How much should one pay for a chip or a component? Lytica, a Canadian supply-chain pricing analytics company, has the answer. Founded by former Nortel chief procurement officer Ken Bradley, who, like many others in the IT industry, was once bemused by component pricing, Lytica is transforming itself into a software-as-a-service (SaaS) company, helping OEM and EMS make well-informed deals when buying or selling.Digitimes recently had a talk with Bradley about the value that business intelligence, such as pricing information and well-thought-out planning, can bring to a company.Q: I guess many OEMs have used Lytica's services when deciding what components to use. What inspired you to create this company that offers such a unique service?A: I spent many years at Nortel Networks, managed to negotiate for some joint ventures for Nortel in Guangdong province in China and in Shanghai, and founded Lytica after I retired from Nortel. My background was semiconductor physics and developing components. After I helped Nortel set up some joint ventures in China, they made me chief procurement officer. I was brand-new in purchasing, leading 1,600 people around the world, doing buying, sourcing, negotiating, and component engineering. What I discovered was that nobody knew how much I should be paying for a component.At the time Nortel was growing really quickly. I think we had the largest market capitalization in the world at that time. My suppliers and my staff told me I had the best pricing in the world. I sent a circuit card, with all the components on it, to suppliers, or EMS, like Flextronics or Sanmina, and I would get back a card with prices for its materials and components. Many of those prices were cheaper than I was paying! So I asked my staff: "Yesterday you told me I had the best pricing in the world! How did this happen?" They would not know. It turns out no body knows. Pricing is a private and secret kind of thing. It is not necessarily confidential, but nobody would share their pricing because they think it is their strategic advantage. This really bothered me. I even went to good consulting firms. But after charging me like US$100,000, they would tell me, "Ken you got really good prices." They didn't know, either.When I retired, I formed Lytica in 2005 as a consulting company. Because we are not suppliers, I got to see what people paid for components. I thought: Wouldn't that be great if we can set up a new product line in Lytica? We can gather all the prices from companies that we dealt with. They would share their pricing, and Lytica could statistically analyze where each price fits in the ranking of each component.Q: What was your methodology?A: The idea is to tell how good somebody's spending on electronic components is compared to market. Lytica predicts when something doesn't seem to fit. For example, one company's prices on a component may rank the 10th position most of the time, but for specific parts and components they ranked at the 30th. Lytica would advise them to negotiate prices consistent with the 10th position for those specific parts and components. And 80% of the time they would get prices close to what we suggested. So, the method has proven to work.Lytica founded this service because everybody is having the same problem of not knowing how much they should pay for the same component. People think they have good pricing just like I did while I was at Nortel, but there is always opportunity.Lytica started doing this, and the database grew bigger and bigger to be the largest in the world of this type. So we hired more staff and came up with different ways of doing the modelling. The reality is, our model is based on the prices people are actually paying, and not by crawling on the web. We formed statistical distribution diagrams, knowing all the prices of the same components used by all companies in Lytica's database. When you come back with a price from a supplier, we can see where you fit in the ranking of that component's distribution, and if it's an outlier in your basket of components, Lytica recommends a price appropriate for you to renegotiate.Q: But there are so many factors that can influence the price.A: Yes. There are big companies and small companies. There are companies which are good negotiators, and those that are poor negotiators. Or those which have great business processes versus those that have not. And high volumes and low volumes. We counted some 60 factors that can influence the price, including freight, volume, channel mark-up, payment-terms, commodity spend and not so obvious factors such as financial risk, inventory reservation levels and geography, etc.There is no way you can come up with an equation for people to do something like price analysis. It is easier to calculate the cost, but that is not true for price. Market demand, negotiating skills, tariffs… you can list all of the things that would affect price. But it turns out cost has no correlation with price. You cannot figure out the price with all those 60 factors, either. Take weather for example, people have not come up with a perfect way to predict weather which needs fewer variables. So it is impossible to try using a bottom-up way of calculating the price. By characterizing how you perform in a marketplace Lytica takes it into account all of the things that you do such as being a skilled negotiator, having a good business practices, and all those other things you do. This characterization use statistical distributions to analyze what you're buying.All a customer needs to do is to give us their list of component pricings using a template. They can achieve great savings because we identify components that, for some reason, the prices are not normal for them. Our clients love this because that saves them time and money.Let me give you an example from the Nortel days. We were purchasing a strip of special metal which is about an inch wide and six inch long. Our development price from the supplier was US$1,100. When it went into production, nobody changed the price. Nortel was buying thousands and thousands of these things at the price which was supposed to be for new product introduction. Lytica is able to find those problems right away.Q: How can component and parts manufacturers in Taiwan or EMS companies make Lytica their tool to improve their operations?A: We started with many OEMs as customers because we didn't think the contract manufacturers would like our service as they're selling parts. But EMS do a lot of buying for the OEMs. Now EMS people come to Lytica, saying, "You've got to help us." Their customers demanded a lower price for components because they feel they are paying too much, but EMS have both sides of the equation. They have both the buy side and sell side problems. They were not always getting the best prices themselves. They now are using Lytica to test their prices and make sure they are not charging too much to their customers. So for EMS, they have advantage on both sides using Lytica.Lytica has more OEM customers than EMS. But EMS is the largest single market segment that we have, representing 30-35% of our businesses. Our smallest customer has US$5 million in sales a year, and the largest is among the largest corporations in the Fortune 500 or Global 2000 list. In terms of verticals, we've got customers from computer, automotive, medical, Internet of Things (IoT), etc. We are well-represented on the high end, and our goal is to capture the other 180,000 companies out there using electronic components.Q: Are any of those customers from Asia?A: Most of our customers are US and European companies. Some of the computer companies are from Asia. Some are from Taiwan. Pretty much every customer we have sources somewhere from Asia or the rest of the world. It is hard to get electronic components not associated with China, Taiwan, Malaysia, Vietnam and Singapore.Q: How big is your team? Do you have offices in Asia?A: There are 25 people in the company and growing rapidly, and are all located in Ottawa, Canada. If we wanted to set up offices in Asia, Shenzhen, Taiwan and Shanghai would be great places to do that.Lytica uses technologies so that we don't need to have remote offices everywhere. One of the things we need to have in order to do this job properly, is to have information on componentry. Some companies have offices set up in India or China, with 300 people to do component engineering and manage data information. Lytica does this using artificial intelligence technology. Around 80% of our people have technical degrees, such as science and engineering, mathematics. Lytica has PhDs who've worked with technology and research organizations. We've got a good mix of older people with production and distribution experiences, and young people with skills in AI and data analytics. Our employees speak 18 different languages. The diversity of the company forced us to venture on solutions in different ways than the ones currently adopted in the market.What we've done with technology is to read documents and organize information by use of machines. We do very sophisticated modelling. That really cuts down on our need to have a lot of people doing those tasks. On the sales front we're growing fast. We have transitioned from being a traditional consulting company into a world class software-as-a-service (SaaS) platform company.Lytica is expanding our global presence with marketing and sales initiatives. Asia is an important area of focus and we're bringing in a lot of business leads from there. The majority of our leads come from Asia, Europe, and the US. Our customer profile looks like an inverted pyramid. Most of our customers are huge companies with revenues over US$5 billion a year, and some of them are in the world's largest market cap group. We've got a very good position, because we receive a lot of quality data from those customers, and the opportunity to grow with these data. As the roadmap goes forward, we are able to increase our capabilities and help them get better pricing and improve their processes.The data that our customers give us are confidential, and we do not disclose customers' names. Some customers would mention that they use our service, and some would not.Q: Efficiency seemed to be the golden rule before Covid-19. However, the pandemic sems to have shifted the attention towards supply chain resilience and risk diversification. How has the pandemic affected your business?A: When Covid hit, everybody tightened up, because nobody knew what the impact's going to be and how long to last. We saw our sales go down, and customers delayed spending. There was this six-month period that nobody knew what's going to happen. In the last three or four months, all of our customers have started buying again, and we got new customers at a tremendous rate as we changed to a SaaS model. We actually are having our third record quarters in a row.Your comment about efficiency is interesting. I think efficiency is still the golden rule. Companies still need efficiency, but the problem is, they're not quantifying risk adequately. As they have been hit with risk from a couple of sources - one is the pandemic, one is the trade tension, one is the semiconductor shortage - so they are demanding diversity and resilience. From my perspective, resilience is a result of well-thought-out processes. So the companies that are succeeding are the ones with the most solid practices. Many times, I see a company having to pay more, it is because they did not forecast well enough to be able to ensure the supply they needed. Or they did not have adequate thinking into their sourcing strategy. When you are good at business intelligence and planning, you are in better shape. When the world was short of capacitors in 2018, we saw some companies which were getting solid pricing and supplies. They were the ones with the best planning systems. Other companies which have "I will be able to negotiate when I need it" mindsets were paying prices that are 2-3-4 times than they should in premiums just to get components. It is the practices that are the root cause that drive diversity and resilience. Companies need to have appreciations for diversity and use Lytica's SaaS platform.Lytica chairman Ken BradleyPhoto: Company

Third-party foundries will continue to play an important role in Intel's freshly unveiled IDM 2.0 strategy. TSMC is currently a major contract chipmaker for the US chip giant, and is expected to remain so in the foreseeable future. Chip shortages are worsening, spreading across many different market segments. Sercomm has disclosed networking chip lead times have now extended to a year. In Taiwan, some major companies, such as Compal and Asustek, are eyeing smart pole business opportunities. They have formed an alliance aiming to set smart pole standards for the worldwide market.TSMC to remain major foundry partner of Intel: TSMC has been a major foundry partner of Intel and will remain so in the foreseeable future, despite the ambitious foundry plan that the US chip vendor has just unveiled, according to industry sources.Delivery lead times for networking chips extend to 50 weeks, says Sercomm: Global chips shortages have spread to the networking segment from notebook and automotive sectors, with delivery lead times for networking chips from vendors already extended to as long as 50 weeks due to extremely tight foundry capacity, according to James Wang, president at Sercomm, a Taiwan-based broadband networking solutions provider.Alliance formed in Taiwan to promote 5G smart pole standards: A total of 25 companies have newly formed a 5G Smart Pole Standard Promotion Alliance, aiming to draft technological standards and tap huge smart-city business opportunities worldwide.

Intel has unveiled its so-called IDM 2.0 strategy, with a US$20 billion plan to build two fabs in the US to beef up its manufacturing capacity, plus a new business unit for providing foundry services. The move comes amid tight supply across many different semiconductor segments. Some analog IDMs have had to extend their delivery lead times to as long as 40 weeks for many segments in order to accomodate urgent demand for automotive applications. In the backend sector, demand for wire-bonding capacity is expected to stay robust throughout 2021.Intel announces US$20 billion fab expansion plans in foundry revamp: Intel has announced plans to invest about US$20 billion to build two new fabs in Arizona in line with what it calls its IDM 2.0 strategy, seeking to become a major provider of foundry capacity in the US and Europe to serve worldwide customers.Analog IDMs extending delivery lead times: Analog IDMs have prolonged their delivery lead times to as long as 40 weeks as they probably have given priority to producing automotive chips that have been in severe short supply, according to sources at Taiwan-based OEMs.Wire-bonding capacity to sustain full utilization throughout 2021: IC backend houses are expected to run their mainstream wire-bonding packaging lines at full capacity through the end of the year to meet strong demand for automotive and consumer chips, and their supply of flip-chip (FC) packaging remains constrained due to substrate shortages, according to industry sources.

Taiwan-based QT Medical and Lubn Taiwan have joined Qualcomm Advantage Network (QAN), with both participating in Qualcomm's Smart Cities Accelerator Program and focusing on developing IoT innovations.QT founder and CEO Ruey-Kang Chang pointed out that the company is mainly developing technologies for remote and home medical care and has developed solutions to bridge medical staff and patients with heart diseases.Lubn founder and CEO YC Chung noted that his company offers IoT and automation technologies for managing real estate including access control. As demand for managing real estate has been picking up, the company's automated services will be able to assist its clients to achieve their targets in a very short time, he said.

PatSnap, a Singaporean provider of innovation intelligence, has announced it has secured US$300 million in Series E funding, led by SoftBank Vision Fund 2 and Tencent Investment with participation from CPE Industrial Fund and existing investors Sequoia China, Shun Wei Capital, and Vertex Ventures.PatSnap said it plans to use the funds to advance its innovation intelligence platform, accelerate product development, and acquire additional domain expertise in the industry sectors where its technology is used by R&D and IP teams. The funds will also enable PatSnap to expand its sales presence around the world.PatSnap said its intelligence platforms use machine learning (ML), computer vision, natural language processing (NLP), and other artificial intelligence (AI) technology to provide innovation teams at many of the world's largest companies, global brands, and universities and research institutions with access to market, technology, and competitive intelligence as well as patent insights needed to take their products from ideation to commercialization. PatSnap said its clients include Dyson, Spotify, Oxford University Innovation, and Dow Chemical."PatSnap's mission is to empower innovators to make the world a better place," said Jeffrey Tiong, founder and CEO of PatSnap, as cited in a company press release. "Our global footprint, leadership, and strategic position in the innovation economy have enabled us to attract top investors, customers, and talent. Adding SoftBank Vision Fund 2 and Tencent to our notable roster of investors will help solidify PatSnap as the industry standard for innovation intelligence. Both have deep investment expertise with AI-led companies and proven track records supporting sustainable company growth."PatSnap said it has more than 10,000 customers around the world, supported by more than 700 employees working from the company's Asian headquarters in Singapore, European headquarters in London, and North American headquarters in Toronto.

Renesas saw one of its plants in northern Japan struck by fire and the incident could lead to a further tightening of global IC capacity for automotive chips, but Taiwan-based PTI is expected to see shifted orders from the Japan-based IDM as part of Renesas' backend bumping capacity was also damaged by the fire. At the same time, ASE and ShunSin are optimistic about their shipments for optical communication modules used in 5G device applications starting the second quarter.Fire at Renesas fab to further constrain auto IC supply: A recent fire that struck one of Renesas' plants in Japan may lead to a further tightening in the world's available production capacity for automotive chips, according to industry sources.PTI to offer Renesas additional backend capacity support: Backend house Powertech Technology (PTI) will provide automotive chips vendor Renesas Electronics with additional capacity support starting April, as the IDM's backend bumping capacity is constrained by a fire that has halted one of its chip factories in northern Japan, according to industry sources.ASE, ShunSin gearing up for optical module shipment boom: Backend houses ASE Technology and ShunSin Technology are both expected to enjoy a surge in orders for optical communication modules for 5G device applications starting the second quarter, according to market sources.

The penetration rate of Samsung Display's flexible AMOLED panels at the high-end smartphone segment in China in the first quarter of 2021 was higher than that seen a year earlier, Digitimes Research has found.Among Chinese vendors, only Huawei and its spun-off Honor and ZTE have opted to adopt flexible AMOLED panels rolled out by local panel makers. Others, including Xiaomi and Vivo, were primarily using locally made AMOLED panels to produce mid-tier models.Samsung Display's newly launched E4 AMOLED panels improve the brightness of the displays to 1,500 nits from the 1,200 nits found in the previous model and reach a contrast ratio of 5-million:1, which has enabled the Korean panel maker to continue to make headways in the China market, Digitimes Research says.Samsung Display has also lowered its rigid AMOLED panel prices, making those products more affordable than rival Chinese vendors' LTPS TFT LCD panels.Honor has launched its first smartphone, V40, after being spun off from Huawei. The Honor V40 features a MediaTek Dimensity 1000+ CPU, with its AMOLED panels coming from BOE Technology and Visionox.ZTE has been using AMOLED panels from Visionox since 2017. Visionox has managed to roll out full-screen AMOLED panels with in-display camera modules.In addition to using E4 AMOLED panels from Samsung Display, Xiaomi has also purchased flexible AMOLED ones from China Star Optoelectronics Technology (CSOT) for its updated Xiaomi 10S model.

Based in Montreal, Quebec, one of Canada's three artificial intelligence (AI) innovation clusters that host world-leading AI technology firms and talent, startup Dataperformers endeavors to incorporate AI technologies into the day-to-day operations of the manufacturing industry. The Dataperformers team specializing in deep learning and computer vision has decided to begin with end-to-end automated material and process defect inspection solutions with a focus on maximizing the corporation's ROI for its adoption of AI technologies so that Dataperformers' customers can reap the biggest AI benefits.Dataperformers started out focusing on joint development projects entrusted by customers and has gained ground in aerospace and electronics applications. Transitioning to targeting automated inspection solutions, Dataperformers names its new product Macula AI. Featuring easy installation and usability, Macula AI does not require the customer to employ an AI expert to make use of the system. Under the circumstances that AI talent is hard to find, this lowers the barrier for customers to incorporate Dataperformers' AI solution. Furthermore, Macula AI only needs a small amount of data for AI model training and algorithm creation to achieve industry-grade precision. This is a compelling advantage that Dataperformers' AI solution offers to customers.Macula AI has made significant progress in addressing challenges facing the manufacturing sector and engineering technologies. Successful use cases include the collaboration with Autodesk and Aisin Seiki, Japan's second-largest auto part supplier. Dataperformers recently began working with a disposable medical product manufacturer that seeks a perfect balance between product quality and production yield in view of its manufacturing process and brisk customer orders. As soon as it detects a defective product, Macula AI prompts the laser marking machine to mark the product. This allows the production line to keep operating uninterruptedly and the production equipment to keep churning out products at a high speed. The laser marking also accelerates the screening of defective products at the step of quality control sorting and inspection. Macula AI also includes an anomaly detection system that automatically notifies the site manager when discovering the product quality has started to deteriorate so that critical decisions can be made to stop unnecessary waste of materials.In view of Taiwan's pivotal role in the global electronics manufacturing sector and semiconductor supply chain, Dataperformers entered into a partnership with Advantech at year-end 2020, from which an outstanding product portfolio combining software and hardware systems is created. Leveraging Macula AI's appealing features such as the detection of scratches, stains, pores and other surface defects, Dataperformers has also initiated the search for potential business opportunities in Taiwan. Macula AI's expansion in Taiwan targeting low latency automated inspection and quality control applications based on cloud and edge computing technologies, coupled with private network infrastructure, has made eye-opening achievements.Canada-based startup Dataperformers CEO, Mehdi Merai

China's IC design sector saw out grow 24% to hit a record high in 2020, with the top-five vendors accounting for nearly 50% of the total, Digitimes Research has found.Chips related to communication, consumer devices and IT products contributed over 80% of IC output value in 2020, with communication applications alone taking up nearly 50%.The number of IC design houses is fast rising in China thanks to policy support and ample funding. Ten IC design firms went public in 2020, and most of them have been performing positively so far.Even so, China suffered a deficit of CNY200 billion (US$30.73 billion) in international trade of IC parts, a level which was on par with that seen a year earlier. The deficit is likely to remain at high levels in coming years, as it relies heavily on imports of high-end chips, while still focusing mainly on developing and producing entry-level to mid-tier IC products locally.

Supply in the semiconductor market has been so tight that prices in many segments, including foundry services, have been rising. Taiwan's pure-play foundry houses are now in contract talks with clients for 2022, and they are looking to introduce floating price schemes to make sure their pricing will reflect the volatility in costs, supply and demand. Tight capacity at foundry partners have undermined revenues gains at MCU vendors, but they are expected to see substantial sales growths starting in March and in second-quarter 2021. In China, memory makers CXMT and YMTC are keen to incease output that could raise their global market share significantly as early as year-end 2021.Pure-play foundries mulling floating price contracts for 2022: Taiwan-based pure-play foundries are already in talks with clients about contracts for 2022, and intend to implement a floating pricing policy amid tight capacity, according to industry sources.MCU makers poised to see significant growths in March and 2Q21: Taiwan-based MCU suppliers are likely to see their revenues increase significantly in March and second-quarter 2021 thanks to capacity support from wafer foundry houses and IC backend service firms, according to industry sources.China memory chipmakers gearing up for output ramp-ups: ChangXin Memory Technologies (CXMT) and Yangtze Memory Technologies (YMTC) are both looking to ramp up their chip output with substantial production yield rate improvements this year, and may start contributing meaningfully to the global memory industry bit output as early as year-end 2021 or 2022, according to industry sources.