United Microelectronics Corporation ("UMC" or "The Company"), a leading global semiconductor foundry, today announced its consolidated operating results for the fourth quarter of 2022. Fourth quarter consolidated revenue was NT$67.84 billion, decreasing 10.0% QoQ from NT$75.39 billion in 3Q22. Compared to a year ago, 4Q22 revenue grew 14.8% YoY from NT$59.10 billion in 4Q21. Consolidated gross margin for 4Q22 was 42.9%. Net income attributable to the shareholders of the parent was NT$19.1 billion, with earnings per ordinary share of NT$1.54.

Jason Wang, co-president of UMC, said, "In the fourth quarter, due to a significant slowdown across most of our end markets and inventory correction in the semiconductor industry, our wafer shipments fell 14.8% QoQ while overall fab utilization rate dropped to 90%. Average selling price increased slightly during the quarter as a result of our ongoing product mix optimization efforts, moderating the decline in revenue."

Cerebras Systems, the pioneer in accelerating artificial intelligence (AI) compute, today unveiled Andromeda, a 13.5 million core AI supercomputer, now available and being used for commercial and academic work. Built with a cluster of 16 Cerebras CS-2 systems and leveraging Cerebras MemoryX and SwarmX technologies, Andromeda delivers more than 1 Exaflop of AI compute and 120 Petaflops of dense compute at 16-bit half precision. It is the only AI supercomputer to ever demonstrate near-perfect linear scaling on large language model workloads relying on simple data parallelism alone.

With more than 13.5 million AI-optimized compute cores and fed by 18,176 3rd Gen AMD EPYC processors, Andromeda features more cores than 1,953 Nvidia A100 GPUs and 1.6 times as many cores as the largest supercomputer in the world, Frontier, which has 8.7 million cores. Unlike any known GPU-based cluster, Andromeda delivers near-perfect scaling via simple data parallelism across GPT-class large language models, including GPT-3, GPT-J and GPT-NeoX.

A major unnamed customer of TSMC has reportedly cut their order for 3 nm wafers. Foundry customers usually place orders for cutting-edge foundry nodes several quarters in advance, in exchange for priority foundry allocations, and preferential rates, while foundries use revenues from these orders to develop the capacity for manufacture these chips. The 3 nm customer could be anyone—Qualcomm, Mediatek, NVIDIA, AMD, or Intel. Order cancellations have reportedly had a domino-effect on the upstream supply-chain of TSMC, hitting suppliers of raw materials, manufacturing equipment, and other consumables. There is an industry-wide slump in demand for consumer electronics and PC hardware, which reflects in the slump in revenues and/or guidance in quarterly financial results releases by prominent companies.

Based on a report by Tom's Hardware, AMD's CEO Lisa Su is planning a trip to Taiwan in the next couple of months. It is said that she is planning to meet with multiple partners in Taiwan, such as ASUS, Acer and maybe more importantly, ASMedia, which will be the sole maker of chipsets for AMD, once the X570 chipset is discontinued. AMD is apparently also seeing various less well known partners that deliver parts for its CPUs, such as Nan Ya PCB, Unimicron Technologies and Kinsus Interconnects.

However, it appears that the main reason for Lisa Su herself to visit Taiwan will be to meet with TSMC, to discuss future collaboration with CC Wei, TSMC's chief executive. This is so AMD can secure enough wafer allocation on future nodes, such as its 3 nm and 2 nm class nodes. The move to these nodes is obviously not happening in the near future for AMD, but considering that TSMC is currently the leading foundry and is operating at capacity, it makes sense to get in early, as the competition is stiff when it comes to getting wafer allocation on cutting edge nodes. It's unclear which exact 3 nm class node AMD will be aiming for, but it might be the N3P node, which is said to kick off production sometime next year. Lisa Su is also said to have meetings with TSMC, SPIL and Ase Technology when it comes to advanced packaging for AMD's products. This includes technologies such as chip-on-wafer-on-substrate (CoWoS) and fan-out embedded bridge (FO-EB), with AMD already being expected to use some of these technologies in its upcoming Navi 3x GPUs.

Andreas Schilling with Hardwareluxx.de, as part of the Intel Tech Tour Israel, got to hold a 12-inch wafer full of "Raptor Lake-S" dies. These are dies in their full 8P+16E configuration. The die is estimated to measure 257 mm² in area. We count 231 full dies on this wafer. Intel is building "Raptor Lake" on the same 10 nm Enhanced SuperFin (aka Intel 7) node as "Alder Lake." The die is about 23% larger than "Alder Lake" on account of two additional E-core clusters, possibly larger P-cores, and larger L2 caches for both the P-core and E-core clusters. "Raptor Lake" gains significance as it will be the last client processor from Intel to be built on a monolithic die of a uniform silicon fabrication node. Future generations are expected to take the chiplets route, realizing the company's IDM 2.0 product development strategy.

According to the latest TrendForce investigations, moving into the second half of 3Q22, the lack of a peak season has led to a delay in inventory destocking. Transactions in the NAND Flash market have been frosty. Buyers are watching passively and tend not to negotiate pricing. Pressure on factory inventory has reached a breaking point and manufacturers are bottoming out pricing in order to make a deal. This move will lead to a further decline in manufacturer pricing. TrendForce once again revises downward 3Q22 NAND Flash wafer contract prices and the decline of pricing is estimated to balloon to 30-35% from the original estimate of 15-20%.

In the past two years, the pandemic has promoted digital transformation and notebook computers and servers have stimulated rapid growth in NAND Flash consumption. In order to satisfy demand, manufacturers have been expanding aggressively, with their processes accelerating the output of 128-layer+ products. However, the 2H22 NAND Flash market situation has deteriorated sharply with the acute correction in purchase order demand for smartphones and laptops indicative of a market oversupply. Looking forward to 2023, the conservative attitudes of various consumer electronics brands may lead to difficulties in improving market conditions in the next year and stimulate suppliers to step up efforts to seize market share.

Hsinchu, Taiwan-based global semiconductor silicon wafer company, GlobalWafers, announced today that it plans to build a state-of-the-art 300-millimeter silicon wafer factory in Sherman, Texas, which is the first of its kind in the USA over twenty (20) years. Construction is expected to commence later this year. This 300-millimeter greenfield investment is consistent with the Company's announcement on February 6th of this year of brownfield and greenfield expansions totaling NTD 100 bn. This new Texas investment could also support as many as 1,500 jobs with production volumes ultimately reaching 1.2 mn wafers per month after multiple stages of equipment installation, in alignment with market demand.

300-millimeter silicon wafers are the starting material for all advanced semiconductor fabrication sites (or fabs), including recently announced United States (US) expansions by GlobalFoundries, Intel, Samsung, Texas Instruments and TSMC. Most of these wafers are currently manufactured in Asia, forcing the US semiconductor industry to highly rely on imported silicon wafers. This investment will represent the first new silicon wafer facility in the US in over two decades and close a critical semiconductor supply chain gap.

Cerebras Systems, the pioneer in high performance artificial intelligence (AI) computing, today announced, for the first time ever, the ability to train models with up to 20 billion parameters on a single CS-2 system - a feat not possible on any other single device. By enabling a single CS-2 to train these models, Cerebras reduces the system engineering time necessary to run large natural language processing (NLP) models from months to minutes. It also eliminates one of the most painful aspects of NLP—namely the partitioning of the model across hundreds or thousands of small graphics processing units (GPU).

"In NLP, bigger models are shown to be more accurate. But traditionally, only a very select few companies had the resources and expertise necessary to do the painstaking work of breaking up these large models and spreading them across hundreds or thousands of graphics processing units," said Andrew Feldman, CEO and Co-Founder of Cerebras Systems. "As a result, only very few companies could train large NLP models - it was too expensive, time-consuming and inaccessible for the rest of the industry. Today we are proud to democratize access to GPT-3 1.3B, GPT-J 6B, GPT-3 13B and GPT-NeoX 20B, enabling the entire AI ecosystem to set up large models in minutes and train them on a single CS-2."

According to TrendForce research, although demand for consumer electronics remains weak, structural growth demand in the semiconductor industry including for servers, high-performance computing, automotive, and industrial equipment has not flagged, becoming a key driver for medium and long term foundry growth. At the same time, due to robust wafer production at higher pricing in 1Q22, quarterly output value hit a new high for the 11th consecutive quarter, reaching US$31.96 billion, 8.2% QoQ, marginally less than the previous quarter. In terms of ranking, the biggest change is Nexchip surpassed Tower at the ninth position.

TSMC's across the board wafer hikes in 4Q21 on batches primarily produced in 1Q22 coupled with sustained strong demand for high-performance computing and better foreign currency exchange rates pushed TSMC's 1Q22 revenue to $17.53 billion, up 11.3% QoQ. Quarterly revenue growth by node was generally around 10% and the 7/6 nm and 16/12 nm processes posted the highest growth rate due to small expansions in production. The only instance of revenue decline came at the 5/4 nm process due to Apple's iPhone 13 entering the off season for production stocking.

According to TrendForce research, looking at NAND Flash wafers, the pricing of which more sensitively reflects the market, suppliers are increasingly motivated to cut prices in exchange for sales due to weak retail demand since March and a more conservative outlook for shipments of other end products. The price of NAND Flash wafers is expected to begin falling in May and the supply of NAND Flash will gradually overtake demand in 2H22. The price decline of NAND Flash wafers in 3Q22 may reach 5~10%.

At the same time, TrendForce indicates that February's contamination incident at Kioxia was expected to tighten the market in 2Q22 and 3Q22. However, as a consequence of rising inflation and the war between Russia and Ukraine, market demand for consumer products in the traditional peak season of the second half of the year is trending conservative and the prices of client SSD, eMMC, and UFS in 3Q22 will be flat compared to 2Q22, breaking from the original expectation that prices may rise. In terms of enterprise SSDs, as demand for data centers remains strong, no significant correction in demand has yet been observed. However, as the overall NAND Flash market gradually moves into oversupply, prices will only grow slightly by approximately 0~5% in 3Q22.

United Microelectronics Corporation ("UMC" or "The Company"), a leading global semiconductor foundry, today announced that its Board of Directors has approved a plan to build a new advanced manufacturing facility next to its existing 300 mm fab (Fab12i) in Singapore. The first phase of this greenfield fab will have a monthly capacity of 30,000 wafers with production expected to commence in late 2024.

The new fab (Fab12i P3) will be one of the most advanced semiconductor foundries in Singapore, providing UMC's 22/28 nm processes. The planned investment for this project will be US$5 billion. UMC has operated as a pure-play foundry supplier in Singapore for more than 20 years and the location is also the company's designated R&D center for advanced specialty technologies. To account for the Fab12i expansion, the company's 2022 capex budget will be revised upward to US$3.6 billion.

From 2020 to 2025, the compound annual growth rate (CAGR) of 12-inch equivalent wafer capacity at the world's top ten foundries will be approximately 10% with the majority of these companies focusing on 12-inch capacity expansion, which will see a CAGR of approximately 13.2%, according to TrendForce's research. In terms of 8-inch wafers, due to factors such as difficult to obtain equipment and whether capacity expansion is cost-effective, most fabs can only expand production slightly by means of capacity optimization, equating to a CAGR of only 3.3%. In terms of demand, the products primarily derived from 8-inch wafers, PMIC and Power Discrete, are driven by demand for electric vehicles, 5G smartphones, and servers. Stocking momentum has not fallen off, resulting in a serious shortage of 8-inch wafer production capacity that has festered since 2H19. Therefore, in order to mitigate competition for 8-inch capacity, a trend of shifting certain products to 12-inch production has gradually emerged. However, if shortages in overall 8-inch capacity is to be effectively alleviated, it is still necessary to wait for a large number of mainstream products to migrate to 12-inch production. The timeframe for this migration is estimated to be close to 2H23 into 2024.

Toshiba Electronic Devices & Storage Corporation ("Toshiba") today announced that it will construct a new 300-milimeter wafer fabrication facility for power semiconductors at its main discrete semiconductor production base, Kaga Toshiba Electronics Corporation, in Ishikawa Prefecture. Construction will take place in two phases, allowing the pace of investment to be optimized against market trends, with the production start of Phase 1 scheduled for within fiscal 2024. When Phase 1 reaches full capacity, Toshiba's power semiconductor production capacity will be 2.5 times that of fiscal 2021.

Power devices are essential components for managing and reducing power consumption in every kind of electronic equipment, and for achieving a carbon neutral society. Current demand is expanding on vehicle electrification and the automation of industrial equipment, with very strong demand for low-voltage MOSFETs (metal oxide semiconductor field effect transistors) and IGBTs (insulated-gate bipolar transistors) and other devices. To date, Toshiba has met this demand growth by increasing production capacity on 200-milimeter lines, and expediting the start of production on 300-milimeter production lines from the first half of fiscal 2023 to the second half of fiscal 2022. Decisions on the new fab's overall capacity and equipment investment, the start of production, production capacity and production plan will reflect market trends.

In a surprising outcome, Taiwanese GlobalWafers purchase of German Siltronic has failed, due to the German Ministry for Economic Affairs having failed to approve the deal. The €4.35 billion deal had been approved by all other regulators globally, but the German Ministry for Economic Affairs claims that they didn't have enough time to go over the review process locally by the deadline that was set. It appears that they're blaming their counterparts in the PRC for having been too slow in their approval process, as the Germans wanted to go over the Chinese approval, before committing to their own approval.

The deal might not be dead in the water though, but it looks like GlobalWafers is going to have to make a new offer and start the entire process over again. Had the deal gone through, GlobalWafers would've become the world's second largest 300 mm silicon wafer producer, behind Japanese Shin-Etsu, overtaking Sumco, which is also based out of Japan. GlobalWafers already owns a majority stake in Siltronic, but the question now is if GlobalWafers will maintain that share, or look to invest elsewhere. One option on the table for GlobalWafers is apparently the US, but at the same time, the majority shareholder of Siltronic wants to sell, ideally to GlobalWafers. Although it's only speculation at this point, it has been suggested that the German government wanted to retain Siltronic as a local company, especially due to growing interest by the EU to expand foundry type businesses in Europe. GlobalWafers is said to make an announcement by the end of this week as to whether they will continue with the Siltronic purchase or not.

AMD in its 8-K filing with the SEC, disclosed that it has updated its wafer supply agreement (WSA) with GlobalFoundries. Under the latest agreement, AMD commits to buy $2.1 billion worth wafers from GlobalFoundries between 2022 and 2025. The previous version of the WSA saw commitments up to 2024, and wafers worth $1.6 billion. The update hence adds another year and $500 million worth supply.

AMD currently sources 12 nm and 14 nm wafers from GlobalFoundries, which go into making cIOD and sIOD components in its processors, and motherboard chipsets. The move to extend the WSA indicates that the company may continue to use 12 nm-class I/O dies in its processors for the foreseeable future. It will be very interesting to see if 12 nm-class I/O dies make it to next-generation products such as "Genoa" and "Rapael," which integrate the latest IP blocks such as PCI-Express Gen 5 root-complexes, DDR5 memory controllers, and 3rd Gen Infinity Fabric. Processors with 12 nm I/O dies, such as "Milan" and "Vermeer" could be retired only by 2023-24, as AMD will use 2022 to spread across its next-gen product launches.

The world's third largest supplier of silicon wafers—GlobalWafers—has told its customers that it's unlikely that it will be able to fulfill all of their orders until at least 2024, despite trying its best to improve production efficiency. The company is currently delivering around 90 percent of its customer orders and is doing its best to meet customer demand, while at the same time investing heavily in increasing production capacity of 12-inch wafers.

GlobalWafers has orders lined up along the tune of US$3.6 billion and many of its customers have increased their supply agreements from five to eight years. Like many other industries, the silicon wafer manufacturers are having supply problems with raw materials and have seen a price increase for many of the materials used in the production of silicon wafers. That said, global shipments of silicon wafers are said to see a growth of over six percent next year and should continue to increase at a slightly lower rate the following two years. It would appear the shortage of various ICs will continue, maybe even longer than expected, unless we see an industry wide change into what is being produced and there's a focus on fewer products, which might in turn allow for at least some of the backlogs to be cleared.

Intel Labs recently opened the Intel Research Center for Integrated Photonics for Data Center Interconnects. The center's mission is to accelerate optical input/output (I/O) technology innovation in performance scaling and integration with a specific focus on photonics technology and devices, CMOS circuits and link architecture, and package integration and fiber coupling.

"At Intel Labs, we're strong believers that no one organization can successfully turn all the requisite innovations into research reality. By collaborating with some of the top scientific minds from across the United States, Intel is opening the doors for the advancement of integrated photonics for the next generation of compute interconnect. We look forward to working closely with these researchers to explore how we can overcome impending performance barriers," said James Jaussi, senior principal engineer and director of the PHY Research Lab in Intel Labs.

GLOBALFOUNDRIES (GF ), the global leader in feature-rich semiconductor manufacturing, today announced it is expanding its global manufacturing footprint with the construction of a new fab on its Singapore campus. In partnership with the Singapore Economic Development Board and with co-investments from committed customers, GF's more than US $4B (S$5B) investment will play an integral role in meeting the growing demand for the company's industry-leading manufacturing technologies and services to enable companies worldwide to develop and scale their business.

In a virtual groundbreaking ceremony, Singapore Minister for Transport and Minister-in-charge of Trade Relations S. Iswaran and Mubadala Investment Company Managing Director and Group CEO H.E. Khaldoon Khalifa Al Mubarak, were joined by: UAE Ambassador to Singapore H.E. Jamal Abdulla Al Suwaidi; Singapore Ambassador to the UAE H.E. Kamal R Vaswani; Singapore Economic Development Board Managing Director Chng Kai Fong; GF Board Chairman Ahmed Yahia Al Idrissi; along with GF executives including CEO Tom Caulfield; CFO David Reeder; SVP and Head of Global Operations KC Ang; SVP of Global Sales Juan Cordovez; VP of Human Resource for APAC and International Fabs Janice Lee; and VP of Technology Development in Singapore Dr. Soh Yun Siah.

GLOBALFOUNDRIES (GF ), the global leader in feature-rich semiconductor manufacturing, and GlobalWafers Co., Ltd. (GWC), one of the top silicon wafer manufacturers in the world, today announced an $800 million agreement to add 300 mm silicon-on-insulator (SOI) wafer manufacturing and expand existing 200 mm SOI wafer production at GWC's MEMC facility in O'Fallon, Missouri.

The silicon wafers produced by GWC are key input materials for semiconductors and an integral part of GF's supply chain. The wafers are used in GF's multi-billion dollar manufacturing facilities, or fabs, where they are used to manufacture the computer chips that are pervasive and vital to the global economy. Today's announcement expands GF's domestic silicon wafer supply from the United States.

Korean DRAM and NAND flash giant, SK Hynix, admitted that a rather big batch of its DRAM wafers is defective and in circulation. The size of this defective batch is rumored to be 240,000 wafers according to a Yonhap report, although the company downplays the scale of the problem citing its monthly production output of 1.8 million wafers.

The company said that it is working with its customers who received these wafers, for recall and replacement. "We're currently talking to a limited number of customers affected by this to address the issue. While it's too early to estimate the potential losses, we don't think they would be that significant as the defect is within the range of typical quality issue check." Besides this, the company is battling rumors surrounding the scale of defective DRAM wafers by the company, in circulation. "The scale of the potential losses mentioned in the rumor is absolutely not true and exaggerated," the company said, in a statement to The Register.

AMD in a filing with the U.S. Securities and Exchange Commission (SEC), revealed that its wafer supply agreement with GlobalFoundries has been amended. Under the new terms, AMD places orders for wafers from GlobalFoundries up to 2024, with purchase targets set for each year leading up to 2024. Beyond meeting these targets, AMD is free from all other exclusivity commitments. The agreement was previously amended in January 2019, setting annual purchase targets for 2019, 2020, and 2021, while beginning a de-coupling between AMD and GlobalFoundries. This enabled the company to source 7 nm (or smaller) chips, such as CCDs and GPUs, from other foundries, such as TSMC, while keeping GlobalFoundries exclusive for 12 nm (or larger) nodes.

The updated wafer supply agreement unlocks many possibilities for AMD. For starters, it can finally build a next-generation sIOD (server I/O die) on a more efficient node than GlobalFoundries 12LP, such as TSMC 7 nm. This transition to 7 nm will be needed as the next-gen "Genoa" EPYC processor could feature future I/O standards such as DDR5 memory and PCI-Express Gen 5, and the switching fabric for these could be too power-hungry on 12 nm. The "Zen 4" CPU core complex dies (CCDs) of "Genoa" are expected to be built on TSMC 5 nm.

NAND Flash demand continues to rise as strong sales of notebook (laptop) computers spur PC OEMs to place additional orders for client SSDs, according to TrendForce's latest investigations. Also, the supply-side inventory for NAND Flash memory has already fallen considerably due to the aggressive stock-up activities of some smartphone brands. With customers in the data center segment expected to ramp up procurement in 2Q21, NAND Flash suppliers have decided to scale back the supply of NAND Flash wafers. Compared with other product categories, wafers have a lower gross margin. As a result of these factors, the decline in contract prices of wafers has been easing over the past two months (i.e., from December of last year to January of this year).

Semiconductor foundries across the board are preparing to raise price quotes of their 8-inch wafers from 2021. A DigiTimes report sheds light on various foundry companies, including UMC (United Microelectronics), Global Foundries, and Vanguard International Semiconductor (VIS) have raised their 8-inch foundry quotes by 10-15% in Q4-2020, with the quotes set to rise by another 20-40% in 2021. Foundries don't tend to use flat pricing, and instead rely on quotes specific to the size and design requirements of an order (by a fabless chip designer).

The foundry industry operates broadly on silicon fabrication nodes and wafer sizes. This article by Telescope Magazine provides insights into the typical use-cases for each wafer size. Although pertaining strictly to pricing of 8-inch (200 mm) wafers, an impending price-rise across the semiconductor industry can be extrapolated on the basis on significant labor cost increases. TSMC is planning to implement a 20% pay hike for its personnel in 2021.

TSMC, the world's leading semiconductor manufacturing company, is rumored to start production of its 6th generation Chip-on-Wafer-on-Substrate (CoWoS) packaging technology. As the silicon scaling is getting ever so challenging, the manufacturers have to come up with a way to get as much performance as possible. That is where TSMC's CoWoS and other chiplet technologies come. They allow designers to integrate many integrated circuits on a single package, making for a cheaper overall product compared to if the product used one big die. So what is so special about 6th generation CoWoS technology from TSMC, you might wonder. The new generation is said to enable a massive 12 stacks of HBM memory on a package. You are reading that right. Imagine if each stack would be an HBM2E variant with 16 GB capacity that would be 192 GB of memory on the package present. Of course, that would be a very expensive chip to manufacture, however, it is just a showcase of what the technology could achieve.

Update 16:44 UTC—TheEnglish DigiTimes report indicates that this technology is expected to see mass production in 2023.

TSMC has had quite a good time recently. They are having all of their capacity fully booked and the development of new semiconductor nodes is going good. Today, thanks to the report of DigiTimes, we have found out that TSMC is ramping up the production lines to prepare for 3 nm high-volume manufacturing. The 3 nm node is expected to enter HVM in 2022, which is not that far away. In the beginning, the new node is going to be manufactured on 55.000 wafers of 300 mm size, and it is expected to reach as much as 100.000 wafers per month output by 2023. With the accelerated purchase of EUV machines, TSMC already has all of the equipment required for the manufacturing of the latest node. We are waiting to see more details on the 3 nm node as we approach its official release.