However, due to the original development plan may be used or emissions of chemical substances change, environmental analysis of the difference in the report to be re-sent…. Review, in November, after the initial success, and finally in the 28th EIA Committee adopted, announced the expansion of plant step, the future can be a good sprint 10-nanometer production.

Related to Taiwan Semiconductor Manufacturing Company 10 nm, 7-nanometer manufacturing process in the Science Park (formerly Daishan ammunition sub-library) development case, the four environmental impact assessment in February 2015, to the Section 1 and 2 health risk assessment also included In five under the assessment, conditional. However, the plant will be used or derived chemical substances and the number of projects to be changed, TSMC and then re-send pieces of consideration, November 3 this year, the Environmental Protection Department ad hoc group adopted preliminary re-submitted to the EIA Assembly, 28, 306 The EIA meeting finally passed the case, declared the dust settles. gmf32024abtw

According to the Environmental Information Center of the Taiwan Environmental Information Association, the EIA General Assembly has also requested that the SEGA conduct regular and occasional checks on the use of TSMC chemicals. TSMC also needs to disclose the air, water and waste Such as sewage situation. TSMC senior director Zhuang Zishou pointed out that the ring through the case, but also to the current test run of TSMC 10 nm eat reassurance, TSMC will start mass production, and expressed the hope that 2017 has good news.

The development scale of the development project of Zhongke Park (formerly Da’er Shan Ammunition Sub-library) is 53.08 hectares. TSMC applies for an area of ​​about 32 hectares. It is estimated that P5 and P6 fabs will be completed in 2017. Except for 10- Mass production, 7-nanometer process is expected to be arranged in the official. msp430f436ipzr

TSMC Kaohsiung campus expansion ring difference case is also the end of November this year through the first trial of the ring, the expansion is expected to spend 500 billion Taiwan dollars to build a new plant, to start next year, the development of 5 nm and 3 nm process.

According to Sigmaintell earlier analysis, brand manufacturers in 2017, product planning, will decisively stop Sharp 60-inch new layout, will demand to 65-inch. So, this change in the first quarter of 2017 panel industry, especially Chinese manufacturers have what effect?

First, Samsung Electronics is SDP 65-inch main customers, while Samsung Display (SDC) is the world’s largest supplier of 65-inch. Samsung Electronics to 65-inch transfer demand, we expect will lead directly to Samsung Electronics and SDC in 2017 the supplier relationship has been strengthened.

Second, 65-inch panel resource allocation face adjustment. Currently the world’s 65-inch panel suppliers mainly for the Korean plant SDC, LGD, Taiwan and AUO factory Chong Chong. SDP 65-inch shipments are calculated on an average of 250K, Samsung and other brands by the single-affect, these numbers will be gathered to the 65-inch suppliers. The first is SDC and LGD, followed by Taiwan and AUO Group. Therefore, in 2017 the Chinese brands available 65-inch resource allocation will be compressed, small and medium brand pressure will be greater. xc4020xl-3pq160c

Third, the short-term supply will cause 65-inch tension, and produce butterfly effect, the spread to the adjacent size.

Fourth, the Chinese panel makers BOE and CEC Panda also have a small amount of 65-inch shipments. BOE in G8.5 generation does not have the economic cutting capacity, can expand the space is limited. But CEC Panda G6 generation has economic benefits, CEC Panda can seize this time window, expand the TV customer base and market discourse. xc17s05pd8i

Finally, the long term, even if the SDP panel to stop external suppliers, but its production capacity is still to maintain full production marry, regardless of the final brand to the sea for Sharp, or Foxconn OEM release number, the global total shipments will not cut back. The future of more than 65-inch large-size panel supply and demand situation remains to be cautious.

TSMC, Samsung 10nm production process has entered the countdown stage, Apple A11 / A10X, Huawei Unicorn 970, Qualcomm Xiaolong 835, MediaTek Helio X30, etc. a new flagship chip are also highly anticipated.

According to the plan, Apple’s next-generation iPad processor A10X will be the first in March 2017 TSMC production in the second quarter will continue to mass production of iPhone 8 processor A11, several other are also arranged in the first half of next year.

However, according to the latest news from Taiwan media, TSMC, Samsung 10nm process yields are not as expected, production is facing an embarrassing situation, and even affected the future of new technology research and development. ms3348-8-10

TSMC has been a senior R & D director Wu Xian Yang (16 / 7nm) and Cao Min (20 / 10nm) turns the leader, but the future 5nm to March this year before joining the TSMC R & D department of the former Qualcomm senior technology Director Geoffrey Yeap. TSMC plans to mass production in 2020 5nm, and then Intel is also expected to launch 7nm, in order to better than rival TSMC will be launched a lot of new technologies, including optical / UV lithography, germanium / 35 materials instead of silicon and so on.

Samsung, said its 10nm yield is not even as TSMC, leading to high tension and the relationship is very tight.

Qualcomm because of dissatisfaction with the progress of TSMC and defected to Samsung, 14nm on the co-operative fairly happy, so continue on the 10nm, is expected to return to TSMC 7nm, Samsung 10nm is not smooth, leading to Qualcomm chips have to return next year 14nm, such as Xiao Long 660. ms-470n

Vertical Channel Filled Metal 3D NAND Flash One of the key issues in the development of vertical stacking over 64 layers, the Atomic Layer Deposition (ALD) process can be formed in large areas in a high aspect ratio vertical channel Film, and has a good step coverage (Step Coverage), it applies to more layers of 3D NAND Flash vertical channel filler metal, but its deposition rate is slow and the higher the cost of materials and other issues.

ALD chemical vapor deposition (CVD), and another CVD technology plasma-assisted chemical vapor deposition (Plasma-Enhanced CVD; PECVD) compared to ALD using sequential injection of the first and second Precursor as a reactive gas to vaporize the film, which is different from the PECVD using plasma evaporation technology, while evaporation of two or more precursors for chemical reaction. ms3348-8-10

The advantage of ALD is that the film thickness can be accurately controlled and a large area uniform film can be formed at the atomic level precision, and it is suitable for depositing a thin film with excellent step coverage in the concavo-convex structure. On the other hand, PECVD is not easy in the concave-convex structure or the deep hole pattern Reaching the same thickness of the film, it ALD process is expected to become more than 64 layers of 3D NAND Flash vertical channel filler metal solution.

However, the deposition rate of ALD is slower than that of PECVD because of the long process time, and the amount of precursor needed by ALD is high, which makes the deposition cost increase. ms-470n

3D NAND Flash to 64 layers above the development of the need to improve the deposition and etching of thin film deposition, vertical through channels, channel filling metal and other processes, of which, film deposition process time, vertical through the channel need to change the cost of dry etching Technology, and channel filling metal ALD process required for the process time and cost increases and other issues, so how to cope with rising costs will become the 3D NAND Flash technology to develop more layers of important issues.

With this year is coming to an end, in the 3D NAND, advanced foundry and advanced packaging investment driven, making the equipment spending power than the beginning of the strong forecast …

The book-to-bill ratio (B / D) for the North American semiconductor equipment manufacturers was $ 1.55 billion in November 2016, according to the latest book-to-bill shipping report from the International Semiconductor Industry Association (SEMI) B value) of 0.96, on behalf of the semiconductor equipment manufacturers in the month of $ 100 per shipment of products, you can receive orders worth $ 96. b32529c182k289

SEMI report pointed out that the North American semiconductor equipment manufacturers in November 2016 the world received an order estimate of $ 1.55 billion, compared with $ 14.9 billion in October growth of 4%, and compared with $ 1.24 billion in the same period last year, growth of 25.1 / RTI & gt;

Shipments in November were $ 1.61 billion, a slight decrease of 1.1% from $ 1.63 billion last month, but up 25.2% from $ 1.29 billion for the same period last year.

“As this year is drawing to a close, driven by strong investment in 3D NAND, advanced foundry and advanced packaging, equipment spending is expected to be stronger than expected at the beginning of the year,” said Cao Shih-lun, president of SEMI’s Taiwan region. Expenditure growth in 2017. tda7580

SEMI’s published B / B values ​​are based on the average order book value of North American semiconductor equipment manufacturers for the past three months, divided by the average amount shipped in the past three months.

Xilinx Announces New Amazon Elastic Cloud Compute (Amazon EC2) F1 Enabler Types for its 16-nm UltraScale + Family of FPGAs to Accelerate Gene Research, Financial Analysis, and Imaging in Amazon’s Cloud-Based Web Services (AWS) Processing, massive data, security, and machine learning to infer workloads.

In addition to the Amazon EC2 F1 implementation, AWS has released an Amazon Machine Image (AMI) design tool that provides FPGA developers with pre-built development tools and scripts that include the Xilinx Vivado Design Suite and authorizations. pps272

“We believe that FPGAs will be the dominant application for cloud computing,” said Steve Glaser, senior vice president of Xilinx’s Enterprise Strategy Group. “The AWS announcement last week confirms that this trend is taking place and is unstoppable. lm2936mp-5.0/nopb

Storage-class memory (SCM), also known as persistent memory, is a composite storage technology that combines both traditional storage and memory characteristics. Since storage-level memory can provide more memory than flash Faster read / write speeds are cheaper than DRAMs in terms of cost and are expected to be the next generation of revolutionary storage technologies, both from the performance side and from the application side.

Strictly speaking, the storage level memory is not entirely memory, not entirely traditional storage device, although it is like DRAM can be embedded into the motherboard slot, but its computing speed slower than DRAM, and even in the Loss of power in the state can still continue to save data, at this point is relatively close to the characteristics of traditional storage devices. NJU7221L20

However, compared to traditional storage devices, storage-level memory not only in the read and write data to provide faster speed, while the erase and write data when the program also has greater flexibility, unlike the general flash memory is very Easy to wear and tear because of high frequency read and write times.

Interestingly, the storage-level memory can process data in bytes or blocks at the same time, giving the operating system, software, and virtual machine hypervisor the developer , In the use of storage-level memory as a medium application device to obtain greater flexibility.

For example, the operating system would initially view storage-level storage as a block storage device formatted by the file system and a database with compatibility capabilities, but the next-generation application might choose to map directly through memory Memory-mapped files to access storage-level storage, and the virtual machine monitor can allocate separate areas of the storage-level memory to different virtual machines for use as program execution memory, or the like Flash memory storage resources. 2SD466

Compared with the use of traditional DRAM, the biggest difference between the storage-level memory is that it can be stored continuously without loss of data due to the interruption of external power supply. In addition, it can process data in units of bytes, The need to store data packages in blocks of 512 consecutive bytes, the combination of these two features is bound to change the future of the application design architecture and manner.

On the current status of development, the storage level of memory technology is expected to be in the end of 2016 into the practical application stage, the beginning will be made by Intel (Intel) 3D XPoint technology architecture come out on top, and Hewlett-Packard (HP) and SanDisk In the past has also been announced to cooperate in the development of SCM technology, but the fastest may have to wait until 2017 or later have the opportunity to really come out.

As with any emerging technology, the early use of storage-level storage may be limited to specific industries and uses, especially in the performance of price and performance considerations, most of this technology can only be applied to specific areas, Including in-memory computing, high-performance computing, and server-side caching, are likely to be early adopters of storage-level storage.

Drive and touch chip manufacturers to launch integrated drive and touch functionality TDD chip victory, including Huawei, ZTE, OPPO, VIVO, millet and other top ten mobile phone factory in mainland China all adopted, to this end, the fourth quarter single Month TDDI shipments can be estimated to reach 3 million, quarter by 50%, and led the fourth quarter revenue is expected to buck the trend in the first quarter of next year, shipments is reached 4.5 million, 50%, the first quarter revenue is optimistic.

TD-TDDI chips shipped in heavy volume, since June revenue growth momentum to reproduce, and even in the third quarter suffered a shortage of small panel supply, but the overall third-quarter operating results in line with expectations, especially with TDDI chip official heavy volume shipments , Single-month shipments reached 200 million, and help the third quarter gross profit margin is to return to two percent of the mark. pps272

After entering the fourth quarter, although the small panel side is still tight supply pressure, but the Thai-chip in the TDDI side, repeatedly heard the news received orders, in particular, has been identified to include ZTE, Huawei, millet, OPPO, VIVO, etc. China’s top ten mobile phone factory orders in hand, in this injection, the fourth quarter of Dunhua TDDI single-chip shipments will reach 3 million, compared with the third quarter of the month of 200 million, quarter by five percent.

In addition, in response to the Lunar New Year next year’s procurement season demand, Dunt is currently the first quarter orders are almost all in place, the first quarter of next year, TDDI chip monthly orders has reached 4.5 million, compared with the fourth quarter of this year, , With considerable growth, to the second quarter of next year, shipments are expected to continue to climb up to 700 million, quarterly more than 55%. lm2936mp-5.0/nopb

In other words, with the high price and high margin of the TDDI chip shipments are showing quarterly growth rate of more than 50%, Dunnet fourth quarter, next year’s first quarter and next year’s second-quarter revenue is quite visible, Gross margin will be 20% mark for the bottom line, a significant quarterly upward climb.

Following the acquisition of NT $ 130 billion of shares in the company, Micron decided to expand its investment in Taiwan and plan to build Micron’s first 3D memory IC packaging and testing facility in China and to recruit former Amkor, General Manager Liang Mingcheng as general manager of the business in Taiwan, the new investment plan is scheduled to be announced on the 12th.

Liang Mingcheng on the 4th confirmed at the end of October to leave Aiker, but Micron has not yet officially released personnel on the cloth, he said, still resting low-key, detailed positions such as Micron announced.

It is understood that the acquisition of Micron shares Micron has been the end of the equity case, Subfamily has been on November 30 to stop trading, Micron scheduled to hold the United States on the 12th to join Micron’s Micron celebration, Micron CEO Duncan (Mark Duncan) Will once again personally arrived in Taiwan to host, is expected to also be announced simultaneously related personnel and expand investment in Taiwan layout. tps74801drcr

Semiconductor equipment manufacturers to disclose, Micron completed the acquisition of subfamily, the future will be incorporated into the Taiwan Micron after the company, a Micron DRAM advanced process pioneer. Micron plans to set up in the next step in the most advanced 3D DRAM IC packaging and testing plant, is currently negotiating with the relevant government departments and related land purchase operations.

In addition, Micron also plans to HuaYaKe in Taoyuan HuaYaKe park next to the land, for expansion. In late September of this year, Micron CEO Duncan personally called Mayor Taoyuan Zheng Wencan, that is, for the relevant expansion and EIA case, hoping Taoyuan city government to help.

Semiconductor equipment manufacturers have revealed that Micron has completed in Ruijin, Taiwan Branch is the largest share of Micron production DRAM production after the acquisition, coupled with the Taiwan financial institutions to borrow 80 billion yuan, in the Taiwan’s total investment reached 600 billion yuan.

Micron follow-up to expand the investment case, including 3D DRAM IC packaging and testing plant and subfamily sub-expansion case, the estimated total investment amount is 100 billion yuan. gmf32024abtw

Long-term and Micron’s Formosa Plastics executives also stressed that the focus of US DRAM production focus bet in Taiwan, the main fancy talent in Taiwan and semiconductor supply chain integrity, short-term can not turn to cooperation with the purple, the focus shifted to mainland China, Taiwan is not yet To be marginalized.

South Asia Branch in the disposal of subfractional stocks, the decision of which 445.7 billion yuan to 17.25 US dollars per share investment in Micron private equity common stock, made Micron 5.02% stake, as Micron third Large shareholders, and made Micron 1X / 1Y advanced DRAM technology licensing, the two sides still maintain close cooperation.

KAIST IP filed a patent infringement lawsuit against the Texas Federal District Court on November 30, alleging unauthorized use of Samsung Electronics, Qualcomm and GlobalFoundries, according to Korean media. KAIST IP, a patent management subsidiary of the Korea Institute of Science and Technology (KAIST) Its own “FinFET” technology patents, require payment of royalties.

“In addition to Samsung, Qualcomm and Grosvenor, the future plans for Taiwan’s TSMC, Apple (Apple), Apple and Apple (Apple),” said KAIST IP said, “Samsung has long been ongoing and Samsung to pay the royalties on the issue of consultation, but Samsung fully rejected, resulting in the breakdown of negotiations, )Sue”. mbrb2035ct

KAIST IP said that Samsung, Grosvenor, TSMC use FinFET technology to produce, sell mobile phone chips, but do not pay royalties. Samsung, Groot grid supply chip to Qualcomm, TSMC is to help Apple to produce iPhone chips. j1011f01p

According to reports, Samsung said, “Samsung is currently in charge of the contents of the complaint.Samsung from the early 2000s on the development of 3D semiconductor technology, and has a FinFET technology-related patents.”