,000 wafers/month year-end target and whether outsourcing CoWoS packaging steps to OSAT partners eases the AI-accelerator packaging bottleneck.

Mid-October 2026

Track Record

1 of 3 announced dates hit · 2 slipped

A16 volume production in H2 2026

Target H2 2026 · Resolved Apr 23, 2026

Announced in April 2024 for production in 2026, A16 was pushed to 2027 in the roadmap TSMC presented at its April 2026 North America Technology Symposium - the company's most visible recent node slip.

Slipped

N2 (2nm) volume production in H2 2025

Target H2 2025 · Resolved Dec 30, 2025

N2 entered volume production in Q4 2025 exactly as promised, confirmed December 30, 2025 - TSMC's first gate-all-around node, ramping at Fab 22 Kaohsiung and Fab 20 Hsinchu with reported yields around 70%.

Hit

Start construction of Kumamoto Fab 2 by end of 2024

Target End of 2024 · Resolved Oct 28, 2025

Construction was promised for end-2024 but began only in late October 2025, roughly ten months late, after local traffic and infrastructure concerns; the end-2027 operations target was retained and the planned process later upgraded to 3nm.

Slipped

Operations & Revenue

StatusOperational - record growth on AI demand

Record run continues: Q2 2026 net profit jumped 77% YoY to a record NT$706.6B (~US$22B) on record revenue of NT$1.27T (+36% YoY), TSMC's fifth straight record quarter, with advanced CoWoS packaging reported sold out through year-end. On the July 16, 2026 call, management raised full-year 2026 revenue-growth guidance to more than 40% in USD terms (from more than 30% in April) and lifted 2026 capex guidance to $60–64B (from $52–56B), while guiding Q3 2026 gross margin down to 65-67% (from Q2's record 67.7%) as the N2 ramp and overseas-fab dilution bite. July 2026 revenue of NT$467.58B (+44.7% YoY) already beat that pace. TSMC is effectively the sole manufacturer of leading-edge AI silicon - Nvidia, Apple, AMD, Broadcom, and now Intel as a foundry customer - holding a record 73% of the pure-play foundry market in Q2 2026 (Samsung a distant second at 7%). It is also diversifying beyond logic: on August 11, 2026, TSMC and Sony executed a definitive agreement for a ~$6.3B joint venture (Sony 60%/TSMC 40%) to build a Kumamoto image-sensor fab targeting 2029 mass production. TSM shares (and the broader semiconductor sector) slid as much as 3.2% on August 24 amid pre-earnings profit-taking ahead of top customer Nvidia's August 26 results, with TSMC's market cap pulling back to roughly $1.97T from an early-August peak near $2.1-2.2T. Key risks: extreme geographic concentration in Taiwan amid China tension, now including a Chinese proposal (reported mid-August 2026, not yet enacted) to bar TSMC from making chips based on Chinese-company designs; US tariff policy (a January 2026 Section 232 25% tariff on advanced logic, mitigated by the US-Taiwan deal's quota exemptions tied to the now-$265B US buildout); margin dilution from overseas fabs; and any AI-capex correction, which CEO C.C. Wei has publicly dismissed while pledging disciplined investment.

Revenue Streams

High-performance computing (AI accelerators & data center)

Fabricating GPUs, AI accelerators, and server CPUs for Nvidia, AMD, Broadcom, and hyperscaler custom-silicon programs - 66% of Q2 2026 revenue and up 20% quarter-over-quarter, the dominant and fastest-growing stream, now two-thirds of the company.

Smartphone SoCs

Application processors and modems for Apple and other handset makers - 22% of Q2 2026 revenue (down 4% sequentially), historically TSMC's largest segment before AI overtook it.

IoT, automotive & consumer electronics

Specialty and mature-node chips for IoT devices (5% of Q2 2026 revenue), automotive (4%), and digital consumer electronics (1%), largely from 8-inch fabs and the Japan/China sites.

Advanced packaging (CoWoS / SoIC)

2.5D/3D packaging that assembles GPU dies with high-bandwidth memory - the binding constraint on AI accelerator supply. Capacity is being quadrupled toward ~130-140k wafers/month by late 2026, absorbing 10–20% of the 2026 capital budget; TSMC began outsourcing part of the process to OSAT partners in August 2026 to ease the bottleneck.

Key Metrics

Employees

90,557 (as of Dec 31, 2025)

Est. Annual Revenue

$122.42B (FY2025, +35.9% YoY; net income $55.21B); FY2026 revenue guided to grow more than 40% YoY in USD terms (raised Jul 16, 2026, from more than 30% in April). July 2026 monthly revenue of NT$467.58B (+44.7% YoY) is already outpacing the Q3 guided 36-37% growth range; Jan-Jul 2026 cumulative revenue reached NT$2,872.06B (+37% YoY)

Foundry market share

73% of the pure-play foundry market (Q2 2026, per Counterpoint Research) - Samsung a distant second at 7%, a 66-point gap, as the sector grew 29% YoY on AI-driven demand. Different research firms' foundry-share figures vary by methodology (TrendForce put Q1 2026 at 72.3%/69.9% FY2025)

Latest quarterly revenue

NT$1.27T / ~US$40B (Q2 2026, +36% YoY); net profit NT$706.6B (+77% YoY, record)

Jan-Jul 2026 cumulative revenue

NT$2,872.06B (+37% YoY); July alone NT$467.58B / ~US$16.03B (+44.7% YoY, +5.6% MoM), a record month on AI chip demand

Market cap

~$1.97T (Aug 26, 2026; TSM ADR ~$417.69), down from the ~$2.1-2.2T level reached earlier in August after a semiconductor-sector pullback (TSM -3.2% on Aug 24) ahead of Nvidia's Aug 26 earnings

2026 capex guidance

$60–64B (raised from $52–56B on Jul 16, 2026); no further revision as of the next scheduled update at Q3 2026 earnings (~mid-October 2026)

2027 wafer price increase

Base prices to rise 5-10% across advanced nodes (5nm/4nm/3nm) from Jan 2027, plus a 10-15% surcharge on incremental HPC orders (up to ~25% combined); also TSMC's first mature-node (12/16/28nm) price increase in over three years

Q3 2026 revenue guidance

$44.6-45.8B guided (+37% YoY at midpoint, +12% QoQ), issued Jul 16, 2026; gross margin guided 65-67%, operating margin 56-58%

Gross margin

67.7% (Q2 2026 actual, record; up from 66.2% in Q1). Guided down to 65-67% in Q3 2026 as the N2 ramp and overseas-fab dilution weigh on margins

Advanced-node share of wafer revenue

77% from ≤7nm (Q2 2026): 2nm 3%, 3nm 30%, 5nm 33%, 7nm 11%. Q3 2026 breakdown not yet reported

Fabs

18 (incl. six 12-inch GIGAFABs); >17M 12-inch-equivalent wafers/yr

Customers / products (2025)

534 customers, 12,682 distinct products

Arizona subsidiary H1 2026 profit

NT$36.07B (~$1.1B), up ~663% YoY and already above full-year 2025 profit; Japan subsidiary JASM also swung to profit in 2026, its first profitable stretch since Kumamoto production began in late 2024

Timeline

2026Record FY2025 results and history's largest chip capex budget

On January 15, 2026, TSMC reports record 2025 revenue of $122.42B (+35.9%) with net income up 51.2%, and guides 2026 capex to $52–56B - the largest single-year capital budget in semiconductor history. The same day, a US–Taiwan trade deal exempts US-investing Taiwanese chipmakers from new Section 232 tariffs.

2026Roadmap through 2029 unveiled; A16 slips to 2027

At the April 2026 North America Technology Symposium, TSMC lays out N2P and A16 (now 2027, from H2 2026), A14 next, and A12/A13 for 2029 - notably planned without ASML's High-NA EUV tools.

2026Q2 2026 profit jumps 77% to a fifth straight record quarter

On July 16, 2026, TSMC reports second-quarter net profit of NT$706.6 billion (about US$22B), up 77% YoY to a record and far above the ~NT$632.6B analyst consensus, on record revenue of NT$1.27 trillion (+36% YoY). Demand for leading-edge N3/N5 nodes and advanced CoWoS packaging - reported sold out through year-end - kept capacity near full as the AI buildout continued.

2026Pledges ~$100B more for Arizona; raises 2026 capex to $60–64B

On its July 16, 2026 earnings call, TSMC said it will invest roughly $100B more to expand its Arizona campus with at least four additional fabs for 2nm-and-below nodes, lifting its total planned US investment to about $265B - the largest foreign direct investment in US history. It also raised 2026 capital-spending guidance to a record $60–64B (from $52–56B), with 70–80% earmarked for advanced process technology, and expects ~30% of its 2nm-and-below capacity to eventually sit in the US.

2026Kumamoto earthquake evacuates JASM fab; operations resume

A magnitude 7.1 earthquake strikes Japan's Kumamoto region on July 28, prompting TSMC to evacuate its JASM fab under standard safety protocols. TSMC says all employees are safe and post-quake inspections find the fab structure unaffected, with operations gradually resuming from July 28 - the first major earthquake to hit JASM since the majority TSMC-owned joint venture began production in 2024. The affected fab runs mature nodes (12-28nm) and represents under 3% of TSMC's total wafer capacity (about 1% of 2026 revenue), so TSMC and analysts expect only a limited impact on Q3 2026 results.

2026Develops EMIB-like packaging with Kinsus to counter Intel

With CoWoS-L capacity sold out through 2026 and into 2027 (lead times reportedly stretching to 78 weeks), TSMC is reported to be developing an EMIB-like advanced-packaging technology internally, working with Taiwanese IC-substrate maker Kinsus to scale it from development to volume production. The move responds to Intel's rival EMIB packaging, a cheaper substrate-embedded-bridge approach that has reportedly drawn interest from Nvidia, Google, and OpenAI as an alternative to TSMC's silicon-interposer-based CoWoS.

2026Finalizes 2027 price increases of up to 10% across chip nodes

After negotiations that began in June, TSMC locks in base wafer-price increases of 5-10% across advanced nodes (5nm/4nm/3nm) effective January 2027, plus a 10-15% surcharge on incremental HPC orders (pushing some increases toward 25%); it is also TSMC's first price increase on mature nodes (12/16/28nm) in more than three years. Apple, Nvidia, AMD, and Qualcomm are among the customers facing higher wafer costs.

2026Pulls forward 3nm and 2nm capacity ramps

TSMC is on track to reach 180,000 3nm wafers per month by early Q4 2026, two to three months ahead of its prior schedule, on strong demand from customers including Nvidia, AMD and Broadcom, per an early-August 2026 TrendForce report. 2nm capacity is expected to reach 100,000 wafers per month by year end, building on the record $60-64B 2026 capital-spending guidance TSMC raised in July.

2026Second Kumamoto fab's 3nm upgrade approved; timeline moves to 2028

Taiwan's government formally approves TSMC's February 2026 plan to upgrade JASM's second Kumamoto fab from 6/7nm to 3nm - Japan's first leading-edge node - raising the investment from ~$13.9B to ~$17B for 15,000 wafers/month of capacity. Equipment installation and mass production are now targeted for 2028, later than the fab's pre-upgrade late-2027 goal.

2026Announces $6.3B Sony image-sensor joint venture

TSMC and Sony Semiconductor Solutions announce plans to jointly invest roughly $6.3B (about JPY 1 trillion) in a new fab inside Sony's Kumamoto image-sensor campus, following a May 2026 MOU. Sony holds a 60% controlling stake and TSMC 40%; the venture targets 2029 mass production of next-generation image sensors for automotive and robotics customers - TSMC's first move beyond logic and specialty chips into image sensors.

2026Board approves $29.4B capex tranche; Sony JV made definitive

At a board meeting concluding August 11, 2026, TSMC approves roughly $29.44B in capital appropriations for advanced-technology, packaging and fab construction spending, and executes a legally binding definitive agreement with Sony Semiconductor Solutions to establish Advanced Vision Semiconductor Manufacturing Corporation, the Kumamoto image-sensor joint venture first announced August 10. Sony will contribute about JPY 465B and TSMC about JPY 282B to the venture in phases, consistent with the previously disclosed 60%/40% split and 2029 volume-production target.

2026Expands CoWoS outsourcing to OSATs as 5.5-reticle packaging tops 99% yield

In early August 2026, TSMC is reported to expand outsourcing of the CoW (chip-on-wafer) front-end packaging step to OSAT partners including ASE - beyond its earlier practice of outsourcing only the later WoS step - to relieve CoWoS capacity pressure driven mainly by Nvidia, which has reserved an estimated 800,000-850,000 CoWoS wafers for 2026 (over half of TSMC's total). Separately, TSMC reports its 5.5-reticle-size CoWoS-L is in volume production with yields topping 98-99% across multiple AI customer products, while flagging memory and ABF substrates - not the CoWoS process itself - as the AI supply chain's next bottleneck; a 14-reticle CoWoS package is slated for 2028.

2026A16 development and verification completed; Q4 2026 production readiness confirmed

TSMC confirms it has finished development and verification of the A16 (1.6nm-class) process, the industry's first angstrom-class node with Super Power Rail backside power delivery, and reiterates it is on track for initial Q4 2026 production. This is manufacturing readiness rather than a reversal of the April 2026 roadmap update: TSMC's own guidance still ties broad-based volume production, dependent on customer product qualification, to 2027.

2026China weighs export controls that would bar TSMC from Chinese-designed chips

China's Ministry of Commerce is reported, as of mid-August 2026, to be soliciting industry feedback on export controls covering advanced AI models, training data, and foreign chipmaking access - including a measure that would bar overseas foundries like TSMC from manufacturing advanced processors based on designs from Chinese companies such as Alibaba, ByteDance, and Huawei. The proposal, not yet enacted, would mirror US-led export controls in reverse and adds a new geopolitical risk vector alongside existing US Section 232 tariff exposure.

2025$100B additional US investment announced ($165B total)

On March 4, 2025, TSMC announces three more fabs, two advanced-packaging plants, and an R&D center in Arizona, lifting its total US commitment to $165B - the largest foreign direct investment in US history.

2025N2 (2nm) enters volume production on schedule

TSMC's first gate-all-around nanosheet node enters volume production in Q4 2025 at Fab 22 (Kaohsiung) and Fab 20 (Hsinchu), confirmed December 30, 2025 - with reported yields around 70% and a claimed 10–15% speed gain or 25–30% power cut versus N3E.

2024JASM Fab 1 opens in Kumamoto, Japan

TSMC's Japanese joint venture with Sony, Denso, and Toyota opens its first fab in February 2024, producing 12–28nm chips - the anchor of Japan's semiconductor revival.

2024Arizona Fab 21 enters volume production

Phase 1 of Fab 21 begins high-volume 4nm (N4) production in Q4 2024 - TSMC's first US leading-edge output, with yields on par with Taiwan fabs and customers including Apple and Nvidia.

20223nm (N3) volume production begins in Tainan

TSMC starts volume production of N3, the industry's most advanced node, extending its leadership through the generation that would power the first wave of AI accelerators.

20205nm production begins; Arizona fab announced

Volume production of 5nm chips begins (Apple's A14 and M1 era), and TSMC announces its first US leading-edge fab in Phoenix, Arizona, with an initial $12B commitment.

2018Takes process leadership with 7nm; C.C. Wei becomes CEO

TSMC becomes the first to high-volume 7nm production, seizing process leadership from Intel - a lead it has never relinquished. C.C. Wei becomes CEO the same year.

1994Lists on the Taiwan Stock Exchange

TSMC lists on the TWSE under ticker 2330, followed in 1997 by NYSE ADRs (TSM) - the first Taiwanese company to list on the NYSE.

1987Morris Chang founds the world's first pure-play foundry

TSMC is founded in Hsinchu with Taiwan's government providing 48.3% of the capital and Philips 27.6% ($58M) - the first semiconductor company to manufacture only other companies' chip designs, the business model that reshaped the industry.

Funding

RoundDateAmountInvestorsSource
Founding capital1987Taiwan government 48.3%, Philips $58M (27.6%)Government of Taiwan, Philips, private Taiwanese companies (~20%)
TWSE listing1994IPOPublic markets (TWSE: 2330)
NYSE ADR listingOct 1997ADR listingPublic markets (NYSE: TSM) - first Taiwanese company on the NYSE

/ Related Companies

All Semiconductors
ASML logo

ASML

Semiconductors

ASML, founded in 1984 as a Philips/ASM International joint venture that built its first machine in a shed in Eindhoven, is the world's dominant supplier of photolithography systems (~83% of the market) and the only company on Earth that makes EUV machines - the $180M-$400M tools without which no leading-edge chip can be manufactured. That monopoly has made the Veldhoven company one of Europe's most valuable tech firms (~$601B, July 2026, down from ~$693B in July after a China-competition scare - see below). 2025 revenue reached €32.7B with a record €13.2B in Q4 bookings and a €38.8B year-end backlog (ASML has since stopped disclosing quarterly bookings), and after twice raising 2026 guidance during the year - most recently at Q2 2026 to €43-45B - CEO Christophe Fouquet says chip demand is outpacing supply, with 65 low-NA EUV systems planned for 2026, rising toward 80-85 in 2027. The next act is High-NA EUV: the first EXE:5000 shipped to Intel in December 2023, the production-grade EXE:5200B arrived at Intel in Q4 2025 for its 14A node (risk production still targeted 2027), and on July 15, 2026 Intel became the first company to ship a high-volume commercial product patterned with High-NA EUV - select layers of its Panther Lake (Core Ultra Series 3, 18A) chips are now dual-qualified for either 0.33 NA or 0.55 NA scanners. Samsung, SK hynix and imec also hold early High-NA systems - fewer than a dozen exist worldwide, at roughly $380M each - but Samsung is reportedly holding its units back from mass production to contain foundry costs. TSMC, by contrast, has ruled out High-NA through its A16, A14, A13 and A12 nodes (production planned through 2029), citing cost - a real demand risk for ASML's most expensive tool. ASML also made an uncharacteristic strategic bet in September 2025, investing €1.3B to become the largest shareholder (~11%) of French AI lab Mistral AI. The main headwind is geopolitics: China was 33% of 2025 sales and held at roughly 20% of net sales through Q2 2026 under expanding US and Dutch export controls; a July 27, 2026 report that a state-linked Chinese entity has begun mass-producing a domestic immersion DUV tool (assessed as roughly equivalent to ASML's ~2008-era NXT:1950i, 15-18 years behind ASML's current platform) briefly knocked ASML shares down as much as 8%. Separately, US Commerce Secretary Howard Lutnick told senior ASML executives in a series of meetings beginning in April 2026 that Washington believes an EUV system or EUV-specific components may have reached China in breach of export controls - a claim ASML has categorically denied, backing its denial with an accounting of all 314 operational and 26 decommissioned EUV units worldwide by location; no formal enforcement action had been announced as of ASML's July 15, 2026 earnings call.

BYD logo

BYD

Semiconductors

BYD (Build Your Dreams) is a Chinese high-tech manufacturer spanning automobiles, batteries, electronics, renewable energy, and rail transit. Its auto business is unusually vertically integrated, with in-house batteries, motors, power electronics, and automotive semiconductors underpinning a lineup that now ranges from low-cost EVs and plug-in hybrids to premium Denza and Yangwang vehicles. First-half 2026 revenue fell 7.13% YoY to ¥344.82B (~$50.9B) and net profit dropped 20.54% to ¥12.33B as China NEV weakness and foreign-exchange losses outweighed overseas growth; auto-related revenue was ¥275.34B (-8.98%) while gross margin rose to 18.85% from 18.01%. H1 NEV sales were 1,808,511 (-15.72%), with the Q2 decline narrowing to 3.24% after a 30% Q1 drop, and overseas shipments of about 792,000 (+67.8%) making up roughly 44% of H1 volume. August 2026 NEV wholesales reached 440,293 (+17.84% YoY), the highest month of 2026, with a record 189,466 overseas units (43.03% of volume) offsetting still-soft domestic sales. On August 28, 2026 BYD opened its 10,000th FLASH Charging station in Shenzhen, doubling the network in under five months across 325 cities and reaching halfway to its 20,000 year-end target, with 1.83 million users (nearly one-third non-BYD). BYD's Szeged plant in Hungary began trial production in January 2026, but EVP Stella Li said in June 2026 that full-scale series production has slipped to Q4 2026 as equipment is still being installed; the same month BYD put its planned €1B Turkey (Manisa) plant on hold after Ankara withdrew tax incentives, making Hungary its top European manufacturing priority.

Nvidia logo

Nvidia

Semiconductors

Nvidia, founded in 1993 by Jensen Huang, Chris Malachowsky, and Curtis Priem, invented the modern GPU (GeForce 256, 1999) and then - via the 2006 CUDA platform - turned it into the engine of the AI revolution. It touched a record $5.06 trillion valuation in late July 2026 and then briefly slipped to about $4.77 trillion on July 27, 2026 as Apple reclaimed the world's-most-valuable-company title amid a broader AI-financing selloff, but Nvidia rallied back through early August - its biggest one-week market-cap gain ever - to reclaim the crown at about $5.42 trillion (Aug 7, 2026 close), before easing back to roughly $5.29 trillion (Aug 28, 2026 close, $217.89/share) after Q2 FY2027 gross-margin guidance drew a mixed reaction. Fiscal 2026 revenue hit $215.9B (+65%), and the momentum accelerated: Q2 FY2027 (ended July 2026, reported August 26, 2026) delivered a record $96.2B (+106% YoY, beating the $92.2B consensus) with data-center revenue of $89.0B (+117% YoY), CFO Colette Kress guiding Q3 to ~$108B and roughly 70% revenue growth for FY2028, and supply commitments more than doubling to $279B (mostly HBM memory) against a hyperscaler cloud backlog exceeding $2 trillion. The Blackwell Ultra GB300 platform drove the ramp, and Vera Rubin - the seven-chip platform Huang declared in full production at GTC 2026 - reached customers on July 21, 2026, shipping to OpenAI, CoreWeave, Google Cloud, Microsoft Azure, Meta, and Dell, with OpenAI set to deploy it at scale in Q3 2026 and roughly $1 trillion in cumulative Blackwell-plus-Rubin orders projected through 2027. Nvidia has also become the AI buildout's central dealmaker: up to $100B invested in OpenAI (whose Ohio campus financing was finalized at $105B for an initial 4.25 IT-GW, online in phases from 2028), up to $10B in Anthropic, a key role in the $500B Stargate project, millions of GPUs committed to Meta, an August 10-11, 2026 financing coalition with Apollo, Blackstone, BlackRock, Brookfield, Goldman Sachs, and KKR aiming to mobilize over $500B in third-party AI-compute capital, and - confirmed September 3, 2026 - a $12.93B deal to acquire AI model-hosting platform Hugging Face, Nvidia's largest acquisition ever, alongside talks to invest in Perplexity at a valuation above $30B, plus a confirmed August 31, 2026 $3.5 billion convertible-bond investment in MediaTek that opens NVLink Fusion to MediaTek custom XPUs in Nvidia rack-scale AI factories. That expanding web of AI financing, alongside a recurring circular-financing debate, remains the biggest swing factor in Nvidia's valuation. The China business is the big casualty - export restrictions triggered a $4.5B charge in 2025, and while H200 shipments to China stayed 'trivial' through mid-July 2026 under the deal with a 25% cut to the US government, ByteDance and Tencent each received roughly 10,000 H200 units in the weeks before August 19, 2026, still only about 13% of the 75,000-unit licenses each holds; Nvidia's 10-K still calls the company effectively foreclosed from China's data-center market and its guidance assumes zero China data-center revenue. Competition now comes less from AMD (still ~5-7% of AI accelerator revenue) than from customers' own silicon: Google TPUs, AWS Trainium, Broadcom-built custom ASICs, and OpenAI's own Jalapeno inference chip unveiled August 25, 2026.

SpaceX logo

SpaceX

Semiconductors

Space Exploration Technologies Corp. (SpaceX) designs, manufactures, and launches reusable rockets and spacecraft. The company operates Falcon 9, Falcon Heavy, Dragon, Starlink, and the government-focused Starshield line while continuing to develop Starship for high-cadence heavy-lift missions to the Moon, Mars, and beyond. Starlink crossed 12 million customers across 160+ countries on June 4, 2026 and the constellation now exceeds 10,400 working satellites in orbit. Following the February 2026 xAI merger, SpaceX also operates the Memphis Colossus supercomputers as an AI hyperscaler, disclosing $14.1B of new Cloud Services Agreements signed in Q2 2026 alone (on top of the Anthropic, Google, and Reflection AI compute deals). SpaceX completed the largest IPO in history on June 12, 2026, debuting on Nasdaq under ticker SPCX at its fixed $135-per-share price for a roughly $1.77 trillion valuation and about $75B in proceeds, with Elon Musk retaining ~82% voting control; SPCX slid as low as $107.01 in late July before rebounding to close above its $135 IPO price on August 10, 2026 for the first time since mid-July. Starship returned to flight after the FAA closed the Flight 12 mishap investigation on July 13, 2026: on July 24, 2026 Flight 13 - the second flight of the Version 3 vehicle - deployed the program's first 20 operational Starlink V3 satellites and made the softest ship splashdown yet, and recovery crews towed the intact upper stage (Ship 40) through rough seas to Christmas Island, Australia by August 18, 2026 - the first Starship upper stage ever recovered intact after flight - before the semi-submersible vessel Forte departed Christmas Island on August 28, 2026 to carry it roughly 20,000 km back to Starbase, Texas. Days after its IPO, SpaceX agreed on June 16, 2026 to acquire AI coding startup Cursor's parent Anysphere for $60B in an all-stock deal - the largest acquisition of a venture-backed startup ever - and closed the deal on August 15, 2026, ahead of its late-August target, giving Cursor access to SpaceX's GPU fleet including the Colossus supercomputer. SpaceX posted its first earnings as a public company on August 4, 2026 (Q2 revenue of $7.8B, up 92% year over year, narrowing its net loss to $541M), and on August 11 Elon Musk told employees SpaceX's AI compute business would exceed all its other revenue by September 2026 and account for 99% of the company's value within five years, targeting 10 gigawatts of AI compute capacity - up from 1.4 GW at the end of Q2 - by the end of 2027. On August 25 SpaceX announced a $100 billion Starbase Louisiana spaceport in Vermilion Parish, with construction expected in 2027 and a first launch targeted as soon as 2029. A Falcon Heavy flew its 13th mission on August 30, 2026, launching NASA's Roman Space Telescope with both side boosters landing intact at Cape Canaveral, a day after NASA and SpaceX postponed the planned September 12 Crew-13 ISS launch when an oxidizer leak was found in Dragon's propulsion system during prelaunch processing. A mid-August 13F filing also disclosed Nvidia holds roughly $21B of SpaceX stock via its earlier xAI investment, as SpaceX committed on its August 4 earnings call to building its future AI compute exclusively on Nvidia's Vera Rubin architecture. SPCX shares climbed about 7% to close at $150.84 on September 3, 2026 after Oppenheimer raised its price target to $280 from $250 on growing confidence in the AI-compute business, with investors also looking ahead to the approaching Starship Flight 14 test.

Tesla logo

Tesla

Semiconductors

Tesla is a vertically integrated automotive, energy storage, and AI company headquartered in Austin, Texas. After building its final Model S and Model X on May 10, 2026 and converting that Fremont line to Optimus, it now manufactures three consumer vehicles (Model 3, Model Y, Cybertruck), sells Full Self-Driving (Supervised) software, operates energy storage and solar businesses, runs an unsupervised Robotaxi service across seven US metros (Austin, Dallas, Houston, Miami, Orlando, and Tampa driverless, plus a supervised Bay Area pilot), and is ramping new products including Cybercab (in production at Giga Texas since Feb 2026; commercial paid rides began in Austin on September 3, 2026, the same day NHTSA opened audit AQ26002 on how Tesla self-certified the steering-wheel-free vehicle), Tesla Semi (now in a commissioning phase at the new Nevada line as it works toward a 50,000-unit/year target), and Optimus (entered production on Fremont's converted first-generation line around August 27-28, 2026, feeding an internal training program rather than customer sales; the V3 public reveal remains undated). In Q2 2026, Tesla generated a record $28.24B in revenue (+26% YoY) on 480,126 vehicle deliveries, though GAAP operating margin narrowed to 1.4% amid heavy AI, Optimus, and Robotaxi investment, and guided to over $25B of full-year 2026 capital expenditure; it had 1.48M active FSD subscriptions and operated 8,704 Supercharger stations.

Varda Space Industries logo

Varda Space Industries

Semiconductors

Varda Space Industries manufactures materials in orbit and brings them back to Earth, a business no company had ever run commercially before it. Founded in 2021 by former SpaceX Dragon engineer Will Bruey and Founders Fund partner Delian Asparouhov, it builds the W-Series: free-flying spacecraft that process pharmaceuticals in microgravity and then return the product inside a capsule that reenters at over 18,000 mph and lands under parachute, usually at the Koonibba Test Range in South Australia. The commercial logic is that microgravity produces crystal structures and morphologies that gravity makes impossible on the ground, and Varda proved it on its first flight by growing and recovering the metastable Form III of ritonavir, an HIV antiviral. Six missions have flown since 2023, and five of the six capsules made it home; the fourth, W-4, spent nearly a year on orbit while Varda evaluated a propellant-feed problem before closing out the mission with a disposal burn in May 2026 rather than a recovery. W-5 returned in January 2026 as the first full mission on Varda's own vertically integrated satellite bus, carrying a U.S. Navy payload, and W-6 returned in May 2026 flying thermal-protection and autonomous-navigation experiments for NASA and defense customers under the AFRL's Prometheus program. That second business, selling the capsule as a cheap, high-cadence hypersonic reentry testbed, now runs alongside the pharmaceutical work, and in October 2025 Varda added a third line, agreeing with United Semiconductors to grow semiconductor crystals in microgravity for autonomous-systems, sensing, AI, and defense customers. In May 2026 United Therapeutics signed on to develop microgravity formulations for rare pulmonary disease, the first time a publicly traded pharmaceutical company has spent its own capital on commercial drug manufacturing in space, and Varda leased a former Mattel manufacturing plant in El Segundo in early 2026 to scale spacecraft production to match.