TrendForce's summary of top fabs in Q3. SMIC is really gaining on Samsung.
Chinese semiconductor thread II
- Thread starter vincent
- Start date
TrendForce's summary of top fabs in Q3. SMIC is really gaining on Samsung.
A big issue is how to measure the market share of Samsung's foundry business by itself. Excluding both memory and their IDM business.
Samsung make SoCs for their consumer products unit, at least they used to, would you count that as foundry revenue?
I have seen a lot of numbers where everything Samsung makes is jumbled up together including memory.
I would not be surprised to see SMIC surpass Samsung in foundry revenue. SMIC will surpass Samsung in wafers per month even earlier than that. Maybe in a year or two, when SMIC's 28nm gigafabs come fully online, if they have not done it already.
I expect Hua Hong to surpass GlobalFoundries and UMC in wafers per month and revenue after Hua Hong Wuxi's third fab is fully operational.
Right now Hua Hong announced start of mass production on their second fab at Wuxi. Construction of the planned third fab's shell has not even begun yet.
Each Hua Hong fab at Wuxi will have close to 100,000 wpm capacity with 12" wafers. Three fabs should be around 300,000 wpm. Hua Hong also operates 8" fabs which can produce 79,000 wpm in 12" wafer equivalents. Hua Hong's HLMC subsidiary can produce 75,000 wpm 12" wafers.
GlobalFoundries should have around 305,000 wpm capacity with 12" wafer equivalents. UMC should have around 340,000 wpm capacity with 12" wafer equivalents.
GlobalFoundries might be able to command higher prices per wafer because of specialty SOI and FinFET processes. But UMC's FinFET process seems to have next to no customers. So I expect them to be hit the hardest by the new mainland Chinese competition.
Last edited:
TSMC has a fairly high gross margin, which we expect to get squeezed as competition becomes more exciting. So their revenue will decline rapidly, even if they are making the same amount of chips.
Huachen Equipment plans to set up a joint venture with Changguang Daqi, a subsidiary of Changchun Institute of Optics, Fine Mechanics and Physics
Recently, Huachen Equipment Co., Ltd. and Changchun Changguang Daqi Technology Co., Ltd. ("Changguang Daqi") signed a "Joint Venture Agreement", agreeing to jointly invest in the establishment of a joint venture company, Changguang Huachen Intelligent Manufacturing Technology Co., Ltd. (tentative name, hereinafter referred to as the "Joint Venture"). The registered capital of the joint venture is 5.5 million yuan, of which the company contributed RMB 4 million in cash, accounting for 72.73% of the registered capital of the joint venture, and Changguang Daqi contributed in the form of intangible assets (patent rights), accounting for 27.27% of the registered capital of the joint venture.
The joint venture is mainly engaged in the application verification, sales and service of ultra-precision optical component grinders and ultra-precision guideway (flat and curved) grinders, as well as undertaking the grinding processing service business of ultra-precision parts such as optical components and semiconductors (or silicon carbide ceramics).
Huachen Equipment is responsible for manufacturing ultra-precision optical component grinders and ultra-precision guideway (flat and curved) grinders, and providing equipment process verification technology. Changguang Daqi is responsible for providing grinding processing business and process technology guidance for ultra-precision parts such as optical components and semiconductors (or silicon carbide ceramics), and providing support for the joint venture's business market development.
It is reported that Changguang Daqi is an advanced technology company integrating R&D, production and sales of high-end precision optical components and other ultra-precision parts processing and testing equipment for medium and large aperture aspheric and free-form surfaces. It is a subsidiary of the Changchun Institute of Optics, Fine Mechanics and Physics of the Chinese Academy of Sciences (hereinafter referred to as "Changchun Institute of Optics, Fine Mechanics and Physics"). Changchun Institute of Optics, Fine Mechanics and Physics has long been committed to basic research, applied research and engineering technology development in the fields of optics, optoelectronics and precision machinery. It has a top-notch high-quality scientific research team in the industry, and has leading technology research and development and advanced industrialization capabilities.
The establishment of this joint venture is based on the advantages of both Huachen Equipment and Changguang Equipment, and is aimed at the grinding and processing needs of ultra-precision parts such as ultra-precision optical components, ultra-precision flat (curved) parts, linear guides, semiconductors (or silicon carbide ceramics), etc. It is a strong combination of an advantageous unit in the domestic high-end equipment manufacturing field and a leading unit in high-end technology application process research, and is in line with the support direction of relevant national industrial policies.
Scalable production of ultraflat and ultraflexible diamond membrane
Abstract
Diamond is an exceptional material with great potential across various fields owing to its interesting properties. However, despite extensive efforts over the past decades,,, producing large quantities of desired ultrathin diamond membranes for widespread use remains challenging. Here we demonstrate that edge-exposed exfoliation using sticky tape is a simple, scalable and reliable method for producing ultrathin and transferable polycrystalline diamond membranes. Our approach enables the mass production of large-area (2-inch wafer), ultrathin (sub-micrometre thickness), ultraflat (sub-nano surface roughness) and ultraflexible (360° bendable) diamond membranes. These high-quality membranes, which have a flat workable surface, support standard micromanufacturing techniques, and their ultraflexible nature allows for direct elastic strain engineering and deformation sensing applications, which is not possible with their bulky counterpart. Systematic experimental and theoretical studies reveal that the quality of the exfoliated membranes depends on the peeling angle and membrane thickness, for which largely intact diamond membranes can be robustly produced within an optimal operation window. This single-step method, which opens up new avenues for the mass production of high-figure-of-merit diamond membranes, is expected to accelerate the commercialization and arrival of the diamond era in electronics, photonics and other related fields.
One step closer to Diamond Chips.
One step closer to Diamond Chips.
TSMC has lots of advantages. They have fully depreciated older fabs. They can provide their customer with a complete foundry solution which includes even packaging. But US sanctions on TSMC's supply of chips to the huge mainland Chinese market will erode some of these advantages.
Whole new mainland Chinese fabs using the latest tools, of which Nexchip is one example, will offer much better economics and lower marginal price per chip. Nexchip will be producing 150-55nm chips on 12" wafers with 2.25x the wafer area of the 8" wafers TSMC uses in their older fabs. 2.25x the wafer area with machines operating at similar or higher wpm frequency. There have also been issues with worldwide supply of blank 8" wafers, while no such issue seems to exist for blank 12" wafers for which worldwide production capacity is still increasing. Further increasing cost for 8" fabs.
TSMC, UMC, and GlobalFoundries will have to decide what to do with their 8" fabs. Either they close them down or operate them at a highly reduced capacity.
Last edited:
I don't think HLMC revenues are counted in Huahong's revenue as of now. Maybe that will change when HH buys back HLMC and report it as part of their earnings report
I do wonder if these report miss on some of the smaller Chinese fabs, which probably add together to more than HH and Nexchip's total foundry revenue. Like for example, SMEC, CanSemi, Shanghai Jita, CR Micro, Sanan and a bunch more.
Allegedly 910c. Note the chiplet design.
Attachments
-
FC3A2C6720B3D54652544791185311EB.png -
EB6F76D0E5F494965F18B7E9697B3831.png
What competition？The vast majority of TSMC revenue comes from advanced node，and I don't see anyone capable of challenge TSMC at advanced node.
Isn't SMIC's 7nm fab also a subsidiary that SMIC only partially owns? Are the 7nm revenue included fully in the SMIC numbers?