If you have not heard, the United States and China are engaged in a trade war. On the American side, one of the major problems is the problem of forced partnerships and intellectual property theft that US companies often report the price of doing business in China. China uses many sophisticated technologies, but according to the chip designers themselves, it will be impossible to meet Beijing's ambitious targets for local production if the United States refuses to cooperate.
There are alternatives in China, but the technological gap is too big, "said one of the leaders of one of China's leading artificial intelligence chip manufacturers, which relies on United States technology for chip design, said Nikkei "If we lose access to the US software or can no longer receive updates, our chip development will end up in a dead end."
Officials from NextVPU, an AI company created by former AMD employees, also echoed his remarks, including a representative of Semiconductor Manufacturing International Co, or SMIC, the most important foundry in mainland China. "We would use all the equipment and chip materials we have locally if their performance was sufficient," said a senior official at Semiconductor Manufacturing International Co., the largest chip manufacturing contract in China. "But we still need (US) equipment, materials, IP addresses and chip design software.It is unlikely that any of the world's chip makers .. is getting rid of American sellers soon. "
This chart from IC Insights clearly shows how flawed the market for chip design firms is.
SMIC, for example, is currently focusing on increasing production at the beginning of 14 nm, with plans to build a $ 10 billion smelter to focus on 14nm production for customers local. Assuming this facility is online and fully operational by 2022, China's production of integrated circuits will be about five years behind the rest of the world in terms of node leadership and seven to eight years in terms of volume production. However, according to reports, China 's overall self – sufficiency rate in semiconductor production would be only about 15%, due to the globalized and fragmented nature of the chain of custody. global supply.
Made in China 2025 plan demands 40% of Chinese Semiconductor Built in China by 2020 and 70% by 2025. According to experts, the currently achievable target is 20.5% by 2023. According to Nikkei, one of the obstacles China faces is that its own developers are more reluctant to rely on its local sites. technology produced or pay the higher prices associated with smaller-scale production from local fabs. This is another piece of this puzzle: companies do not perceive the Chinese product as an equivalent product, so they are less likely to buy it.
All of these preferences and positions are of importance, as long as China has a way to buy semiconductors made in the United States. If these roads closed completely, that would be another story. Having no other choice than to rely on its own know-how in semiconductor design, China would probably be late, probably for years. But such a decision from the United States would also attract considerable investment from the Chinese government, which would probably feel like they have no choice but to spend money to reproduce these pieces of the technological puzzle. that he could not buy elsewhere.
This is the central risk of any business, at least from the point of view of commercial enterprises. US companies that spend billions on developing new technologies do not want to give way to Chinese companies because of the cost of doing business. At the same time, however, they do not want to be totally excluded from these markets. Freezing the Chinese global semiconductor market by stifling access to American technology will not bring the Chinese back to vacuum tubes (although I would not mind seeing the first and only phone cell in the world powered by vacuum tubes). However, this could trigger a series of geopolitical events that have led other countries to agree to move away from US technology companies and sources if they see us as uncertain and unreliable partners.
Historically, it is extremely difficult to prevent the technology from wanting to acquire it. The United States could not stop other countries from inventing the atomic bomb after the Second World War. Firms with strong market conditions and near-monopolies in software distribution can benefit from these conditions, but even in these sclerotic markets, the change eventually occurs. Microsoft and Intel defined the PC sector from the late 1980s to the late 2000s, but now share space with various computer peripherals based on other operating systems and running different processors. architectures. Stuxnet – one of the most sophisticated malware programs ever created – has successfully delayed Iran's nuclear ambitions, without however preventing them alone.
Ironically, the impact of the current export control decisions by the United States is as unknown as the conditions governing other long-term bets made by semiconductor manufacturers. As Robert Palmer, former CEO of Digital Eleven, said: Designing microprocessors, it's like playing Russian roulette. You put a gun on your head, pull the trigger and discover four years later if you blew your brains.
It is always possible, of course, that the United States and China agree to resolve their trade disputes and that Chinese companies regain access to US technology. If that does not happen, however, we will be in Russian roulette for the long-term consequences. No country can afford to be completely cut off from the technology and semiconductor markets, which means that China would have little use other than to build its own knowledge base. and its own components, possibly as part of a broad international coalition with partners fearing to be treated. likewise by the United States.
There is no doubt that refusing to work with the Chinese would significantly slow down China's technological development – for a while. If that presents a long-term benefit to the United States – that is the question, is not it? Part of the argument is that the Chinese will recognize that they can not live without access to US semiconductor technology and that they will have enough incentive to play ball. But it is also possible that the Chinese, even if they are forced to return to the negotiating table for the moment, file their own long-term plans to move away completely from the technological dominance of the United States. . None of this means that the US declared objectives in the current trade war are less important, but they clearly show how problem-solving of this type can be complex.
Background image by SMIC.
The United States on Friday added five Chinese companies to a blacklist, limiting their access to US technology. The "list of entities" identifies entities for which there is a reasonable basis to believe, based on specific and explainable facts, that are involved, involved or have a significant risk of being or participating in activities contrary to US foreign national security policy ".
The companies in question are: Sugon, Higon, Chengdu Haiguang Integrated Circuit, Chengdu Haiguang Microelectronics Technology and Wuxi Jiangnan Institute of Computer Technology. One of them, Higon (also spelled Hygon), is a semiconductor joint venture between AMD and THATIC, in charge of selling x86 processors for the Chinese server market. THATIC itself consists of two separate joint ventures: Chengdu Haiguang microelectronics technology and Chengdu Haiguang integrated circuit design. If you look at the list above, these two companies are included.
In short, the US government seems to have just banned THATIC, the two companies that formed it, and the non-factory semiconductor company to design and sell x86 processors in China. According to the government's information document, all of these companies are owned in part by Sugon, which seems to be the main target company. The document of the Ministry of Commerce states It was decided that Sugon and the Wuxi Jiangnan Institute of Computer Technology were involved in activities:
United States … Sugon, the Institute of Computer Technology Wuxi Jiangnan and the National University of Defense Technology (NUDT) are the three main Chinese entities in the development of exascale high performance computing . Sugon has publicly acknowledged a variety of military end-uses and end-users of his high-performance computers.
This would seem to be a serious problem for AMD's efforts to expand its operations in China as part of the THATIC joint venture. We contacted AMD and received the following statement:
We are currently evaluating the addition of five new entities to the list of entities by the Bureau of Industry and Security. AMD will comply with the regulations governing this list, just as we have complied with US laws to date. We are reviewing the order details to determine the next steps for our joint ventures with THATIC in China.
Reports of earlier this month stated that AMD did not license Zen Architecture 2 to THATIC. The future involvement of AMD, its revenue forecast and exposure to it are currently not clear.
The move comes as President Trump prepares to meet Chinese President Xi Jinping in China next week and discusses how export laws can be used to radically change the competitive landscape. These determinations will change the balance of power between AMD and Intel in China, as well as the market available for Semiconductor in this nation the NYT reports that the United States is considering adding Hikvision to the list. Sugon is a leading exascale computer manufacturer in China, with 10 of the fastest Chinese supercomputers according to the TOP500. Sugon is a much smaller company than Huawei, but these specific lists will continue to impact the semiconductor industry given their exposure to the HPC market – a critical area in which AMD hopes to gain market share.
There was a time when computer capabilities were considered a national resource of the United States, and strict restrictions were imposed on the export of computer technologies. As Steve Jobs has already said: in 1999"The Power Mac G4 is so fast that it is classified as a supercomputer by the US government, and we are prohibited from exporting to 50 countries around the world." These restrictions have been considerably relaxed over the years. following, but some steps have been taken to tighten them since then. In 2015, for example, the Obama administration banned Intel, Nvidia and AMD from selling chips to the Chinese government. The Trump Administration's decision to further tighten the rules and lock down some subsidiaries is a welcome expansion of these restrictions, but not without precedent.
According to rumors, Intel could prepare to significantly reduce the price of its processors, with the upcoming launch of the AMD Ryzen 3000 family. What is less clear is if it represents a real potential outcome or simply wishful thinking.
According to DigiTimes (subscription necessary), Intel could reduce processor prices by up to 15 percent in response to AMD's upcoming Ryzen parts and has already informed motherboard manufacturers of this change. It is possible that Intel will take this step, but judging by the company's actions since AMD introduced Ryzen in 2017, it is rather unlikely.
Intel has taken several important steps to improve its products in the months leading up to the launch of Ryzen. He added hyper-threading to some processors in the Pentium class, and then, with the 8th generation family, increased the number of cores on his various processors. Similarly, the HEDT Skylake family of pieces has increased the number of base pieces at the same price. Intel's six-core and eight-core processors are about half the price they were four to six years ago.
What Intel has not done all this time is actually a reduction of its CPU. the prices as such. Instead, the company simply waited until the time came to introduce a new generation of parts and position the new processors more favorably than the old ones. This allowed Chipzilla to adjust its prices to take into account AMD 's competitiveness without launching into titles such as "Intel could reduce the price of its processor before the launch of Ryzen 3000 by AMD".
Intel does not want to see prices directly because it means you have competition in the first place. Not only that, but the fact that Intel's HEDT components are priced, even in the face of AMD's high performance Threadripper competition, means that the company has not faced much of the current threat of these chips. If users aggressively transferred their workstation purchases from Xeon to TR, Intel would have been forced to react. The fact that he did not suggest that they had not done it.
Intel chief executive Bob Swan has previously admitted that his company expects increased AMD competition in the server market and that gross margins could suffer, which would imply that some price cuts could take place, but not where society would prevail. Currently, the high-end parts of Intel are those where are virtually the essential price differences. An eight-core Intel processor can cost around $ 500 versus $ 300 for an AMD equivalent, but a 16-core AMD processor costs $ 829 on Newegg, while a 16-core Intel Core i9 Skylake X processor costs 1,700. USD. It would take more than 15% of the price to bring the two chips to parity.
Based on Intel's actions to date and its competitive response to Ryzen, we believe the company will reduce its prices as much as necessary. It may not be ready to happen before the launch of the Ryzen 3000 family and may choose to focus on lowering prices on specific SKUs for which it needs to improve its competitiveness rather than significantly reducing costs. At the same time, the room for maneuvering Intel to achieve certain price cuts, at least on some parts, is unclear. We will have to see how Ryzen 7nm parts stand out before saying more.
Performance data indicates that AMD Ryzen 7 3800X will give the same result as the 9900K in multicore, while the Ryzen 5 3600, the cheapest of those announced, it will work like the 9700K. The 3800X will cost around 400 euros depending on the filtered prices, while the 3600 will be at 209 euros. This is equivalent to 100 and 180 euros respectively less than the Intel processors against which they will compete in performance.
Therefore, Intel, despite the shortage that suffers, could not remain inactive at some processors that will be on sale at least six months before their, on the basis of Ice Lake 10 nm for desktops, which have not been announced yet and will have to see if they arrive on time.
The price decrease, as revealed by Digitimes, will be up to 15% in the processors of the eighth and ninth generationsWhat are the ones that the company currently has to sell. Intel would have already informed component and motherboard manufacturers that they would already receive the chips at a reduced price.
Intel processors are down for a few months after the 9900K reached 600 euros. At the moment he is in 488 euros on Amazon, and a decrease of 15% would place it at 425 euros, which could compete in price with AMD chips.
In recent months, Two-thirds of processors sold on PC in a German department store are AMD, only one-third for Intel. The difference could be even steeper since the launch of the Ryzen 3000, where, for a lower price, there will be more powerful processors, not only multicore, but also multicore, which will make them ideal for playing and working without any problems. type. of sacrifice. In addition, AMD will create exclusive PCIe 4.0, something that Intel will not have until at least 2020, or even 2021, if they decide to go directly to PCIe 5.0. AMD also improved its performance under Windows 10and is safe from all the defects that appear in Intel processors.
Therefore, all this will result in consumer benefits because we will have next month more powerful and cheaper processors. AMD gets rid of the stock of forced marches of Ryzen 2000, with prices that make you laugh, like the 141 euros, which was worth 125 euros a few days ago, including two promotional games (The Division 2 and World War Z), so the chip goes out almost for free.
Today, PCI-SIG has announced that it has already started work on PCI-Express 6.0. If you've followed the latest news about PCIe 4.0 and 5.0, this may seem excessive. PCIe 4.0 is not yet withdrawn from the market, PCIe 5.0 will not arrive until 2020 at the earliest, and the PCI-GIS wants to finish PCIe 6.0 by 2021. Given the typical delay of Adoption of 12 months, put the PCIe 6.0 on the market by 2022.
This represents a significant improvement rate compared to the relatively poor speeds with which we have lived for sixteen years. When I started out in IT, the now venerable ISA bus offered a bandwidth of 8.33 MB / s. By the time I was going to college, PCI and its dedicated graphic cousin, AGP, were the standards of choice, but PCI's bandwidth limitations have become major limitations. Eleven 3D video cards came on the scene. PCI-Express, which debuted in 2003, initially offered significant bandwidth improvements, but for many years, it was unclear what could practically stretch the interface. Simple graphics cards have never pushed PCIe particularly hard the configurations showed higher scaling in the paired x8 / x8 slots, which had more to do with the latency chipset caused by the suspension at the x16 / x4 asymmetric configuration of the southbridge.
Since then, two things have changed the market. First, SSDs are being upgraded to PCI Express as a storage medium, which means that storage performance improvements are now tied, at least in part, to improvements in the underlying standard. It is obvious that no one who has been irritated by the restrictions imposed by a PCIe 3.0×4 configuration will be able to stay much longer. If PCIe 6.0 arrives in 2022, an x1 connection would offer the same bandwidth as the current high-end PCI X8 x8 SSD. The few readers who even offer this type of interface today are business products.
Increasing PCIe 6.0 bandwidth up to 256 GB / s bidirectional bandwidth would put an x16 connection at par with VRAM bandwidth available in a low-end GPU like the current RX 560. This is significantly more bandwidth than what even the DDR5 should provide by 2022 in a dual-channel configuration. I do not want to stress too much the comparison – the latency is going to be very different between the two solutions – but that partly explains why companies like Intel are pushing towards the idea of non-volatile storage taking over some of the roles mainly played by the DRAM.
Although latency-aware workloads still respond well to high-speed caches and fast interfaces, there is a proven window of opportunity for high-bandwidth, high-latency products. GPUs are often cited as examples of products that are not particularly latent. PCI-Express 6.0 is also likely to be useful in auto-drive systems, industrial IoT and any system that juggles with multiple sensors in order to combine the inputs of multiple devices into a coherent whole. The increase in bandwidth per pin can give the manufacturer the opportunity to perform the same job with a smaller size or dedicated cabling to the bus itself.
If the PCI-GIS succeeds, it will offer an 8-fold increase in effective bandwidth in just four years. Considering that it took 15 years to achieve this type of gain so far, the jump would be considerable. We did not expect GPUs to be the best beneficiaries, but storage arrays and nonvolatile memory could generate significant gains, as well as AI accelerators and FPGAs in deployments with limited bandwidth. If PCIe 6.0 is really ready by 2022, fewer and fewer companies may be adopting version 5.0 or 4.0 – or smarter, so there may be just one new product that allows them to easily increase performance. Whatever the case may be, if PCI-SIG reaches this delivery rate, storage controllers will again become the limiting factor in storage performance.
Arrival of PCIe 5.0 in 2019 with 4 times more bandwidth than PCIe 3.0
The PCIe 5.0 specification is now available even before PCIe 4.0 ships
How the Makimoto Wave explains the tsunami of market-leading AI processors
New rumors are running that Intel may seek Samsung's assistance at 14 nm, although there are also reasons to doubt it. If this is true, it suggests that Santa Clara will remain stuck at 14 nm for a significant amount of time for at least a few games, notwithstanding recent discussions on Ice Lake.
according to SE daily (via Google Translate), Intel and Samsung are in the final stages of trading for additional capacity. Intel would have chosen to work with Samsung rather than TSMC because of concerns regarding the improved competitive performance of Huawei and AMD. TSMC stated that it thought it could continue to manufacture chips for Huawei, and that it allegedly prompted Intel to prefer Samsung as a partner, due to the possibility that new business decisions regarding retaliation would be taken against companies that do business with Huawei.
I do not want to go so far as to say that it's wrong, but the chronology seems extremely compressed. Negotiations on smelting capacity between two large companies will not be negotiated in a weekend, and the US government total blockade Huawei is still quite new. In addition, taking action against TSMC for believing that it could continue to manufacture systems on a chip for Huawei would, in some respects, be excessive. Huawei is facing enormous problems in bringing its products to market for reasons that have nothing to do with its ability to supply SoC. Even with perfect support for the smelter, its manufacturing supply chain is threatened in an existential way, not to mention its access to software and support tools.
The idea that Intel would choose to use a smelter other than its main competitor, AMD, possible Intel could be sensitive to the idea that it was passed, hat in hand, to the same company that supplies its competitors. The partnership with Samsung – whose 14nm node is generally in excellent condition and has been used for AMD hardware at GlobalFoundries after GF fired it many years ago – is a little less direct.
The biggest reason to look down on this rumor is that it suggests that Intel would launch at 14 nm the competitor "Rocket Lake" on silicon Samsung. In the past, Intel had signed agreements with TSMC for the production of Atom processors or chipsets (as is often said). Building "big hearts" in a rival foundry would be a major change. That's one of the reasons I do not want to weigh heavily in this rumor, but there's a way to make sense of this rumor.
One of the difficulties associated with setting up a new process in an existing plant is the disruption of ongoing production. If you want to replace a capacity of 14 nm by 7 nm, you may have to disable the lines to perform the upgrades. To do this, Intel has always operated its production lines in tight rates, but we know that the demand for 14 nm has been extremely high. Just last year, Intel announced the allocation of additional funds to boost production of 14 nm. At the same time, the long 10-nm delay has plagued Intel's installations. The company expects a relatively fast switchover to 7nm (production being scheduled by 2021), which means that it needs a fairly fast volume rollout at a time when demand for 14 nm can already be very busy.
If this rumor is true, it may be true to the extent that Intel has reached an agreement with Samsung to operate certain products from its own factories while aggressively upgrading its own factories. The company undoubtedly wants to restore the story of the supremacy of the process that it had 20 years before its 10 nm slide and it might prefer to run at 7 nm by taking advantage of the production of a competitor rather than conducting it alone.
The Daily SE suggests another reason why Intel and Samsung could conclude this type of agreement: prices. From the story:
The Samsung smelter recently announced that it had submitted to TSMC an unexpected unit price of 60% for some companies. Samsung has offered TSMC a complete set of less expensive masks than the "multi-layer mask" (MLM) set up to reduce low-volume production costs. A mask is a kind of film used to draw a circuit on a wafer.
While the dramatic cost reductions we've heard about were 7nm, it's quite possible that Samsung and Intel will also reach a 14nm agreement. Samsung Foundry will probably be hungry for customers and build for Intel would be a prestigious victory. Intel (again, assuming this rumor is accurate) would obviously want a good deal for the products and could find Samsung more acceptable than TSMC – or simply worry about more prosaic issues regarding parts availability.
At the present time, Intel has given limited windows to its 10-nm and 7-nm roadmaps. The company said that 10nm ++ and 7nm would overlap in 2021 and that it would result in a 7nm GPU. Deliveries of Ice Lake in notebooks are expected to begin in June, and volume shipments by the end of the year. No timeline has been provided for office rooms and roadmaps that have leaked (which may not be accurate) indicate that 14 nm hang on the desk until 2020. With the launch of 7 nm by AMD in a few weeks, the hike risks Intel.
Updated (18/06/2019): There is reason to believe that if such an agreement is concluded – and nothing has yet been publicly announced – it could be the same type of contiguity products that Intel has sometimes put to the point with partners before. This type of allocation is the kind of maneuver we expect from Intel while trying to maximize the in-house manufacturing of the highest margin parts with limited foundry space.