TSMC

TSMC has surveyed suppliers on using rare-earth-free motors in factory-automation equipment...effort driven by concerns about potential supply ​disruptions from China. Of course they are explorting, viable solutions are another thing…

Just in Taiwan to shut Friday as Typhoon Champi approaches. Out of an abundance of caution, TSMC canceled its annual Sports Day scheduled for Saturday. https://t.co/wUrYhlqzKS

Even Taiwan's second tier chipmakers are printing money right now. According to Taiwanese media, Powerchip just reported third quarter revenue up 80% from a year ago and net profit up 338%. Its gross margin jumped from 28% to 48% in a…

Expert call covering GUC – Google CPU & TPU v10 ASIC outlook, Microsoft Maia 200 ramp, TK1 vs. TK3 margin dynamics, TSMC capacity support, and 2027 hyperscaler volume trajectory. $MFFT $META $GOOGL https://t.co/VvdNhBSJFv

Multiple outlets: Nvidia wants next-gen custom HBM base dies on TSMC 3nm, and Samsung may use rival TSMC lines for that die. Unconfirmed. Memory makers keep the stack; the logic die’s foundry choice is becoming a Nvidia-spec decision.…

Demanding Korean suppliers beat Chinese pricing while building 100,000 wafers per month misunderstands semiconductor supply economics. Wafer fab tools and pure chemical consumables are priced by yield and purity, not procurement leverage.…

Elon Musk just put dates on Terafab. According to Reuters, Musk said on the SpaceX earnings call that Terafab's first working SRAM test chip is targeted for May or June 2027 at the research fab in Austin. Volume production is targeted for…

After eight long years, TSMC is finally shrinking their 6T High Density SRAM bitcell for their A14 process! At IEDM, they will present the world's smallest bitcell, <0.017μm2 on their 2028 node. TSMC has been stuck at 0.021μm2 since N5 in…

$SPY rose about 0.2% today to around $768, bouncing from an early low of $764.25 as the Dow led and tech lagged. Here are the top stories you need to know. 1. South Korean neocloud Elice plunged 32% in its trading debut, dragging the KOSPI…

$MU $SNDK $SKHY $NBIS Nomura’s memory team just put out a really interesting note on the AI scare, the numbers underneath are some of the strongest I’ve seen in this drawdown. Nomura sees AI going from 10GW of data center power in 2025 to…

"Intel is 80% done." That is how Lip Bu Tan described the turnaround at $INTC in a new interview with Nikkei. Since taking over in March 2025, he has replaced most of the senior leadership team, including bringing in the former CEO of SK…

Intel's CEO, Lip-Bu Tan, said China could catch up to U.S. semiconductor technology within 5-10 years. Regarding AI, he cited Moonshot AI's and DeepSeek as not technologically inferior to Anthropic or OpenAI and are significantly cheaper.…

Well said Yeok! -your neighborhood engineer who is also a licensed engineer in China H/t: @walterpchen https://t.co/fVZHeRylPh

Samsung Electronics collaborating with TSMC to develop NVHBM. $TSM https://t.co/xHuBksxsqm https://t.co/M15yv0KZIG

Samsung is teaming up with its biggest foundry rival to build Nvidia's next memory chip. According to Korean media, Samsung will use $TSM's 3nm process for the base die in $NVDA's custom NVHBM, at Nvidia's request. The base die is the…

China is on track to domestically produce most key semiconductor manufacturing equipment for mature nodes by 2030, led by NAURA, AMEC, and SiCarrier, according to TrendForce. Except lithography where Chinese domestic market share is…

Two GF moves in packaging: a five-year deal worth about $2B supplying silicon interposers from Malta, NY into TSMC's CoWoS ecosystem, and $955M of its CHIPS award reallocated by Commerce to silicon photonics, SiGe and advanced packaging.…

J.P. Morgan: We expect A14 to go into mass production in 1H28, which could generate some revenue in 2H28. We believe the tight leading edge wafer supply will extend into 2027-208 and expect TSMC to continue its accelerated capacity build…

TSMC Q3 revenue $46.7bn beat its own $44.6-45.8bn guide, HPC/AI 68% of sales, Q4 guided to $52.3-54.3bn, about 14% sequential growth. Gross margin fell 130bp to 66.4%; overseas-fab dilution widens from ~2% to 3-4%. The stock fell 4.5%.…

JPMorgan expects $TSM to keep growing its CoWoS packaging capacity. - 130,000 wafers a month by the end of 2026. - 260,000 by the end of 2027. - 320,000 by the end of 2028. $INTC says its EMIB packaging is 30% to 40% cheaper. But JPMorgan…

Taiwan is planning industrial parks in the United States for $NVDA's AI supply chain. The first one is in Texas and could cost up to $4 billion. More parks are planned, at $2 to $4 billion each. The goal is to bring the smaller suppliers…

Analysis: TSMC's chipmaking edge strengthens C.C. Wei's hand — from advanced-node capacity to Musk's Terafab https://t.co/CVsF9sIYED

BofA sees SpaceX consuming ~20% of TSMC's advanced node capacity in 2029, up from ~1% in 2026, due to limited support from Terafab. $SPCX $TSMC https://t.co/kSutYfLLSI

According to Korean media, Microsoft, Google, Amazon and Meta are speeding up their in house AI chips. That could open up a SECOND group of HBM buyers beyond $NVDA. Microsoft's Maia 300 is expected to ship more than 1 MILLION units next…

I asked TSMC chairman CC Wei today whether he can share more on his company’s possible deal with @elonmusk and @TerafabCorp Ever the joker, he told me “I’ll share more at the next investor conference” Friends, they just held an IR call…

‘Doing What It Takes’ in AI. OpenAI, Meta & TSMC. ARD #185 …OpenAI’s ad sprint, Meta’s open Muse on Apple’s Macs, TSMC on its own terms, and Apple’s Ternus era. $TSM $META $AAPL Full writeup and source links: https://t.co/phJT1JuRTJ…

This is it. This is the key quote from TSMC on @TerafabCorp And yet, everyone missed it — too caught up on CC Wei’s comment about running the fab to notice what he said about excess capacity at the fab. And note what @elonmusk said about…

Two weeks ago @Culpium broke news of talks between TSMC & @TerafabCorp @elonmusk has said his team will run the fab TSMC just said it will run the fab Both sides have dropped hints as to how it may work. Two things can be true at once…

TSMC finally speaks about Terafab, and paints a picture of a new future https://t.co/OpS1jNtbAW

This is the way ❤️ https://t.co/ypUyyfMvc3

TSMC’s wafer ASP gap with UMC, GlobalFoundries and Samsung keeps widening. TSMC is setting new blended ASP highs while most peers remain below their previous peaks. Per our Foundry Model, from 2Q22 to 2Q26, TSMC’s wafer revenue share…

TSMC, Intel, Samsung seek to break AI chipmaking's next barrier through early technology integration https://t.co/WYOjq9fqQh

TSMC is outsourcing part of CoWoS. A five-year $2B deal puts CoWoS-S silicon interposers at GlobalFoundries' Malta, New York fab, volume production H1 2028. TSMC's Arizona CoWoS arrives before 2029. Packaging capacity is being duplicated…

CoWoS interposers have never left TSMC's own line. Per igor'sLAB, TSMC plans to have GlobalFoundries manufacture them — the first external partner for the packaging behind Nvidia and AMD AI chips. Unconfirmed. Packaging, not wafers, is the…

Maia 200 shipped in January. Maia 300 lands nine months later: TSMC N2P, 288GB HBM4 at 18TB/s vs 216GB at 7TB/s, 96 per rack vs 48, 950W vs 750W. Microsoft claims 40% more tokens per megawatt than GPU systems; production 2027. The cadence…

Damnang not only has good writing, but good visuals! This is a great visual showing the difference in HBM4 between Samsung and SK Hynix. Please have a read at this article! https://t.co/OMa9Q1boO6 https://t.co/vrxucwF34V

J.P. Morgan: Recent investor discussion has been focused on delays in TPUv8t ramp, v9 tapeout and uncertainty on v10 allocation. Based on our checks, MediaTek's TPUv8t Zebrafish continues to ramp on schedule at TSMC, and remains on track…

Just in: CASCI Posts First Drop in 12 Months as Constraints Bite Midstream suppliers get squeezed from both sides, Culpium AI Supply Chain Index shows. https://t.co/cFkMLGO8El https://t.co/pEJZoh1Ewz

The 6×6-inch photomask has not changed in nearly half a century. High-NA EUV's 4x/8x anamorphic optics halve the exposure field, so ASML and TSMC are pushing a 6×12-inch format: pilot line 2031, production support 2033. The mask supply…

$AMAT just opened the largest joint chip research facility in the world, and all the major chipmakers are in it. The EPIC Center in Sunnyvale, California covers about 600,000 square feet, with nearly four acres of cleanroom space. Spending…

The chip sell off just spread across ALL of Asia. South Korea's KOSPI is down about 6% and has dropped below 6,000. Japan's Nikkei is down about 2.6%, and MSCI's Asia Pacific index outside Japan is down about 2%, according to Reuters. Not…

Cantor Fitzgerald on TSMC's CapEx: "As for CY27, while TSMC has historically saved its out-year guide for January, our work now suggests ~$90-100B of Capex (vs $80B+ in July, when we flagged upside to as high as $90B), supported by…

TSMC, ASML eye 12-inch photomasks to unlock High-NA EUV's full-field potential https://t.co/TrrqvTChCj

This is the reason Microsoft is cutting its Claude token usage. Microsoft just announced it will scale its own AI chip, Maia, to multigigawatt deployments starting in 2027. Maia will become the main system running AI inside Microsoft's…

“The limiting factor will be turning the chips on,” Musk said. “You need power and transformers and cooling.” Inside TSMC, Texas, and the Terafab Tango https://t.co/VS2hdLGuR0

How is a lightning-fast pace of 10 chip designs a year even possible? 🤯 💡 Here’s my take: Terafab’s path to faster chip development starts with the research fab at Giga Texas. By bringing mask-making, fabrication, packaging, and testing…

Inside TSMC, Texas, and the Terafab Tango https://t.co/Vauee9FDQ1

With the Singapore Grand Prix in a few hours, I can't help but draw comparisons to the financial markets and the AI trade. One slow pitstop or a miscalculated overtake can undo weeks of preparation and cost millions of dollars. So with AI…

Coming soon. Exclusive to @culpium subscribers later today. It's been a busy week of news, speculation, reaction and responses surrounding Intel, TSMC and Terafab. Make sure you're signed up when this piece drops. https://t.co/vIR00A0QoG…

Morgan Stanley has TSMC's optical engine shipment increasing to 73.87 million units in 2028, up from 0.39 million in 2026. $TSMC https://t.co/i3BhqwJdGX

BofA has CPO TAM growing to $14.1 billion in 2030E, up from $0.2 billion in 2026E. 206% CAGR '26-'30. $LITE $NVDA $COHR $MRVL $AVGO $TSM https://t.co/1xHSwbPFQT

🇺🇸EXCLUSIVE: @elonmusk says @Tesla’s real edge isn’t the cars; it’s the custom manufacturing equipment most people never see. After touring Intel, Samsung, and TSMC chip fabs, he told @teslaownersSV at @theXtakeover he was shocked to…

Nvidia is now redesigning how AI memory works and it is pulling TSMC right into the middle of the HBM business. According to TheElec, Micron will use TSMC to make the base die for its HBM4E. Nvidia's custom NVHBM moves the memory…

🇺🇸Intel will remain involved in @elonmusk 's Terafab chipmaking project, CEO Lip-Bu Tan told Bloomberg, despite discussions about potential cooperation with Taiwan’s TSMC. Musk has been exploring partners for the venture as he develops…

Sabi designed its own non-contact EEG chip, had TSMC fab it, and put it in a cap that turns your thoughts into text! Custom ASICs, custom sensors, and a brain foundation model trained on their own neural data. $50M from Khosla, Accel, DST,…

Global Silicon Photonics/CPO Industry Ecosystem $GLW $TSM $NVDA $AVGO $MRVL

HW K9050 Logic Folding thus far use conservative layout where compute unit sit on top/main die produced w/ N+3 process & memory/IO unit are on lower die produced w/ N+2 process. This allows heat to be spread out rather than gather in 1…

Nomura: total CPO OE shipments are estimated at 2.0 million units in 2027F, scaling significantly to 22.5 million units in 2028F. $NVDA $AVGO $TSM $MRVL

GlobalFoundries reaches agreement to establish U.S.-based supply of silicon interposers for advanced AI packaging https://t.co/eFUvVNoZKf

Capital only buys the cleanroom. Specialized production equipment only gets a facility to the starting line. Commercial output depends on daily process discipline on the floor. Critical toolmakers remain the true operational foundation of…

Employees want more compensation and they have a right to it 💯 Great expert call from Damnang. Go check it out! $MU $DRAM $EWY. . https://t.co/6cDE6TekVu https://t.co/oa7gGZMne1

Expect some level of price increases for TSMC's advanced-node in 2027 with minimal pushback from enterprise clients. $TSM https://t.co/CreG6VBrfp

"Once China figures out advanced SME (by ~2028-31), its fully verticalized and highly integrated chip ecosystem will pose by far the greatest competitive threat to TSMC." https://t.co/ZTr6R8j3Kn

Bernstein: AI server supply chain $NVDA $AMD $AVGO $TSM $MU $AMKR $GOOGL

IBM’s Quantum Fab Anderon Aims to Be the TSMC Of Qubits With $2 Billion Bet https://t.co/TAPdPP5vhB

This.... https://t.co/svAX7knacq

Cadence and Samsung expanding the design ecosystem tackles chiplet architecture up front. In automotive physical AI, the bottleneck is downstream packaging: handling continuous thermal dissipation and rigorous AEC-Q100 stress testing…

Intel shipped the first commercial backside power delivery and its first GAA transistors in Panther Lake. SemiAnalysis's teardown puts 18A compute logic density on par with TSMC N3E, and not ahead of N3P, N2 or Samsung SF2 — while the…

BREAKING 🚨 Elon Musk just drew a hard line in the sand: Tesla and SpaceX are taking full command of the Terafab. Full stop. From blueprint to active production, his teams run the entire show. If an industry giant like TSMC gets involved,…

Here we go! Anthropic plans to adopt Samsung's 2nm process for its ASIC project to cut development cost and avoid capacity competition with TSMC. https://t.co/s5ugahAp2q

Lightmatter at HSBC's Silicon Valley tech tour indicated that the broader ecosystem is dealing with delayed CPO adoption, with expected timelines moving towards 2028-30. $NVDA $AMD $LITE $COHR $AAOI $TSM

Tower Semiconductor has the best SiPho foundry 👍 $TSEM $TSM $UMC $GFS $INTC https://t.co/mK82Hx1hDi

JPM on the potential TSMC collaboration with TSLA/SPCX 👀 https://t.co/1QFV0Zoejk

Expert Takeaways: TSMC > Monthly 2-nm capacity across Fab 20 and Fab 22 is on track to surpass 120k wafers by end-2027 (above the 110k consensus), with Fab-22 Phase-4 ramping faster than expected. > ~20% of TSMC's 7-nm capacity at Fab-15B…

Capital only buys the cleanroom. Advanced lithography and deposition tools get a fab to the starting line. Sustained wafer yield comes down to the process engineers on the floor managing drift. https://t.co/PZIPnPS6Yw

I can confirm what other media have reported regarding TSMC and Texas. Actually, I've heard about these discussion for a while. Various teams at TSMC have been tasked with evaluating a Texas project.

Mobilint's second-gen NPU Regulus slips six months to June 2027 after TSMC 12nm optimization with Asicland missed performance targets. The revenue target is halved to 15bn won from 30bn, while the Asicland contract was extended a year and…

Is EMIB really chipping away at CoWoS? CoWoS has owned advanced packaging for years. TSMC keeps adding capacity aggressively but it still isn't enough to satisfy customer demand. Intel stepped in with a compelling alternative with EMIB —…

Here is CXMT G5 compared to Big 3's D1b & D1c products. Note the feature size & cell size + bid density likely to slot in bw 1b & 1c. Consider that 1b DRAM wafers now fetch even higher ASP than TSMC N2. CXMT will be scaling up G4 & G5…

CoWoS vs. EMIB comparison 🔥 Reticle size limit - EMIB Cost - EMIB Yield - CoWoS Scale - CoWoS $TSM $INTC https://t.co/Rlwc9qwDlo https://t.co/A1NwlYPZAE

We took the silicon out of the silicon. Quick turn from SemiAnalysis STEEL Teardown Lab: iPhone 18 Pro Max, A20 silicon on TSMC N2. more to follow... https://t.co/RURlD49MP1

TSMC's foundry capacity shortage will signficantly worsen in 2027. TSMC's N3 expansion isn't nearly enough to bride the demand-supply gap - demand is already massive...and will grow through 2027. A major driver is Nvidia's accelerated Vera…

TSMC may double CoWoS capacity by 2028 on AI chip demand, per Communications Today. In the same week, Google is reportedly weighing CoWoS as a fallback for a TPU designed around Intel EMIB-T. A backup plan only works if the backup has…

Samsung plans to open its custom HBM base die to TSMC's foundry, per https://t.co/d0BzaGna0R. Single-outlet, unconfirmed. A vertically integrated memory maker buying logic wafers from its largest foundry rival would be a change of posture,…

AMD, TSMC and OSAT Channel Check: 1. AMD's AI GPU production (MI455/MI450) is expected to remain 100% at TSMC CoWoS in 2027. Production estimates land at a mid-point of 600k units for both SKUs next year per recent MS forecasts. 2. While…

Piper Sandler: All the big customers (i.e. Amazon, Apple, AMD, Google, Tesla, Microsoft, Nvidia, Qualcomm etc) are evaluating Intel's 14A process as companies look for alternative foundry/packaging suppliers to TSMC considering it is…

TSMC leading-edge allocation is where the real automotive compute timeline gets decided. Hyperscalers and handset vendors take the first tranches of 3nm and CoWoS packaging, which pushes automotive SDV controller roadmaps into later supply…

The semiconductor industry is not a market. It’s a stack of monopolies. Everyone talks about TSMC and NVIDIA. Almost nobody talks about the companies without which neither can ship an advanced chip. • ASML: lithography tools • Tokyo…

Jensen Huang just mapped out the chain of companies needed to power the AI revolution (Save this). His message is that AI is becoming a new industrial system that requires manufacturing, infrastructure, energy, financing and an enormous…

MediaTek launches new mobile chip using TSMC's most advanced technology MediaTek said the Dimensity 9600 Pro ​is its first mobile processor built using Taiwan Semiconductor Manufacturing Co's cutting-edge 2-nanometre process. ⁠It also…

A senior TSMC executive likened AI to a toddler with extraordinary powers that doesn’t know right from wrong, explaining why the company adopts a cautious approach to using the technology https://t.co/biv8gJM2nb

Per TrendForce, TSMC is reportedly targeting 22% more 2nm capacity and over 16% more 3nm by mid-2027, with CoWoS doubling by 2028. The leading edge expands by a fifth; packaging doubles. The pacing item is still what goes around the die,…

Kirin 9050 is a huge upgrade over 9030 Pro in GPU (2.2 TFLOPS vs 1.4 TFLOPS) See this in various gaming performance: Genshin Impact & 2 other games Mate XT2 vs XT vs Samsung Galaxy Z Trifold XT2 had the higher fps than Galaxy Z despite…

Morgan Stanley has CapEx of CSPs ~12% in 2028, with spending on AI semiconductors expected to increase even more. > CoWoS/CoPoS/EMIB-T capacity is estimated to grow by 50% in 2028, while infrastructure spending (such as factories and AI…

TSMC would take CoWoS from 130k wafers/month at end-2026 to 260k by end-2028, per Taiwanese media, expanding Arizona and AP7. Analysts put Intel's EMIB-T at 40k-45k CoWoS-equivalent in 2028; overflow already goes to Amkor, UMC and ASE.…

Samsung will fabricate a custom AI chip for OpenAI — likely the successor to Jalapeno, designed with Broadcom and fabbed by TSMC. The two will also co-develop memory for Stargate, where SK Hynix already sits. OpenAI is splitting its custom…

TSMC's CFO remains confident the company can maintain a gross margin of 56% in the long term. Also mentioned CoWoS capacity remains tight in capacity. $TSM https://t.co/oW1xsQoq9w

Two bets on HBM interconnect. TSMC, per ETNews, has asked suppliers for 5-micrometre-class microbump bonding and underfill — keep scaling bumps. Samsung is going the other way: bumpless hybrid copper bonding plus multi-wafer bonding for…

Please give @awodias a follow. They are at CIOE 2026 and providing a lot of useful information 👍 CPO market expected to grow from ~$90 million in 2025 to >$110 billion in 2031, or ~230% CAGR '25-'31. $TSM $LITE $COHR $AVGO $NVDA $AMD…

OpenAI's first inference ASIC, Jalapeno, is made at TSMC with Samsung HBM4. OpenAI Korea now says next-generation chips could also run at Samsung foundry and packaging. Samsung and Broadcom already have a $200B+ partnership through 2030.…