AMD Archives - Counterpoint

Qualcomm Snapdragon X Elite Unveiled: ARM-based SoC for Windows-powered AI PCs

Qualcomm’s new Snapdragon X Elite SoC targets the AI PC market as the PC market bottoms out.
Made on TSMC’s 4nm process node, the new Snapdragon X Elite features a powerful 12-core Oryon CPU.
Between the Oryon CPU, Adreno GPU, and the Hexagon NPU, the Snapdragon X Elite can deliver up to 75 TOPS of AI computing performance.

Qualcomm announced its latest Snapdragon X Elite compute platform at the 2023 Snapdragon Summit held in Hawaii on October 24-26, during which the company also launched the Snapdragon 8 Gen 3 mobile platform. The Snapdragon X Elite is an ARM-based processor designed for personal computers and aims at delivering powerful performance and better power efficiency along with cutting-edge on-device generative AI features.

Qualcomm looking to expand its ARM SoC Smartphone Success to the PC Platform

Based on its brand-new ARM CPU core ‘Oryon’, developed from its Nuvia acquisition, Qualcomm’s Snapdragon X Elite SoC is built on TSMC’s 4nm process node. The CPU uses ARM’s 8.7 instruction set and features 12 high-performance ‘Oryon’ cores clocked at 3.8GHz. There is also a dual-core boost feature offering peak clock speeds of up to 4.3GHz. Qualcomm says the Snapdragon X Elite can offer 2x faster CPU performance than the competition while consuming one-third of the power.

Source: Qualcomm

Combined with 136GB/s LPDDR5x memory bandwidth, and up to a total of 42MB cache, including three 12MB L2 cache, and the remaining 6MB shared with the overall CPU. The SoC comes with an integrated Adreno GPU as well as Hexagon NPU and sensing hub to enable multiple AI functions. This is effectively moving AI inferencing to the network edge (laptop). The Adreno GPU supports DirectX, OpenCL, and Vulcan APIs.

Built for on-device AI experiences, the Hexagon NPU can deliver peak AI computing performance of 45 TOPS, which Qualcomm claims is 4.5 times faster than the competition. And between the CPU, GPU, and NPU, the SoC can offer peak AI computing performance of 75 TOPS.

The SoC was designed to run large language models on-device, with up to 13 billion parameters, thus offering the claimed fastest Stable Diffusion performance by a laptop chip on the market. It can generate 30 tokens per second for seven billion large language models (LLMs).

Source: Qualcomm

The SoC is designed in a way that it can be used on laptops, tablets, and even desktop PCs. Qualcomm showcased a 12W fanless reference design, along with 23W, 45W, and even the one with 80W thermal design power (TDP). TDP measures the maximum amount of heat produced by the chip in terms of watts.

PC vendors including Lenovo, HP, Dell, Microsoft, and Acer are all working to bring Snapdragon X Elite-powered laptops to the market by mid-2024.

WATCH: “AI PC” Era Beckons with Snapdragon X Elite: Deep Dive with Qualcomm’s Kedar Kondap

Snapdragon X Elite Benchmarks Show Impressive Gains over Apple Silicon, Intel and AMD Processors

At the Snapdragon Summit, we also had an opportunity to take a sneak peek at the Snapdragon X Elite benchmarking, where Qualcomm tested the SoC on both Windows and Linux (using Geekbench 6.2). Cinebench, UL Procyon AI, Wildlife Extreme, Aztec Ruins, and PC Mark were some of the other popular benchmarks on which the Snapdragon X Elite was tested.

Source: Qualcomm

The reference design laptops were loaded with some popular benchmarks to show how they performed over the competition. There were two TDP configurations — Config A laptop with 80W max and Config B laptop with 23W.

Below are some of the scores that Qualcomm shared:

Source: Qualcomm

ARM-based Laptops to Further Gain Share at x86’s Expense

Qualcomm’s Snapdragon X Elite SoC was built completely on an ARM IP structure, instead of the long-lasting x86 IP structure in the PC industry, which supports our view of the double-digit YoY percentage growth in ARM-based laptop shipments.

After Microsoft and Qualcomm’s exclusive agreement to develop ARM-based Windows-compatible chips expires in 2024, we are expecting more chip vendors to enter the market and that will make the PC market more competitive. Intel could face more challenges to its long-lasting dominance in the PC CPU market.

We believe there are still challenges for the ARM-based PC ecosystem and apps/software support. Software developers have spent decades and billions of dollars writing code for Windows that runs on the x86 architecture. Even if Microsoft smoothly migrates its Windows software portfolio to ARM-based processors, it could take a lot of time for the ARM-based ecosystem to see similar migrations and maturity.

AI PC to Drive Another Wave of Shipment Growth in 2024

In 2023, PC OEMs and chip vendors are all dedicated to developing products/solutions for the AI universe. Over the past five quarters, PC OEMs have been working hard to resolve inventory issues and had a hard time searching for a new growth engine for the PC business. Now, the AI PC market is witnessing a surge, underpinned by Intel and Qualcomm’s new PC CPU platform, which is just around the corner. These AI-enabled PC models will likely be available around mid-2024.

We now expect AI PCs to have an over 50% 10-year CAGR from 2020, and after 2026, they will dominate the PC market. Intel, Qualcomm, and other PC CPU makers are working closely with PC OEMs toward the next-generation mainstream models, marking a new chapter for the PC industry.

Data Center CPU Market: AMD Surpasses Intel in Share Growth

The global data center CPU market’s revenue declined 4.4% YoY in 2022.
AMD registered a 62% YoY growth in its data center CPU revenue to hold a 20% market share.
Intel’s data center CPU revenue dropped by 16% YoY in 2022, while its market share fell to 71%.
ARM-based CPUs gained traction with Ampere, Graviton (Amazon) and Yitian (Alibaba) to surpass $1 billion in revenues for the first time.

New Delhi, Beijing, Seoul, Hong Kong, London, Buenos Aires, San Diego – February 27, 2023

The global data center CPU market’s revenue registered a 4.4% YoY decline in 2022, according to the latest research from Counterpoint’s Semiconductor Service. Macroeconomic headwinds and increased energy costs impacted the sales of data center CPUs during the year. Besides, from the architecture perspective, the addition of accelerators in the servers for workloads restricted the demand for additional CPUs for servers.

Commenting on the companies’ 2022 performance, Senior Research Analyst Akshara Bassi said, “Even though Intel is still the market leader, its market share loss points to AMD’s rising product portfolio and better performance over Intel. AMD surpassed Intel in market share growth in 2022. Intel suffered due to continued delays in the release of its next-generation product Sapphire Rapids, generationally comparable to AMD’s Milan launched in 2021.

As demonstrated by hyperscalars AWS and Alibaba, ARM-based architecture chips continue to gain steam due to the ROI offered on varied workload deployments and off-the-shelf solutions from Ampere Computing, and shipping of data center CPUs from NVIDIA in H1 2023.”

Talking from the foundry perspective, Associate Director Dale Gai said, “As evidenced by wafer demand and foundry capacity of advanced nodes from TSMC, the total wafer sales at 5/4nm rose by 85% YoY in 2022. One of the demand drivers for the increased demand of advanced nodes is data center CPUs”

Market summary for 2022

Intel remained the market leader with a 71% share, although far from the share that it commanded till 2018. Its revenue from the segment dropped 16% YoY in 2022. The market share declined primarily due to delays in next-generation products and weakness in enterprise spending due to macroeconomic conditions.

AMD came second with a 20% market share primarily driven by increased adoption of its EPYC processor Milan. AMD is becoming a dominant force in the x86-based CPU for data centers, being increasingly adopted by cloud providers and SKUs of server companies. AMD registered a 62% YoY growth in its data center portfolio in 2022.

AWS’ in-house ARM-based chip Graviton, now in its third generation, has been among the early adopters of ARM architecture in a data center. AWS has increased Graviton’s penetration in its offerings and also expanded it to support ML-based instances with in-house accelerators, representing a shift from general-purpose compute to specific workloads.

Ampere Computing started to gain more traction in 2022 with its expansion from traditional cloud providers to enterprises by having its CPU in off-the-shelf servers from OEMs.

The comprehensive and in-depth “Data Center CPU Tracker” report is available. Please contact Counterpoint Research to access the report.

Background

Counterpoint Technology Market Research is a global research firm specializing in products in the technology, media and telecom (TMT) industry. It services major technology and financial firms with a mix of monthly reports, customized projects, and detailed analyses of the mobile and technology markets. Its key analysts are seasoned experts in the high-tech industry.

Analyst Contacts
Akshara Bassi

Dale Gai

Neil Shah

Counterpoint Research

press@counterpointresearch.com

report.counterpointresearch.com

@CounterPointTR

AMD Records Stellar Q1 2022 Numbers, Completes Xilinx Acquisition

AMD reported its best-ever quarterly revenue at $5.9 billion in Q1 2022, a 71% YoY increase driven by the cloud computing, automotive, gaming and data center segments. This revenue includes six weeks of performance of Xilinx, whose acquisition was successfully completed by the company during the quarter. Xilinx is an FPGA and networking compute solution provider. Excluding Xilinx, AMD’s Q1 2022 revenue was $5.3 billion, a 54.7% YoY increase. We give below some of the key takeaways from AMD’s Q1 2022 earnings statement:

The gross margin was at 48%, up 190 bps YoY.
Revenue for the Computing and Graphics segment was $2.8 billion, up 33% YoY.
Revenue for the Enterprise, Embedded and Semi-custom segment was $2.5 billion, up 88% YoY.
The company intends to acquire Pensando, a distributed services platform provider.
Wafer shortage for 16nm and below will remain constrained till H2 2022.

All values in USD Million

Client Compute revenues
The company launched seven new CPUs for desktop processors and first desktop processors with 3D chiplet architecture. In the notebook segment, the company extended the mobile processors to cater to customers across the spectrum, including gaming and creators, with features like better battery life and compute synergies.

Desktop GPUs grew significantly owing to mobile GPU launches, but data center GPUs, including the new launch of Instinct MI210 accelerators, remained flat.

The revenues across the client compute PC market will see a correction in the next quarter, driven by inflationary pressures and dampened demand.
The company will focus more on premium, high-margin segments as the PC market softens in the next quarter.
GPU demand will also remain muted due to lowered prices and improved supply in the market. Weakening crypto markets will also contribute to the growth pace of the product portfolio.

Enterprise, Embedded and Semi-custom segment

Gaming console, automotive and data center units were the growth drivers for this segment, including a new launch by Valve in the gaming console space. The company also launched its first processors for the cloud and HPC markets, featuring 3D-stacked chiplets architecture.

The Data Center Compute segment will continue to drive growth as hyperscalars continue to expand cloud infrastructure and enterprises upgrade their IT infrastructure.
In networking and communication, due to enhanced synergies with the Xilinx portfolio, the company’s FPGA-as-a-Service and Smart NICs will help in penetration and diversification of the market to mainly telecom and fintech players.
The company will focus more on AI workload-based GPUs and expanding its software stack.

TSMC’s Quarterly High-performance Computing Sales Surpass Smartphones for First Time in Q1 2022

TSMC reported its Q1 2022 revenue at $17.6 billion (up 12% QoQ). Among its different segments, the high-performance computing (HPC) segment posted the highest growth rate at 26% QoQ.
HPC accounted for 41% of TSMC’s Q1 2022 revenue, while the smartphone segment accounted for 40%. It is the first time in TSMC’s history that HPC has become the largest sales contributor during one fiscal quarter.
TSMC’s major HPC customers in 2021 likely included AMD (including Xilinx), Nvidia, Broadcom, Marvell and Intel besides Apple as the top-selling one due to its Apple Silicon (M series) in PCs and tablets.
Intel appeared to be the next major HPC customer by outsourcing more graphic silicon (discrete Arc series and iGPU tiles) and other CPU cores this year. Accordingly, it would be the second-largest customer in TSMC’s 3-nanometer node in 2023.
TSMC’s HPC segment is expected to account for 42% of its total revenue in 2022, including wafer and 2.5D/3D packaging services.

Before TSMC’s investor conference on April 14, semiconductor industry experts were waiting for the company to tune down its annual growth outlook due to weakening consumer sentiment on consumer electronic products. Counterpoint projects the global smartphone sell-through units to decline by over 10% YoY in Q1 2022, impacted by a slowing replacement cycle, geopolitical uncertainties, COVID-19 lockdowns in China and inflation concerns. Recent data on smartphone order cuts implies the risk for foundry wafer demand in the next few quarters and the previous IC shortage concerns have been improved.

Surprisingly, TSMC beat market expectations not only in its Q1 2022 financial results (with gross margin improving to 56%), but the upside of full-year sales guidance also approached 30% YoY growth. While the demand softness from smartphones has been known, TSMC believes the momentum of HPC and automotive businesses is getting stronger in 2022 and will offset the uncertainties from consumer electronics. Notably, TSMC reported $7.2-billion HPC sales in Q1 2022, exceeding the sales from smartphone clients to become the largest business of the company for the first time in history.

Other than retaining its revenue outlook for 2022, TSMC reiterated its capital spending budget at $40-$45 billion during the year, with a focus on leading-edge technology nodes (5 nm and below). The longer lead time of equipment/tool production becomes more challenging in the light of labour crunch and IC component shortages (like FPGA) from global suppliers. But TSMC has demonstrated its solid execution in ramping up new fabs as the capacity expansion plans are largely unchanged.

HPC silicon demand is on a multi-year, structural growth trajectory

The silicon demand of HPC-related ICs is well above that of the end devices and equipment. In 2022, global PC shipments are likely to decline 0%-5% YoY. The server shipments growth rate is also expected to be modest (5%-8% YoY). At the same time, 5G infrastructure growth is not accelerating in many countries, with the growth cycles of edge computing or open RAN being at the initial stages.

So, why TSMC and its HPC customers are expected to deliver a very robust growth outlook in the next 1-2 years? The answer can be found in the mega trends of digital transformation since COVID-19, which have witnessed a boost in capital spending mainly by US-based cloud service providers (CSPs) expanding their global data centre footprints. The total capex by global CSPs, based on Counterpoint estimates, may increase 23% YoY in 2022. In the three years to 2025, the annual compound growth rate (CAGR) will be in double digits. The acceleration in mega data centre buildouts is for AI services, Metaverse and autonomous driving in need of the most advanced semiconductor components. The AI accelerator IC market, which has been dominated by Nvidia, is a typical example of the market where the demand CAGR is likely to exceed 30% in the next few years as GPU, CPU, FPGA and ASIC compete for AI training and interference applications.

All main processors in server and data centres today apply foundry nodes below 10 nm. As computational performance improves significantly and demands more transistors, the roadmap of node migrations beyond 10 nm will get accelerated in the HPC semiconductor industry. In 2022, both AMD and Nvidia’s new graphic processors will move to 4 nm and 5 nm nodes. Further, we expect the next phase of 3 nm generation processors to start to ramp up after 12-18 months. Besides, the increases in larger die size in processors implies more wafer demand and higher dollar content in advanced packaging technologies that TSMC offers for complete solutions.

Source: TSMC financial reports

In our view, the HPC semi demand in foundry remains the seller’s market in the next few years, as TSMC captures over 90% of market opportunities, including Intel’s CPU outsourcing orders. Competition in the HPC semi market from the global top three chipmakers – Intel, AMD and Nvidia – is intensifying by entering new areas, like Intel’s entry into the discrete GPU market and Nvidia’s plan for ARM-based server CPUs, leading to strong wafer demand in advanced nodes.

TSMC offers the leading-edge technology to enable these HPC applications to pursue the best PPAC (performance, power, area and cost) over each technology geometry. This has become TSMC’s fastest-growing business segment in the past three years from the revenue perspective. In Q1 2022, it reported HPC business sales reaching $7.2 billion with nearly 60% YoY growth, overtaking smartphones as the company’s largest segment. While TSMC did not disclose any client information, we believe AMD, Nvidia, Broadcom, Marvell and Intel were the major traditional HPC customers during the quarter, with Apple as the top-selling one due to its Apple Silicon (M series) in PCs and tablets.

5/4 nm is the mainstream today, 3 nm to ramp up from 2023

The reason why TSMC’s revenue is increasing from the HPC segment is that the majority of wafers in HPC applications apply the leading-edge nodes (from 7 nm and 6 nm to 5 nm and 4 nm in 2022). Wafer prices at these nodes are multiple times higher than the foundry industry ASPs due to more expensive equipment tools. During Q1 2022, the mainstream of HPC ICs was at 7/6-nm, including discrete GPUs, data centre GPUs (AI accelerators), x86 CPUs, ARM CPUs for client devices (tablets and laptops) and specific processors like crypto mining ASICs. TSMC will also start to ramp up multiple new CPUs/GPUs at 5/4 nm in H1 2022, leading to its robust wafer demand likely through the end of the year.

Based on our estimates, in 5/4-nm nodes, the total wafer demand from HPC products (see Exhibit 2) will account for 37% of the total foundry (excluding IDM) industry capacity in 2022. Smartphone AP/SoCs will remain the largest portion of wafer demand (over 50%) in this node during the year. With new product cycles from Nvidia and AMD in H2 2022, more wafers will be procured into 2023 as HPC might reach half of the total wafer capacity at 5/4 nm in the next year.

During its Q1 2022 financial result conference, TSMC reiterated 3 nm would enter mass production in H2 2022, with wafer sales contribution from the beginning of 2023. As we all know that the new iPhone application processor (A16) will adopt 4 nm instead of 3 nm due to initial supply constraints from TSMC, HPC products will be the main body of 3 nm in H1 2023, especially the demand from Intel’s wafer outsourcing in its new CPU line-ups. We believe Intel would be the second-largest client after Apple on TSMC’s 3 nm in 2023 as the volume of outsourced wafer orders is highly subject to Intel’s in-house EUV availability in the early stage. HPC applications will be a stable part (Exhibit 3) in 3 nm with 40%-60% of total foundry wafer demand during 2023-2025.

Conclusion

The crossover point between HPC and smartphone sales at TSMC, observed in Q1 2022, is more like the beginning of the future trend where semiconductors will shift growth drivers from mobile devices to multiple computing focus areas for the next wave of technology applications. We estimate TSMC’s HPC segment to account for 42% of its total revenue in 2022, including wafer and 2.5D/3D packaging services.

AMD in 2021: A Year of Great Profits, Market Gains

AMD, one of the world’s leading chip companies, has reported a 49% YoY growth in its Q4 2021 revenue to $4.8 billion while increasing its gross margin by 560 basis points from the same period in 2020. For 2021, its revenue grew 68% YoY to $16.4 billion and the gross margin increased by 370 basis points due to strong data center revenues and high-end processor sales. Data center business contributed to ~25% of AMD’s overall revenues for 2021.

Company Strategy for 2021
The revenues reflect the company’s strategy to focus on high-margin data center products and high-end Ryzen processors for 2021 as semiconductor shortages persisted throughout the year along with supply chain hiccups.

Source: Company Earnings, Counterpoint Estimates

Annual Segment Updates:

Computing and graphics

The segment’s revenues grew 45% YoY to $9.3 billion in 2021. A 57% increase in the average selling price (ASP) and a higher mix of the latest Ryzen processors along with high-end GPU cards drove the segment’s revenues and profitability to record levels since 2011.

Client launches in 2021

AMD launched the Ryzen 5000 series and Ryzen PRO 5000 series mobile processors powered by the Zen 3 core architecture for business laptops built on 7 nm.

AMD expanded its graphic cards portfolio by launching the Radeon RX 6700 XT built on 7 nm process technology and AMD RDNA 2 gaming architecture, as well as the Radeon RX 6600 XT graphics card that delivers a high frame rate, high fidelity and highly responsive 1080p gaming experience. In mobile graphics, Radeon RX 6000M series was launched for high-performance gaming laptops.

Enterprise, Embedded and Semi-Custom

The revenue grew 113% YoY to $7.1 billion due to EPYC processor and MI200 accelerator sales in HPC (high performance computing), and Cloud and Supercomputer wins. Besides, demand from gaming consoles, automotive and IoT contributed to the segment’s growth.

Data center launches in 2021

AMD launched the Instinct MI200 series accelerators based on the 2nd Gen AMD CDNA architecture, optimized for HPC and AI/ML (artificial intelligence/machine learning) workloads. Next-generation AMD EPYC processors – AMD EPYC 7003 series CPUs boasting 64 Zen 3 cores per processor for HPC, cloud and enterprise customers, were also launched.

Analyst takeaways and company outlook for 2022

AMD estimates a 31% YoY annual revenue growth for 2022 to ~$21.5 billion.
Data center revenues to be ~30% of overall revenues.
Xilinx acquisition creates synergies for its embedded portfolio for other industry segments including IoT, automotive and networking.

AMD will gain market share in the server segment due to the launch of Genoa, its next-generation EPYC processors that feature 96 Zen 4 cores and next-generation memory and I/O technologies. Besides, the ‘Infinity Fabric’, which provides greater efficiencies by connecting CPU and GPU, will play a pivotal role in AMD’s market share considering the accelerator portfolio has established itself as a catalyst in AMD’s data center segment.

The diversification through the successful acquisition of Xilinx has unlocked opportunities for AMD to capture the demand from the networking industry and create an end-to-end solution portfolio for specialized workloads.

For the client side, we see demand correction as the COVID-19 pandemic-led demand subsides in the PC market. However, the compute refresh cycles in enterprise and increased demand from premium ultrathin/thin laptops will offset some of the demand correction, as indicated by the launch of its processors and SKUs at CES 2022.

Samsung Takes Semiconductor Crown From Intel in 2021

London, Hong Kong, Boston, Toronto, New Delhi, Beijing, Taipei, Seoul – January 28, 2022

Semiconductor industry went through significant structural changes in 2021 after entity lists were announced by both China and the US. The overall 2021 semiconductor revenue rankings also varied from the previous year. Samsung took the first position from Intel thanks to its solid growth momentum in both logic IC and memory segments. Memory vendors continued to lead the industry with SK Hynix and Micron taking the third and fourth positions, followed by IC design vendors, including Qualcomm and NVIDIA.

During 2021, many top-tier semiconductor companies reiterated material changes happening in the industry, such as semi content growth, higher inventory level and longer chip lead time due to global component shortages as well as logistical issues. The year saw 19% YoY revenue growth with the largest contribution coming from the memory and IC design sectors.

Semiconductor Industry Top Players Revenues & Rankings – 2021

Samsung took the lead in 2021 with a strong DRAM and NAND flash market performance at the expense of Intel’s relatively flattish results. Major smartphone SoC and GPU vendors also enjoyed strong growth in the year with over 50% YoY revenue increase. In addition, we saw 27% YoY revenue growth among the top 15 vendors, outperforming global semiconductor revenue growth and implying another year of centralized semiconductor industry.

In general, we believe supply constraints will likely persist before H2 2022, though our checks suggest some amount of component shortage easing. Looking ahead, foundries are adding new capacities in 2023 and most of them hold optimistic views on their partnerships and utilization rates even if supply and demand normalize in the foreseeable future. High performance computing, metaverse (AR/VR/XR), 5G and automotive remain key semi content growth drivers for the industry.

Disclaimer: This ranking consists of chip companies. Foundry revenue is not included to avoid double counting.

Background

Counterpoint Technology Market Research is a global research firm specializing in products in the TMT (technology, media and telecom) industry. It services major technology and financial firms with a mix of monthly reports, customized projects and detailed analyses of the mobile and technology markets. Its key analysts are seasoned experts in the high-tech industry.

Analyst Contact:

William Li

Follow Counterpoint Research
press(at)counterpointresearch.com

CES 2022: Chip Biggies See Loads of Action

Against the backdrop of COVID-19 and component shortages, the recently concluded Consumer Electronics Show (CES) kept its date with announcements and launches in the processor ecosystem. We saw multiple releases and announcements from AMD, Intel and NVIDIA for both CPUs and GPUs. Trends that stood out included:

Focus on mobile processors — enabling AAA 1080p gaming in mobile/ultra-mobile form factors.
Focus on device security by building hardware protocols.
Focus on software and hardware offerings that accelerate performance.

AMD: Mobile processors with integrated security, budget GPU

CPU launches

Notebook segment: AMD launched a new series of its Radeon RX 6000 mobile processors that are built on TSMC’s 6nm and have Zen 3+ as core architecture, RDNA2 graphic architecture with LPDDR5, DDR5, Wi-Fi 6E and USB 4.0 support. The processors achieve a clock speed of up to 5GHz, which AMD claims gives up to 69% faster video editing, 125% faster 3D rendering performance and double the 1080p gaming performance when compared to the Ryzen 5000 series. Additionally, these processors are the first ones to integrate Microsoft Pluton Security Processor that helps in eliminating attack vectors and protecting critical data.

AMD also introduced new power management features and an adaptive power control framework to improve the power efficiency of the new processors, promising up to 24-hour battery life, 30% less power consumption during a video conference and 15% less while web browsing.

Desktop segment: AMD launched the Ryzen 7 5800X3D featuring AMD 3D V-Cache technology, claiming it to be the fastest desktop gaming processor. It also teased Ryzen 7000 Series processors, the next-generation “Zen 4” core-based processors built on the 5nm process, to be released later in 2022. These processors promise to be the next big thing in AMD’s kitty since 2017 as they carry major architectural changes, like a pivot to LGA (land grid array) as compared to BGA (Ball Grid array), 5 GHz clock rate for all cores and PCIe 5 support.

GPU launches

Notebook segment: AMD launched a new series of discrete GPUs, Radeon RX 6000S, for the notebook and ultra-slim form factors. The series is purported to achieve up to 100 fps (frames per second) for the ultra-thin mobile gaming segment. AMD also expanded the Radeon RX 6000M series for its extreme gaming laptop to offer entry-level and mid-range options.

Desktop segment: For the GPU category, AMD launched its budget-friendly GPU RX 6500 XT at $199 with a 2.6 GHz game clock and 16MB infinity cache. It also launched the RX 6400 graphic card to bring 1080p gaming to the mainstream market. This can be considered as AMD’s effort to establish a stronghold in the GPU category and its acceptance of FSR technology (AMD FidelityFX™ Super Resolution).

Software announcements

AMD Software Adrenalin Edition will be upgrading its platform to enable low-latency and high-fidelity gaming. To be released in spring 2022, the upgrade includes AMD’s Radeon Super Resolution (RSR), a spatial upscaling technology to enhance gaming experiences at native resolutions. Additionally, it will feature AMD Link 5.0, an application that allows users to play their PC games on a phone, tablet, or Windows PC from virtually anywhere.

Intel: 20+ new mobile processors, 50 processors in total including discrete GPU

(Credit: Intel Corporation)

CPU launches

Notebook segment: Intel released 22 new 12^th generation processors, including a new P series of ultra-portable mobile processors. These bring dedicated levels of performance, earlier available in desktops and extreme gaming laptops, to thin-and-light laptops. Developed on the big.LITTLE architecture with both performance and efficiency cores, this single, scalable SoC architecture is built on Intel 7 process and supports DDR5, LPDDR 5, Wi-fi 6E and USB 4.0. The P series works as an intermediate offering between the H series of laptops, aimed at extreme workloads, and the U series, aimed at the ultra-portable segment. According to Intel, the H-series processors are the fastest mobile processors ever, outperforming even the Apple M1 Max in content creation.

The interesting thing about these 12^th generation mobile processors is the implementation of the big.LITTLE architecture ARM style, as seen in the Alder Lake Desktop CPUs last year. The major upgrade comes due to the number of cores Intel is offering for the CPUs. The H and P series have 14 cores in total, whereas the U series has 10 cores, compared to 8 in the H series and 4 cores in the U series in the 11^th generation.

Desktop segment: Intel launched 22 new processors (both in 65W and 35W), unveiling the complete range of 12^th generation CPU processors that have clock rates up to 5.1 GHz and support DDR5 memory. Besides, it launched a new cooling system, Intel® Laminar Coolers, which accompany the new 65W processors.

GPU launches

Billed as its comeback in the discrete GPU market, Intel announced shipments of Intel Arc graphics (codenamed Alchemist) with 50 new mobile and desktop customer design wins. Intel Arc graphics offers industry-leading advanced features such as hardware-accelerated ray tracing, Xe Super Sampling (XeSS), AI-driven upscaling technology and Intel Deep Link technology. The Intel technologies help in intelligently routing power between CPU and GPU processing engines to boost performance by distributing workloads across multiple engines. This does not represent the introduction of any new technology but Intel’s ace execution of a strategy to gain market share.

Automotive launches

Intel Mobileye announced a new system-on-chip (SoC) that is purpose-built for autonomous vehicles (AVs). The EyeQ Ultra is built on Mobileye’s industry-leading EyeQ technology and brings the work of 10 EyeQ5 SoCs in a single package [176 tera operations per second (TOPS)] to deliver the technology required for a fully self-driving vehicle. The interesting thing is that it is fabbed on 5nm and is built on RISC-V cores.

Mobileye also announced a collaboration with Geely’s Zeekr brand to launch the industry’s first Level 4 consumer AV, expected to begin production in 2024. It also announced the launch and shipment to customers of the industry’s first fully capable Level 2+ vehicle equipped with a 360-degree surround-view sensing system and driving policy for the industry’s most advanced longitudinal and lateral control.

Software announcements

Intel announced the launch of its 3^rd generation Intel Evo platform that meets the specifications and key experience indicators set by Intel’s Project Athena Innovation Program. The platform enables better responsiveness, battery life, instant wake function, fast charge, and intelligent collaboration. The intelligent collaboration aims to deliver enhanced experience through video-conferencing apps by leveraging AI-based background noise cancellation, integrated Intel Wi-Fi 6E (Gig+), Intel® Connectivity Performance Suite6, and optional AI-accelerated camera imaging effects.

Intel vPro platform

Intel announced an upgrade to its vPro platform that provides hardware-based security options to consumers with two variants – Enterprise and Essential – to cater to all businesses.

NVIDIA: New GPU cards for laptops, adoption of Drive Hyperion platform

GPU launches

Notebook segment: NVIDIA launched new GPU cards GeForce RTX 3080 Ti and RTX 3070 Ti for laptops for the first time. Aimed as extreme and mid-level performance variants, these GPUs feature 16GB GDDR6, which is a first for laptops, and promise to deliver higher performance than the desktop TITAN RTX. NVIDIA claims that these perform average seven times faster than the latest MacBook Pro 16 M1 Max in 3D renders.

NVIDIA also launched a budget GPU – RTX 3050 ($249), which allows consumers to play ray-traced games on a 50-class GPU at over 60 fps for 1080p gaming. This GPU is a bit above AMD’s 6500XT, which is more comparable to its GTX 1650. NVIDIA also teased its RTX 3090 Ti discrete graphic card, the most powerful discrete desktop graphic card till now.

Software announcement

NVIDIA announced the adoption of its Drive Hyperion platform that has 12 state-of-the-art surround cameras, 12 ultrasonics, 9 radars, 3 interior sensing cameras and 1 front-facing lidar. NVIDIA Drive Hyperion is computer architecture and sensor set for autonomous vehicles that are open to all. NVIDIA announced Desay, Flex, Quanta, Valeo and ZF as its platform scaling partners and Polestar, IM Motors, Li Auto, NIO, R Auto and Xpeng as the companies adopting the platform.

NVIDIA Studio: NVIDIA Studio, a platform for creators, has expanded its support for NVIDIA Omniverse to enable real-time simulation of complex 3D workflows.

Analyst takeaways

This year’s launches from chip companies focused on:

Introducing chips in an ultra-mobile notebook form factor that can offer desktop-level performances with the least compromise on battery life.
Product offerings spanning the whole spectrum from entry-level and mid-level to extreme performance focusing primarily on mobile/ultra-mobile form factors.
Making GPU affordable and consequently 1080p gaming.
Comprehensive software suites and upgrades to extract maximum efficiency from the hardware while harmonizing with the OS to enhance user experience and collaboration across apps.
A sneak peek into the wonders of the omniverse and, more importantly, how to ‘use’ it.

This year will continue to raise the bar as we await the release of next-generation CPUs and GPUs in the second half. The companies’ enthusiasm towards making technology accessible in all form factors and budgets opens the door to mainstream adoption of the fundamentals of the omniverse.

Foundry Industry’s Robust Revenue Growth to Continue in 2021

The semiconductor industry’s foundry sector delivered above-expectation revenue growth in 2020. With tightening supplies from most global vendors, we expect 2021 to continue this momentum.

In addition to a favorable macro environment, such as COVID-19 leading to logistical challenges and trade tensions prompting increase in wafer bookings, the technology migrations in leading-edge nodes (7-nanometer and 5-nanometer) appear to be accelerating to meet the demand from 5G smartphones, game consoles and AI/GPU in cloud servers. On the other hand, new capacity additions across the industry remain rational, with the second-tier foundry vendors preferring to raise wafer prices against building greenfield fabs.

Here we highlight our four predictions for the global foundry industry in 2021, based on our bottom-up analysis and surveys:

Double-digit YoY sales growth again in 2021

In 2020, the foundry industry revenue reached about $82 billion, representing a 23% YoY growth. Despite this high base of 2020, the double-digit growth will persist in 2021. We forecast a 12% YoY growth with a total revenue of $92 billion.

We expect TSMC, Taiwan’s leading foundry service company, to keep outperforming the industry by posting 13%-16% YoY sales growth in 2021. TSMC is scheduled to hold its quarterly result conference in mid-January.

To be consistent with Samsung’s public information, our foundry forecast includes Samsung Foundry’s internal business (to LSI). We expect Samsung Foundry, driven by more order wins from external customers such as Qualcomm and Nvidia, to post a 20% YoY revenue increase in 2021.

For the overall industry, the double-digit growth in 2021 consists of both wafer shipment increase and like-to-like wafer price (ASP) increases, something which we rarely saw in previous cycles. In particular, 8-inch foundries, which have been reporting supply shortages from H2 2020, are acting as a catalyst in convincing some suppliers to raise their average wafer price by 10% in 2021.

Significant ramp-ups in 7/5nm by largely adopting more EUV layers

Both TSMC and Samsung will beat the average industry growth rate due to their accelerating production ramp-ups of EUV-enabled nodes (7-nanometer and 5-nanometer, or 7/5nm) after passing through the initial stage of learning in 2019 and 2020. The EUV (extreme ultraviolet lithography) adoption is a critical factor in extending the Moore’s Law to consistently increase the transistor density of chips to enable the development of both 5G smartphones and HPC (high performance computing) applications.

5nm: TSMC started its 5nm mass production from Q1 2020, and Samsung followed after 6-9 months. 5nm is considered a fully adopted EUV node for both foundries as Intel’s equivalent 7nm announced another delay in production last year. Based on our estimates, the total wafer shipment volume of 5nm will account for 5% of 12-inch wafers in the global foundry industry in 2021, up from less than 1% in 2020. Apple is the top customer (with all orders to TSMC) in 5nm this year (see Exhibit 1), including both for iPhones (A14/A15) and the newly released Apple Silicon. Qualcomm will be the second-largest 5nm customer as the iPhone 13 may adopt its X60 modem. TSMC is expected to book $10-billion revenue from 5nm in 2021. Samsung Foundry will also gain good traction from 5nm order wins, including its in-house (Exynos) SoC and Qualcomm. In our view, the capacity utilization rate will reach an average of 90% for TSMC and Samsung in 2021, with the upside from stronger flagship 5G smartphone models.

Exhibit 1: 5-nanomater Wafer Shipment Breakdown by Customer, 2021

7nm: Different from 5nm, with over 80% wafers used in smartphones, the 7nm applications are more diversified into AI/GPUs, CPUs, networking and automotive processors. TSMC has a variety of 7nm (DUV only), 7nm plus (with EUV) and 6nm (with EUV) in its 7-nm family, while Samsung has introduced 7nm/6nm with both adopting EUV production. Based on our estimates, the total wafer shipment volume of 7nm will account for 11% of 12-inch wafers in the global foundry industry in 2021. In this geometry, smartphones will only consume 35% of wafers (see Exhibit 2) and the majority will ship to AMD (27% of 7nm shipment volume) and Nvidia (21%). In the light of stronger demand for game consoles, cloud server/AI processors and mainstream 5G smartphones, the capacity for 7nm looks extremely tight through the whole of 2021, with the average utilization rate at 95-100% based on our calculations. Therefore, for emerging demand such as crypto-mining ASIC and ARM-based processors (in server and auto), the chipset vendors and OEMs will find it difficult to get allocation for extra capacities in the near term.

Exhibit 2: 7-nanometer (N7, N7+, N6) Wafer Shipment Breakdown by Customer, 2021

A high level of chip inventory becomes normal through the year

Although the foundry industry is less cyclical than that for IC memory, we still consider inventory level as one of the major factors to predict growth. However, we need to reset the expectations for 2021 or even for early 2022 as long as COVID-19 and global trade tensions persist. Global OEMs, as well as their IC vendors, are willing to prepare for extra levels of components. The procurement lead time of some standard ICs in the channel, according to the component supply chain, has lengthened to 26 weeks (6 months) and above from late 2020. In other words, we might see an ascending wave of double-booking pattern in matured nodes not only on TSMC but also more seriously on the second-tier foundry vendors.

By the end of Q3 2020, based on the financial reports from major global IC fabless companies, the inventory was about 79 days compared to the industry average of 70 days since 2016. If we see recent comments from AMD, Nvidia, Qualcomm, or even smaller players like Dialog Semi (analog ICs), the prospects of 2021 are generally upbeat on 5G smartphones, WFH devices and cloud server spending. The momentum supports high inventory levels. As long as the concern of supply chain disruptions persists, chip vendors would maintain a high level of inventory from Q4 2020. The argument indicates a better-than-normal seasonality in H1 2021, since foundry customers would choose to place wafer orders earlier to avoid capacity risks in the second half. We expect the industry inventory to hit the peak during the middle of this year.

Capital intensity also maintains high level, given upside from wining IDM shares

While we will wait for ASML’s EUV shipment forecast and TSMC’s capital expenditure (Capex) guidance for 2021 (both expected in mid-January) as the key bellwethers of global semi outlook, the market expectation of TSMC’s record-high (over $20 billion) number in 2021 looks reasonable to us. In our view, it will be a crossover year for TSMC’s sales between its two growth pillars – smartphone and HPC.

We see Intel going in for CPU outsourcing soon for its long-term survival. Both TSMC and Samsung are ready for the opportunity by potentially starting to expand their 5nm/3nm capacities during the year. It looks like the capex-to-sales ratio, as an indicator of future revenue growth confidence for chipmakers, will maintain peak levels in 2021 – 40% for TSMC and 70% for Samsung Foundry. Aside from the strong demand of end applications, which we discussed above, IDM outsourcing trends are accelerating for global IC vendors pursuing profitability. Not only from Intel, TSMC is likely to see growing orders from Japanese customers including Sony’s CIS (48MP) and Renesas’ MPU (at 28nm) during late 2021.

Conclusion

The upcycle in foundry industry during 2020-21 contains both multiple cyclical and structural positive drivers, which we have not seen since the tech bubble years (1998 -2000). The incremental demand of foundry industry will come from most sub-sectors, including automotive, to support its double-digit growth in 2021. Moreover, by factoring in IDM outsourcing, we expect the industry to reach and exceed $100 billion of market size during 2022-23.

AMD Q3 2019: Highest Quarterly Revenue in a Decade Driven by Ryzen Shipments

AMD reached its highest quarterly revenue in a decade, at US$ 1.8 Billion. AMD’s recently launched EPYC (2^nd Gen) processor managed to capture business from several major OEMs and cloud providers. In addition, AMD benefitted as the desktop and laptop processor segment grew 36% Y/Y, while Intel’s market position declined. However, Enterprise, embedded, and semi-custom segments declined 27% despite the strong performance of the 2^nd generation EPYC processor. The major reasons for this were declining demand in-game console chips as well as shipments of EPYC processor not meeting expectations due to the US-China trade war. In the coming quarters, AMD is expected to grow market share in both the client processor and data center segment.

Counterpoint’s View on AMD’s Q3-2019 Earnings

AMD’s 2^nd Gen EPYC processor has attracted a lot of business; many server OEMs and cloud providers are now powering their servers and data centers with AMD 2^nd Gen EPYC processors.
AMD will likely take market share from Intel in the data center segment, due to and lower cost. AMD is expected to double up its share in the data center processor segment in 2019.
AMD’s main competitor, Intel, has reported a decline in its client processor sales, in which AMD has gained. AMD is taking share from Intel’s in the desktop and laptop processor segment. AMD is taking advantage of the CPU shortages of Intel.
Due to multiple customers win, AMD’s EPYC and Ryzen families of processors will be the driving force for increasing revenues in the coming quarters.

Key Highlights from AMD:

Partnerships and Customers Wins

Google, Amazon AWS, IBM Cloud, Microsoft Azure, OVHcloud, Twitter and Tencent have all selected 2^nd Gen AMD EPYC processors based on leadership performance and total cost of ownership.
Dell, HPE and Lenovo more than doubled their server portfolio with 2^nd Gen AMD EPYC processor platforms.
Microsoft announced that its new 15-inch Microsoft Surface Laptop 3 will be powered by an AMD Ryzen mobile processor.
HP unveiled its first AMD-powered gaming laptop, the Pavilion Gaming 15 Laptop, featuring the 2^nd Gen AMD Ryzen™ 7 mobile processors.
HP and Lenovo announced commercial desktops powered by Ryzen 3000 PRO series processors.

Product Launches and Announcements

AMD launched 2nd Gen AMD EPYC processors for server and cloud customers.
AMD announced Radeon RX 5500 GPU series, which will be available for OEMs from November 2019.
AMD also launched the AMD Athlon™ PRO processors with Radeon Vega Graphics.

Financial Highlights

Revenue at US$1.80 billion, up 9% YoY, however, up 18% quarter-on-quarter (QoQ), in line with expectations.
Gross margin was 43%, up 3% YoY, primarily driven by increased Ryzen and EPYC processor sales.
Operating income was US$186 million compared to US$150 million a year ago. The increase was mainly due to higher revenue in the Computing and Graphics segment.
Cash, cash equivalents, and marketable securities were US$1.2 billion at the end of the quarter.
Client processor average selling price (ASP) increased YoY and sequentially driven by both Ryzen desktop and mobile processor sales.
GPU ASP increased YoY driven by higher channel sales and decreased QoQ due to a higher proportion of mobile sales.

AMD Q2 2019: Driving High on the 7nm Tech-Train

AMD is riding high under the leadership of its CEO Lisa Su, with increasing traction for its 7nm chipsets and product families. The latest addition in this family is its EPYC Rome processor for servers which will be officially launching on Aug 07, 2019. Although the revenues were down year-on-year (YoY), AMD’s expectation from datacenter CPU products is high in upcoming quarters based on the feedback from the initial usage. In addition, multiple partnership wins in different segments points that AMD is constantly improving in developing relations with customers-partners-suppliers alike, which was the bane for AMD’s demise a decade ago.

Counterpoint’s View on AMD’s Q2-2019 Earnings

AMD is banking on its datacenter CPU offerings with EPYC Rome processor.
AMD’s main competitor, Intel, has reported a decline in its datacenter segment. AMD is expected to grab more share and customers away from Intel in short to mid-term, leveraging superior technology and increasing customer wins.
Key gaming customer wins, Microsoft and Sony, powering their gaming consoles with Ryzen Gen-2 CPU will drive the revenues in the semi-custom segment.
Strategic partnership with Samsung will allow AMD to venture into mobile devices with its graphics solutions, as gaming smartphones get popular and flagship smartphones increase their graphics competency.
Overall, client CPU, datacenter CPU, and datacenter GPU will drive the revenues for AMD.

Key Highlights from AMD:

Overall Financials

Revenue at US$1.53 billion, down 13% YoY, however, up 20% quarter-on-quarter (QoQ), as per expectations.
Gross margin was 41%, up 4 pp YoY, primarily driven by increased Ryzen and EPYC processor sales.
Operating income was US$59 million compared to US$153 million a year ago. The decline was primarily due to lower revenue and higher operating expenses.
Cash, cash equivalents, and marketable securities were US$1.1 billion at the end of the quarter.

Segment Financials

Computing and Graphics segment revenue was US$940 million, down 13% YoY. Revenue was lower YoY primarily due to lower graphics channel sales, partially offset by increased client processor and datacenter GPU sales.
Client processor average selling price (ASP) was up YoY, driven by Ryzen processor sales.
GPU ASP increased YoY primarily driven by datacenter GPU sales.
Enterprise, Embedded, and Semi-Custom segment revenue was US$591 million, down 12% YoY. The YoY revenue decrease was primarily due to lower semi-custom product revenue, partially offset by higher EPYC processor sales.

Product Launches

AMD launched client and graphics processors based on advanced new architectures and 7nm process technology.
AMD Ryzen Desktop Processors are available now with up to 12 cores and 24 threads based on the new “Zen 2” core architecture.
AMD introduced RDNA, which provides up to 1.25X higher performance per-clock and up to 1.5X higher performance-per-watt compared to previous generation Graphics Core Next architecture.
AMD Radeon™ RX 5700 series graphics cards, built on the new AMD RDNA gaming architecture.

Partnerships and Customers Wins

AMD got two important customer wins in the gaming segment, powering Microsoft and Sony gaming consoles, with the Ryzen Zen-2 CPU core.
Samsung and AMD announced a multi-year strategic partnership through which Samsung will license AMD graphics IP to advance graphics technologies and solutions for mobile applications, including smartphones.
Apple announced an all-new Mac Pro featuring AMD Radeon Pro Vega II GPUs that deliver exceptional computational performance by harnessing the power of the 7nm process technology, “Vega” graphics architecture and AMD Infinity Fabric™ Link GPU interconnect technology.
Acer announced the upcoming availability of the new Acer Nitro 5 and Swift 3 laptops, based on 2^nd Gen AMD Ryzen Mobile processors, adding to the more than 40 new consumer and commercial notebooks based on the latest Ryzen Mobile and Ryzen Mobile PRO processors launched this year from all leading global OEMs.
AMD announced that it is working with Cray, the U.S. Department of Energy, and Oak Ridge National Laboratory to build what is expected to be the world’s fastest supercomputer – Frontier. Based on future-generation high-performance custom AMD EPYC CPUs and Radeon Instinct GPUs optimized for artificial intelligence (AI), Frontier is expected to deliver greater than 1.5 exaflops of processing performance.
2^nd Gen EPYC processors are advancing the use of AI across diverse research fields on Indiana University’s Cray Shasta™ supercomputer.

Future Guidance/Outlook

Q3 2019 – US$1.8 billion revenue, expected to be driven by Ryzen, EPYC, and Radeon product sales. Gross margin at 43%.
Full Year 2019 – AMD now expects revenue to increase a mid-single-digit percent over 2018 (5%-6%). Gross margin at 42%.

Top

Qualcomm looking to expand its ARM SoC Smartphone Success to the PC Platform

WATCH: “AI PC” Era Beckons with Snapdragon X Elite: Deep Dive with Qualcomm’s Kedar Kondap

Snapdragon X Elite Benchmarks Show Impressive Gains over Apple Silicon, Intel and AMD Processors

Below are some of the scores that Qualcomm shared:

ARM-based Laptops to Further Gain Share at x86’s Expense

AI PC to Drive Another Wave of Shipment Growth in 2024

Market summary for 2022

Related Posts

Related Posts

HPC silicon demand is on a multi-year, structural growth trajectory

5/4 nm is the mainstream today, 3 nm to ramp up from 2023

Related Posts

Related Posts

Background

Related Posts

AMD: Mobile processors with integrated security, budget GPU

CPU launches

GPU launches

Software announcements

Intel: 20+ new mobile processors, 50 processors in total including discrete GPU

CPU launches

GPU launches

Automotive launches

Software announcements

Intel vPro platform

NVIDIA: New GPU cards for laptops, adoption of Drive Hyperion platform

GPU launches

Software announcement

Analyst takeaways

Related Posts

Double-digit YoY sales growth again in 2021

Significant ramp-ups in 7/5nm by largely adopting more EUV layers

A high level of chip inventory becomes normal through the year

Capital intensity also maintains high level, given upside from wining IDM shares

Conclusion

Related Posts

Coverage

Insights

Social media

Contact us

Term of Use and Privacy Policy