1. Qualcomm will extend the NXP trading period to next Wednesday for approval by China;
Sina Technology News Beijing time on July 20 evening news, Qualcomm announced today that it has extended the validity of the $44 billion acquisition of NXP Semiconductors transaction to 5 pm on July 25.
Prior to this, Qualcomm had delayed the validity of the transaction by 10 times, mainly in order to wait for the approval of the Ministry of Commerce of China.
In October 2016, Qualcomm announced the acquisition of NXP Semiconductors for approximately US$38 billion, or approximately US$110 per share. In February of this year, Qualcomm raised its offer to US$127.50 per share, or approximately US$44 billion.
To complete the transaction, Qualcomm needs to be approved by nine national regulatory authorities around the world. So far, Qualcomm has received approval from eight of them, only to be approved by the Chinese Ministry of Commerce.
Qualcomm CEO Steve Mollenkopf recently said that Qualcomm is still waiting for Chinese regulators to approve its acquisition of NXP Semiconductors. Morankopf said: 'We hope to complete the deal.
Morankopf also said that if the approval of the Chinese Ministry of Commerce is not available before July 25, the transaction will be abandoned. To boost the company's stock price, Qualcomm will choose to buy back 20 billion to 30 billion US dollars of stock. (Li Ming)
2. The EU accused Qualcomm of unfair competition at low prices and sold its products below the cost price;
The EU antitrust agency revealed that it will sue the US chip giant Qualcomm for violating the relevant provisions of the anti-monopoly law, suspected of selling chip modules at a price lower than the cost, and squeezing into the market space of competitors such as Nvidia and Icera.
The European Commission revealed that it will focus on investigating whether Qualcomm's baseband chip modules are below the cost price. The investigation was initiated after the Commission received a complaint from Icera. The complaint began in 2015 when Isra accused Qualcomm of using its market position in 2009. From 2011 to 2011, we used unfair competition to suppress the company.
If Qualcomm is finally found to be in violation of EU anti-monopoly regulations, it may face a maximum penalty of 10% of global revenue. In 2017, Qualcomm's revenue was approximately US$23 billion. In 2015, Qualcomm's battalion began in the second quarter of 2015. Net income and revenue fell significantly.
Before Qualcomm had paid for Apple, the iPhone and the iPad were only allowed to use Qualcomm chips, and were found guilty of violating anti-monopoly regulations. They were fined 997 million euros, the fourth highest in the history of anti-monopoly cases in the EU. Qualcomm is once again recognized as a monopoly, and fines may hit a new high. technews
3. Broadcom’s acquisition of CA at a high price was questioned by Wall Street;
As many Wall Street analysts questioned Broadcom's acquisition of CA Technologies, the share price of Broadcom fell 16% and the market value lost nearly $17 billion.
Chip giant Broadcom recently announced that it will spend $18.9 billion to acquire cloud software and enterprise software developer CA Technologies. However, this transaction has been questioned by many analysts, causing Broadcom's share price to fall sharply in the transaction.
Wall Street investment bank Raymond James analyst Chris Caso said in a report that they were surprised by the deal announced by Broadcom and expressed doubts about Broadcom’s acquisition reasons. Caso said: 'In the semiconductor business of Broadcom and Between CA's software business, we don't see significant business synergies. CA's software business is incompatible with Broadcom's semiconductor business, and the deal is likely to confuse Broadcom's strategy.'
In November last year, Broadcom also proposed to acquire Qualcomm for $70 per share in cash and stock. The transaction size exceeded 100 billion US dollars. But in March of this year, US President Trump issued an order prohibiting the acquisition of Broadcom on the grounds of national security. Qualcomm. Subsequently, Broadcom announced the abandonment of the acquisition of Qualcomm. People's Post and Telecommunication
4. The gospel of the hand sweaty people, Qualcomm ultrasound screen fingerprint recognition or mature commercial next year;
Abstract: Recently, foreign media experienced its on-screen ultrasonic fingerprint recognition technology at Qualcomm headquarters. This technology will have the advantages of faster, safer and not afraid of complex environments such as wet hands compared to optical fingerprints.
Set micro-network news (text / Jimmy), the full-screen concept once proposed, immediately became the mainstream of the entire mobile phone circle, now whether the low-end or high-end flagship machine will inevitably hang the full screen, high-screen ratio in their own conference The first two pages of PPT, the appearance of attracting consumers' attention is also understandable in this society of paying attention to the 'face value'. In order to complete the comprehensive screen design, many old friends on the front panel are forced to be displaced or simply Say goodbye, the most familiar is the fingerprint identification module.
After the iPhone X showed the Face ID sign, the fingerprint recognition disappeared from the iPhone overnight. However, not all users will feel that face recognition can completely replace the status of fingerprint recognition. After generations of design, a batch of After a batch of users' experience, fingerprint recognition is still synonymous with safety, speed and efficiency, but it is still not friendly to those who have more sweat.
Today, Qualcomm will bring the gospel to the people with severe sweat. Recently, foreign media has been fortunate to have a prototype based on the fingerprint of the ultrasound screen.
From the demo view, the left screen is the system screen, and the right side is the screen panel with fingerprint recognition. The unlocking speed is slower. For traditional optical recognition, the gap is still obvious, even than the screen of vivo NEX. The fingerprint is unlocked more slowly.
In fact, the two diagrams are used to complete the whole process. For the purpose of development and debugging, it is inevitable that the speed will be slow. After all, it is still the debugging stage. After the mature and commercial, I believe that Qualcomm will give the user a Satisfied answer sheet.
It is reported that the principle is to use sound waves to generate fingerprints, and pressure waves are reflected from the contours of the skin. Qualcomm product management director Gordon Thomas said that ultrasonic technology has some advantages, even if the hands are wet, you can read fingerprints. Because the ultrasonic wave can penetrate the liquid. It is quickly reflected in the rejection rate of 1%, the delay of 250ms, and the nominal and traditional capacitive fingerprints are comparable. Qualcomm also emphasizes that ultrasonic fingerprints will not affect the sensitivity of the sensor due to external light problems.
Glory 10 uses the fingerprint recognition of ultrasonic technology, except that its sensor is placed under the glass in the 'chin' position, not below the screen.
The data shows that the fingerprint recognition of Qualcomm's ultrasonic screen was released in June last year, and the module thickness is only 0.15mm, but it must be used in conjunction with the OLED panel.
According to statistics agency HIS Markit, the number of mobile phones equipped with screen fingerprints will reach 100 million in 2019, and Qualcomm's ultrasonic technology will mature in the next year.
This technology is likely to start on the Samsung Galaxy S10 early next year. Well-known analyst Guo Minghao pointed out in an investment report this week that the Samsung Galaxy S10 has three large, medium and small sizes, including 6.1-inch and 6.4-inch models. The fingerprint is under the screen. In combination with the news of Korean media, the fingerprints that Samsung wants to carry are different from the current vivo, Huawei Mate RS, the optical fingerprint used by Xiaomi, but based on Qualcomm's ultrasonic technology. If Samsung can really carry this on S10 Technology, for the next year's high-end flagship market, Samsung may be the first to go. (Proofreading / group)
5. Intel enters the year of knowing the destiny: can no longer rely on Moore's Law to fight the world;
According to The Verge Beijing time on July 20th, Intel is welcoming the year of knowing the world. For it, this is an important milestone. Compared with other companies, Intel is synonymous with chips. People who know Intel should know Moore's Law - it has been driving Intel's continuous development over the past half century.
Moore's Law - from the hands of Intel co-founder Gordon Moore - means that the number of transistors integrated in the chip will double every two years. But Moore was originally published in the journal Electronics. In the paper, his prediction is that the number of transistors integrated in the chip will double every year between 1965 and 1975. Later, in a revised paper published in 1985, Moore turned the number of transistors. The time was changed to 2 years.
Intel Sandy Bridge chip
Whether it is intentional or not, Moore's Law and the pace of its development of chips have always been a core part of Intel itself. The pace of development of Intel – in fact, the computing industry – has always been determined by it. In recent years, Intel Established the tick-tock chip release strategy, released a smaller architecture size (increase in the number of integrated transistors) a year, and released the same improved version of the chip every other year.
Unfortunately, Moore's Law is not working so well: transistor size is quite small (Intel is currently developing a 10 nanometer manufacturing process - an atomic size), and the laws of physics have begun to hinder the development of chips. The size of the transistor is not completely impossible, but the speed of further reducing the transistor size (the number of transistors will increase accordingly) will be greatly slowed down, and the cost will be higher and higher.
The pace of development of the chip has begun to slow down. In 2015, Brian Krzanich, then Intel CEO, said, 'The last two technical upgrades have shown that our pace is close to two and a half years instead of two years.' Intel The 10 nanometer process has been repeatedly voted. It is expected that the release time of the 10 nanometer process chip will be 2019, which indicates that the technology upgrade interval will exceed 3 years.
Intel's eighth-generation Core chip
Intel has released a generation of 14nm+ chips (Kaby Lake R) and a generation of 14nm++ (Coffee Lake) chips, all based on the previous generation architecture, dedicated to increasing battery life and increasing the number of integrated cores instead of adding transistors. Quantity.
Although the industry has proposed a variety of solutions to this problem - including innovative new transistors, the use of new materials, and even considering the new computer working principle, it will eventually hit the wall.
Whether you like it or not, Intel is changing. Intel missed the new wave of computing represented by the rise of smartphones, and has to face the situation that Qualcomm Snapdragon series chips dominate the mobile phone field. There is almost nothing in the field of smart phones. gain a foothold.
In addition, in the next few years, the specter of Spectre and Meltdown will continue to linger. AMD and Qualcomm and other rivals are making a comeback, trying to challenge Intel's hegemony in the desktop and server areas. There are even rumors that Apple may be considering developing its own computer. Chip. After the class is not normal and the relationship is not normal, Intel also needs to hire a new CEO who can continue to open.
One might think that when Intel needed it the most, Moore's Law dropped the chain.
However, this may not necessarily be a bad thing. The slower pace will give Intel more time to optimize the existing architecture, and companies such as AMD, which are making a comeback, will have more time to catch up and drive market competition. Ultimately, this will make all People benefit.
But this means that Intel must change and move forward in the future, rather than relying on continuous iteration cycles. This is already on the Intel eighth-generation Core chip, quad-core and six-core processors become desktops. Standard with notebooks: With parallel multi-core processing technology, the potential of the existing technology and the number of transistors can be fully exploited instead of blindly improving the original processing power.
Earlier this year, the unprecedented Intel-AMD collaboration gave birth to a CPU-GPU hybrid chip that combines Intel Core processors with AMD Radeon graphics chips. Intel takes advantage of existing processor technology in a new way to deliver higher Performance and extended laptop battery life without the need to 'dead' with Moore's Law and Molecular Physics.
It may not be current, or even in the next 10 years, but Moore's Law will expire sooner or later. The next step of technology -- or Intel's technology for the next 50 years (no longer relying on Moore's Law) may change the computing world again. Frost leaf)
6. Get 10TB in minutes, Western Digital's second-generation QLC flash memory shipped this year.
Abstract: Today, Western Digital and Toshiba announced the successful development of the second-generation 3D QLC NAND with 96-layer BiCS4 architecture, which is expected to be mass-produced in the second half of 2018. Western Digital's QLC flash core capacity is 1.33Tb, which is better than the previous QMC/Intel QLC release. The 1Tb density of flash memory is also 33% higher, claiming to be the highest capacity in the industry.
According to the micro-network news, this year, Intel, Micron have reported that QLC flash drives are being developed, but these two vendors are mainly for enterprise users, and it will take a while for the consumer market.
Western Digital's second-generation QLC flash memory gets 10TB in minutes
Today, Western Digital announced the successful development of the second-generation 3D QLC NAND (4bits/cell) using the 96-layer BiCS4 architecture, which is now in the sample delivery phase and is expected to be mass-produced in the second half of 2018. Western Digital and Toshiba 96-layer QLC NAND single-die With a capacity of up to 1.33Tb, it is the industry's highest density 3D NAND.
1.33Tb is a very large capacity. Previously, the mainstream core capacity of 3D MLC flash memory was 256Gb, 3D TLC flash memory was also 256Gb, and some could reach 512Gb, while Micron and Intel's QLC core capacity was around 1Tb, and Western Digital's core capacity was 1.33Tb. This capacity conversion is 166GB, which is just a core capacity.
Generally speaking, a 2.5-inch SSD hard disk can generally hold 4 to 8 cores. That is to say, the future Western Digital SSD hard disk can easily reach 3-5TB capacity, and a little more cores reach 10TB.
It is understood that Western Digital's second-generation QLC flash memory uses BiCS 4 technology, which is jointly developed by Western Digital and Toshiba. The production plant is located in Yokkaichi, Japan. Now Western Digital's 96-layer QLC flash memory has been sampled, then Toshiba's 96-story. QLC flash memory is also fast, and it is expected to be announced and shipped.
Western Digital said that under the advantage of 96-layer QLC advanced technology, it will satisfy customers' storage needs in the fields of retail, mobile, embedded, data center/enterprise, etc. It is expected that QLC NAND will occupy the mainstream position in these applications in the future, just like the original. TLC replaces MLC as a mainstream application.
Unlike Micron/Intel, Western Digital's QLC flash drive will be launched in SanDisk's consumer-grade hard drive. Western Digital said that the latest second-generation QLC flash will ship this year, and now it's time to see the price.
Toshiba is close behind, the capacity is more horrible
This afternoon, Toshiba is also close behind, immediately announced its own QLC flash memory. Toshiba's 96-layer QLC flash memory uses the same technology as Western Digital, but due to Toshiba's superb packaging technology, Toshiba's single flash memory chip capacity can reach 2.66 TB. Toshiba is the inventor of NAND flash memory, and was the first to develop 3D NAND flash memory. It is also the first one to discuss QLC flash memory in flash memory chip companies. Toshiba's QLC flash memory specification is the same - accurately, it is the Western Digital QLC. The flash memory is the same as Toshiba, using Toshiba's BiCS 4 technology, 96-layer stack. The QLC flash core capacity is 1.33Tb, which is 33% larger than the previous 1Tb core announced by Intel.
Based on the 1.33Tb core QLC flash memory, Toshiba has developed a 16-core single-chip flash memory with a capacity of 2.66TB for one flash memory. Now with the help of QLC flash memory, the single-chip package achieves 2.66TB capacity, which is the previous 5 More than one.
The number of West has been sampled before, this year's mass production, starting in the SanDisk SSD hard drive, but Toshiba is slower than the Western Digital, only in September this year began to sample to SSD hard drive and the main control manufacturer, for evaluation and development, only began in 2019 Mass production.
With the further development of technology, each original factory focuses on the competition of the new generation QLC and 96-layer 3D NAND technology, and the single Die capacity is increased to more than 1Tb, which is doubled compared with the 64-layer 3D NAND technology.
According to the latest offer of ChinaFlashMarket in China's flash memory market, the price of 240GB TLC SSD has dropped to 41 US dollars. The cumulative decline in 2018 has exceeded 30%, and it has fallen below historical lows. The high-capacity price continues to decline, which is expected to stimulate the mainstream demand from 128GB to 240GB. Upgrade and stimulate market demand.
Previously, the QLC hard drive released by Micron was 7.6TB, and Intel also revealed that it will launch a 20TB SSD hard drive, but the thickness is higher. But the current technology is not mature. For ordinary consumers, they want to use TB level. Low-cost SSD, and then wait for QLC flash drive.
Many players are worried about the reliability of QLC flash memory, just as many years ago worried about TLC flash memory is not the same, but from the manufacturer's point of view they are not likely to give up QLC flash memory, because the temptation of capacity density is too big - Western Digital's QLC flash core capacity of 1.33Tb is 33% higher than the 1Tb density of the previous QMC flash memory released by Micron/Intel. It is claimed to be the highest capacity in the industry.
Of course, it is not realistic to engage in such a large capacity. It is not the technology that can't be done, but the cost of the consumer. In any case, ordinary people want to use the TB-class low-cost SSD hard disk in the future, and can only wait for QLC. Flash hard drive.