TSMC, the world's leading semiconductor manufacturing company, has reportedly started to accelerate research and development (R&D) of its next-generation 2 nm node. Having just recently announced that they will be starting production of a 5 nm process in Q4 of 2020, TSMC is pumping out nodes very fast and much faster compared to competition like Intel and Samsung. Having an R&D budget of almost 16 billion USD, TSMC seems to be spending the funds very wisely. The 5 nm node is going into volume production this year, and smaller nodes are already being prepared.

The 3 nm node is going into trial production in the first half of 2021, while the mass production is supposed to commence in 2022. As far as the 2 nm node, TSMC has recently purchased more expensive Extreme Ultra-Violet (EUV) lithography machines for the 2 nm node. Due to the high costs of these EUV machines, TSMC's capital spending will not be revisited this year and it should remain in the $16 billion range. As far as a timeline for 2 nm is concerned, we don't know when will TSMC start trial production as the node is still in development phases.

ASML Holding NV (ASML) today announced that it has completed system integration and testing of its first-generation HMI multibeam inspection (MBI) system for 5 nm nodes and beyond. The HMI eScan1000 demonstrated successful multibeam operation, simultaneously scanning nine beams on a number of test wafers. With nine beams, the eScan1000 will increase throughput up to 600% compared to single e-beam inspection tools for targeted in-line defect inspection applications.

The new MBI system includes an electron optics system capable of creating and controlling multiple primary electron beamlets and then collecting and processing the resulting secondary electron beams, limiting beam-to-beam crosstalk to less than 2% and delivering consistent imaging quality. It also features a high-speed stage to increase the system's overall throughput and a high-speed computational architecture to process the streams of data from the multiple beamlets in real time.

Samsung Electronics Co., Ltd., a world leader in advanced semiconductor technology, today announced plans to boost its foundry capacity at the company's new production line in Pyeongtaek, Korea, to meet growing global demand for cutting-edge extreme ultraviolet (EUV) solutions.

The new foundry line, which will focus on EUV-based 5 nanometer (nm) and below process technology, has just commenced construction this month and is expected to be in full operation in the second half of 2021. It will play a pivotal role as Samsung aims to expand the use of state-of-the-art process technologies across a myriad of current and next generation applications, including 5G, high-performance computing (HPC) and artificial intelligence (AI).

A DigiTimes premium report, interpreted by Chiakokhua, aka Retired Engineer, chronicling NVIDIA's move to contract TSMC for 7 nm and 5 nm EUV nodes for GPU manufacturing, made a startling revelation about NVIDIA's recent foundry diversification moves. Back in July 2019, a leading Korean publication confirmed NVIDIA's decision to contract Samsung for its next-generation GPU manufacturing. This was a week before AMD announced its first new-generation 7 nm products built for the TSMC N7 node, "Navi" and "Zen 2." The DigiTimes report reveals that NVIDIA underestimated the efficiency gains AMD would yield from TSMC N7.

With NVIDIA's bonhomie with Samsung underway, and Apple transitioning to TSMC N5, AMD moved in to quickly grab 7 nm-class foundry allocation and gained prominence with the Taiwanese foundry. The report also calls out a possible strategic error on NVIDIA's part. Upon realizing the efficiency gains AMD managed, NVIDIA decided to bet on TSMC again (apparently without withdrawing from its partnership with Samsung), only to find that AMD had secured a big chunk of its nodal allocation needed to support its growth in the x86 processor and discrete GPU markets. NVIDIA has hence decided to leapfrog AMD by adapting its next-generation graphics architectures to TSMC's EUV nodes, namely the N7+ and N5. The report also speaks of NVIDIA using its Samsung foundry allocation as a bargaining chip in price negotiations with TSMC, but with limited success as TSMC established its 7 nm-class industry leadership. As it stands now, NVIDIA may manufacture its 7 nm-class and 5 nm-class GPUs on both TSMC and Samsung.

TSMC has reportedly secured orders from NVIDIA for chips based on its 7 nm and 5 nm silicon fabrication nodes, sources tell DigiTimes. If true, it could confirm rumors of NVIDIA splitting its next-generation GPU manufacturing between TSMC and Samsung. The Korean semiconductor giant is commencing 5 nm EUV mass production within Q2-2020, and NVIDIA is expected to be one of its customers. NVIDIA is expected to shed light on its next-gen graphics architecture at the GTC 2020 online event held later this month. With its "Turing" architecture approaching six quarters of market presence, it's likely that the decks are being cleared for a new architecture not just in HPC/AI compute product segment, but also GeForce and Quadro consumer graphics cards. Splitting manufacturing between TSMC and Samsung would help NVIDIA disperse any yield issue arriving from either foundry's EUV node, and give it greater bargaining power with both.

A report via DigiTimes places TSMC as having announced to its investors that exploratory studies and R&D for the development of the 2 nm process node have commenced. As today's leading semiconductor fabrication company, TSMC doesn't seem to be one resting on its laurels. Their 7 nm process and derivatives have already achieved a 30% weight on the company's semiconductor orders, and their 5 nm node (which will include EUV litography) is set to hit HVM (High Volume Manufacturing) in Q2 of this year. Apart from that, not much more is known on 2 nm.

After 5 nm, which is expected to boats of an 84-87% transistor density gain over the current 7nm node, the plans are to go 3nm, with TSMC expecting that node to hit mass production come 2022. Interestingly, TSMC is planning to still use FinFET technology for its 3 nm manufacturing node, though in a new GAAFET (gate-all-around field-effect transistor) technology. TSMC's plans to deploy FinFET in under 5nm manufacturing is something that many industry analysts and specialist thought extremely difficult to achieve, with expectations for these sub-5nm nodes to require more exotic materials and transistor designs than TSMC's apparent plans

Samsung's 3 nm manufacturing has already given fruits to the company, with the South Korean giant already achieving risk production at the start of this year. The company previously projected volume production of their 3 nm process to start in early 2021. However, in a report via DigiTimes, this goal may have slipped to 2022 in wake of the coronavirus pandemic.

According to the news outlet, industry sources point this delay not to Samsung's fault in the manufacturing process, but to the entire logistics movement that has to be conducted in ramping up production of a new node. Impacts on logistics and transportation services are causing delays to deliveries of EUV and other critical production equipment, without which Samsung will be hard pressed to achieve its volume production goal. How this will ultimately affect Samsung's bottom line and revenue projections remains to be seen, but this won't do any favors to the company's high-density fabrication tech - especially if rival TSMC somehow manages to skirt these issues.

Samsung Electronics Co., Ltd., the world leader in advanced memory technology, today announced that it has successfully shipped one million of the industry's first 10 nm-class (D1x) DDR4 (Double Date Rate 4) DRAM modules based on extreme ultraviolet (EUV) technology. The new EUV-based DRAM modules have completed global customer evaluations, and will open the door to more cutting-edge EUV process nodes for use in premium PC, mobile, enterprise server and datacenter applications.

"With the production of our new EUV-based DRAM, we are demonstrating our full commitment toward providing revolutionary DRAM solutions in support of our global IT customers," said Jung-bae Lee, executive vice president of DRAM Product & Technology at Samsung Electronics. "This major advancement underscores how we will continue contributing to global IT innovation through timely development of leading-edge process technologies and next-generation memory products for the premium memory market."

TSMC will be doing good on their previous expectations for a H2 2020 ramp-up for high volume production (HVM) on their 5 nm manufacturing process. The new 5 nm fabrication process is an Extreme Ultraviolet lithography (EUV) one, with up to 14 layers being etchable onto the silicon wafers, as opposed to five and six, respectively, for TSMC's N7+ and N6 processes.

Volume production will start with Apple's A14 SoC, meant to be driving next-generation iPhones that should hit shelves by September this year (should the COVID-19 pandemic let it be so). Apple is using two thirds of TSMC's capacity for 5 nm as is with this SoC; it's currently unclear which client (or clients) are getting the leftover one third capacity. TSMC announced back in December that they were seeing yields upwards of 80% in 5 nm EUV fabrication, so now it's "just" a matter of monetizing the process until their 3 nm iteration comes online, expectedly, in 2022.

With its 7 nm RDNA architecture that debuted in July 2019, AMD achieved a nearly 50% gain in performance/Watt over the previous "Vega" architecture. At its 2020 Financial Analyst Day event, AMD made a big disclosure: that its upcoming RDNA2 architecture will offer a similar 50% performance/Watt jump over RDNA. The new RDNA2 graphics architecture is expected to leverage 7 nm+ (7 nm EUV), which offers up to 18% transistor-density increase over 7 nm DUV, among other process-level improvements. AMD could tap into this to increase price-performance by serving up more compute units at existing price-points, running at higher clock speeds.

AMD has two key design goals with RDNA2 that helps it close the feature-set gap with NVIDIA: real-time ray-tracing, and variable-rate shading, both of which have been standardized by Microsoft under DirectX 12 DXR and VRS APIs. AMD announced that RDNA2 will feature dedicated ray-tracing hardware on die. On the software side, the hardware will leverage industry-standard DXR 1.1 API. The company is supplying RDNA2 to next-generation game console manufacturers such as Sony and Microsoft, so it's highly likely that AMD's approach to standardized ray-tracing will have more takers than NVIDIA's RTX ecosystem that tops up DXR feature-sets with its own RTX feature-set.

AMD at its 2020 Financial Analyst Day event unveiled its upcoming CDNA GPU-based compute accelerator architecture. CDNA will complement the company's graphics-oriented RDNA architecture. While RDNA powers the company's Radeon Pro and Radeon RX client- and enterprise graphics products, CDNA will power compute accelerators such as Radeon Instinct, etc. AMD is having to fork its graphics IP to RDNA and CDNA due to what it described as market-based product differentiation.

Data centers and HPCs using Radeon Instinct accelerators have no use for the GPU's actual graphics rendering capabilities. And so, at a silicon level, AMD is removing the raster graphics hardware, the display and multimedia engines, and other associated components that otherwise take up significant amounts of die area. In their place, AMD is adding fixed-function tensor compute hardware, similar to the tensor cores on certain NVIDIA GPUs.

UNISOC, a leading global supplier of mobile communication and IoT chipsets, today officially launched its new-generation 5G SoC mobile platform - T7520. Using cutting-edge process technology, T7520 enables an optimized 5G experience with substantially enhanced AI computing and multimedia imaging processing capabilities while lowering power consumption.

T7520 is UNISOC's second-generation 5G smartphone platform. Built on a 6 nm EUV process technology and empowered by some of the latest design techniques, it offers substantially enhanced performance at a lower level of power consumption than ever.

Samsung Electronics, a world leader in advanced semiconductor technology, today announced that its new cutting-edge semiconductor fabrication line in Hwaseong, Korea, has begun mass production.

The facility, V1, is Samsung's first semiconductor production line dedicated to the extreme ultraviolet (EUV) lithography technology and produces chips using process node of 7 nanometer (nm) and below. The V1 line broke ground in February 2018, and began test wafer production in the second half of 2019. Its first products will be delivered to customers in the first quarter.

European Processor Initiative (EPI) is a Europe's project to kickstart a homegrown development of custom processors tailored towards different usage models that the European Union might need. The first task of EPI is to create a custom processor for high-performance computing applications like machine learning, and the chip prototypes are already on their way. The EPI chairman of the board Jean-Marc Denis recently spoke to the Next Platform and confirmed some information regarding the processor design goals and the timeframe of launch.

Supposed to be manufactured on TSMC's 6 nm EUV (TSMC N6 EUV) technology, the EPI processor will tape-out at the end of 2020 or the beginning of 2021, and it is going to be heterogeneous. That means that on its 2.5D die, many different IPs will be present. The processor will use a custom ARM CPU, based on a "Zeus" iteration of Neoverese server core, meant for general-purpose computation tasks like running the OS. When it comes to the special-purpose chips, EPI will incorporate a chip named Titan - a RISC-V based processor that uses vector and tensor processing units to compute AI tasks. The Titan will use every new standard for AI processing, including FP32, FP64, INT8, and bfloat16. The system will use HBM memory allocated to the Titan processor, have DDR5 links for the CPU, and feature PCIe 5.0 for the inner connection.

In a 2019 earnings call ASML, a Dutch company that is currently the world's largest semiconductor lithography supplier has been talking about the company's records and what awaits them in the future. In its 2019 earning report, ASML was forecasting the delivery of as much as 35 EUV systems in 2020. It is not a forecast per se, but rather a known fact since factories order their equipment months before they need to use it. Having previously delivered 26 EUV systems in 2019, the plan for the coming years is to boost the EUV system shipments by as much as 40% yearly. With plans to ship between 45 and 50 EUV systems in the year 2021, AMSL sees a strong revenue gain in the coming years. What is driving the demand for these machines is the use of ever-smaller semiconductor manufacturing nodes. Even at 7 nm, there is almost a need to use EUV lithography, and as you drop in size the lithography challenge becomes real, the use of EUV becomes a necessity.

At a Q&A session with the tech press in Las Vegas, AMD CEO Dr Lisa Su raised hopes of a high-end graphics card based on its "Navi" family of GPUs. Responding to a specific question by Gordon Ung from PC World on whether there will be a high-end competitor in the discrete graphics space, Dr Su stated that one should expect a "high-end Navi." Dr Su states: "I know those on Reddit want a high end Navi! You should expect that we will have a high-end Navi, and that it is important to have it. The discrete graphics market, especially at the high end, is very important to us. So you should expect that we will have a high-end Navi, although I don't usually comment on unannounced products."

For months now, it's been speculated that AMD has been working on a larger GPU die than "Navi 10." In 2020, AMD is expected to release the "Navi 20" familly of GPUs built on 7 nm+ (EUV) node, based on the RDNA2 graphics architecture. The key design goals of RDNA2 are expected to be support for at least tier-1 variable-rate shading (VRS), and possibly hardware-accelerated ray-tracing. It's possible that "high-end Navi" belongs to this family of GPUs.

AMD in the second half of 2020 could outpace Apple as the biggest foundry customer of TSMC for its 7 nm silicon fabrication nodes (DUV and EUV combined). There are two key factors contributing to this: AMD significantly increasing its orders for the year; and Apple transitioning to TSMC's 5 nm node for its A14 SoC, freeing up some 7 nm allocation, which AMD grabbed. AMD is currently tapping into 7 nm DUV for its "Zen 2" chiplet, "Navi 10," and "Navi 14" GPU dies. The company could continue to order 7 nm DUV until these products reach EOL; while also introducing the new "Renoir" APU die on the process. The foundry's new 7 nm+ (EUV) node will be utilized for "Zen 3" chiplets and "Navi 2#" GPU dies in 2020.

Currently, the top-5 customers for TSMC 7 nm are Apple, HiSilicon, Qualcomm, AMD, and MediaTek. Barring AMD, the others in the top-5 build mobile SoCs or 4G/5G modem chips on the node. AMD is expected to top the list as it scales up orders with TSMC. In the first half of 2020, TSMC's monthly output for 7 nm is expected to grow to 110,000 wafers per month (wpm). Apple's migration to 5 nm in 2H-2020, coupled with capacity-addition could take TSMC's 7 nm output to 140,000 wpm. AMD has reportedly booked the entire capacity-addition for 30,000 wpm, taking its allocation up to 21% in 2H-2020. Qualcomm is switching to Samsung for its next-generation SoCs and modems designed for 7 nm EUV. NVIDIA, too, is expected to built its next-gen 7 nm EUV GPUs on Samsung instead of TSMC. These moves by big players could free up significant foundry allocation at TSMC for AMD's volumes to grow in 2020.

A prominent Taiwanese newspaper reported that AMD will formally unveil its next-generation "Zen 3" CPU microarchitecture at the 2020 International CES. Company CEO Dr Lisa Su will head an address revealing three key client-segment products under the new 4th generation Ryzen processor family, and the company's 3rd generation EPYC enterprise processor family based on the "Milan" MCM that succeeds "Rome." AMD is keen on developing an HEDT version of "Milan" for the 4th generation Ryzen Threadripper family, codenamed "Genesis Peak."

The bulk of the client-segment will be addressed by two distinct developments, "Vermeer" and "Renoir." The "Vermeer" processor is a client-desktop MCM that succeeds "Matisse," and will implement "Zen 3" chiplets. "Renoir," on the other hand, is expected to be a monolithic APU that combines "Zen 2" CPU cores with an iGPU based on the "Vega" graphics architecture, with updated display- and multimedia-engines from "Navi." The common thread between "Milan," "Genesis Peak," and "Vermeer" is the "Zen 3" chiplet, which AMD will build on the new 7 nm EUV silicon fabrication process at TSMC. AMD stated that "Zen 3" will have IPC increases in line with a new microarchitecture.

AMD's upcoming Ryzen 4000 series processors will be based on the company's Zen 3 design, which will feature a deeply revised architecture aiming to offer increased performance (surprising no-one). AMD themselves have already said that Zen 3 will offer performance increases in line with the release of new architectures - and we all remember the around 15% increase achieved with the release of Zen 2 Ryzen 3000 series, which surprised even AMD on its performance capabilities. Several sources around the web are quoting an around 17% increase in performance, taking into account increased operating frequencies of Zen 3 (100 to 200 MHz at least for the enterprise solutions, which could pave the way for even higher increases in consumer-geared products) and increased IPC of its core design. The utilization of EUV in the 7 nm process shouldn't have much to do with the increased frequencies of the CPUs, and will mostly be used to reduce the number of masks that are required for production of AMD's Zen 3 CPUs (which in turn will lead to increased yields).

Sources are claiming an increase of up to 50% in Zen 3's Floating Point Units (FPU) compared to Zen 2, while integer operations should make do with a 10-12% increase. Cores should remain stable across the board - and with that increase in performance, I'd say an upper limit of 16 physical and 32 logic cores in a consumer-geared CPU is more than enough. Increased IPCs and frequencies will definitely make AMD an even better proposition for all markets - gaming in particular, where Intel still has a (slightly virtual) hold in consumer's minds.

At its recent SC19 talk, AMD touched upon its upcoming "Zen 3" CPU microarchitecture. Designed for the 7 nm EUV silicon fabrication process that significantly increases transistor densities, "Zen 3" could post performance gains "right in line with what you would expect from an entirely new architecture," states AMD, referring to the roughly 15 percent IPC gains that were expected of "Zen 2" prior to its launch. "Zen 2" IPC ended up slightly over 15 percent higher than that of the original "Zen" microarchitecture. AMD's SC19 comments need not be a guidance on the IPC itself, but rather performance gains of end-products versus their predecessors.

The 7 nm EUV process, with its 20 percent transistor-density increase could give AMD designers significant headroom to increase clock speeds to meet the company's generational performance improvement targets. Another direction in which "Zen 3" could go is utilizing the additional transistor density to bolster its core components to support demanding instruction-sets such as AVX-512. The company's microarchitecture is also missing something analogous to Intel's DLBoost, an instruction-set that leverages fixed-function hardware to accelerate AI-DNN building and training. Even VIA announced an x86 microarchitecture with AI hardware and AVX-512 support. In either case, the design of "Zen 3" is complete. We'll have to wait until 2020 to find out how fast "Zen 3" is, and the route taken to get there.

TSMC has been very aggressive with its approach to silicon manufacturing, with more investments into its R&D that now match or beat the capex investments of Intel. That indicates a strong demand for new technologies and TSMC's strong will not drop out of the never-ending race for more performance and smaller node sizes.

According to the sources over at DigiTimes, TSMC has acquired as much as 30 hectares of land in the Southern Taiwan Science Park to begin the construction of its fabs that are supposed to start high-volume manufacturing 3 nm node in 2023. Construction of 3 nm manufacturing facilities are set to begin in 2020 when TSMC will lay the groundwork for the new fab. The 3 nm semiconductor node is expected to be TSMC's third attempt at EUV lithography, right after the 7 nm+, and 5 nm nodes which are also based on EUV technology.

In a shocking piece of news, Intel has reportedly scrapped plans to launch its 10 nm "Ice Lake" microarchitecture on the client desktop platform. The company will confine its 10 nm microarchitectures, "Ice Lake" and "Tiger Lake" to only the mobile platform, while the desktop platform will see derivatives of "Skylake" hold Intel's fort under the year 2022! Intel gambles that with HyperThreading enabled across the board and increased clock-speeds, it can restore competitiveness with AMD's 7 nm "Zen 2" Ryzen processors with its "Comet Lake" silicon that offers core-counts of up to 10.

"Comet Lake" will be succeeded in 2021 by the 14 nm "Rocket Lake" silicon, which somehow combines a Gen12 iGPU with "Skylake" derived CPU cores, and possibly increased core-counts and clock speeds over "Comet Lake." It's only 2022 that Intel will ship out a truly new microarchitecture on the desktop platform, with "Meteor Lake." This chip will be built on Intel's swanky 7 nm EUV silicon fabrication node, and possibly integrate CPU cores more advanced than even "Willow Cove," possibly "Golden Cove."

TSMC today announced that its seven-nanometer plus (N7+), the industry's first commercially available Extreme Ultraviolet (EUV) lithography technology, is delivering customer products to market in high volume. The N7+ process with EUV technology is built on TSMC's successful 7 nm node and paves the way for 6 nm and more advanced technologies.

The N7+ volume production is one of the fastest on record. N7+, which began volume production in the second quarter of 2019, is matching yields similar to the original N7 process that has been in volume production for more than one year.

With its next-generation "Zen 3" CPU microarchitecture designed for the 7 nm EUV silicon fabrication process, AMD could bid the "Zen" compute complex or CCX farewell, heralding chiplets with monolithic last-level caches (L3 caches) that are shared across all cores on the chiplet. AMD embraced a quad-core compute complex approach to building multi-core processors with "Zen." At the time, the 8-core "Zeppelin" die featured two CCX with four cores, each. With "Zen 2," AMD reduced the CPU chiplet to only containing CPU cores, L3 cache, and an Infinity Fabric interface, talking to an I/O controller die elsewhere on the processor package. This reduces the economic or technical utility in retaining the CCX topology, which limits the amount of L3 cache individual cores can access.

This and more juicy details about "Zen 3" were put out by a leaked (later deleted) technical presentation by company CTO Mark Papermaster. On the EPYC side of things, AMD's design efforts will be spearheaded by the "Milan" multi-chip module, featuring up to 64 cores spread across eight 8-core chiplets. Papermaster talked about how the individual chiplets will feature "unified" 32 MB of last-level cache, which means a deprecation of the CCX topology. He also detailed an updated SMT implementation that doubles the number of logical processors per physical core. The I/O interface of "Milan" will retain PCI-Express gen 4.0 and eight-channel DDR4 memory interface.

"Moore's Law" is a term coined in 1965 by Gordon Moore, who presented a paper which predicts that semiconductor scaling will allow integrated circuits to feature twice as many transistors present per same area as opposed to a chip manufactured two years ago. That means we could get same performance at half the power than the previous chip, or double the performance at same power/price in only two years time. Today we'll investigate if Moore's Law stayed true to its cause over the years and how much longer can it keep going.