On the morning of February 18, Beijing time, Intel held the 2022 Investor Conference, intel CEO Pat Kissinger and the relevant responsible person announced the product roadmap and development strategy of its major business units such as data center and artificial intelligence division (DCAI), client computing division (CCG), accelerated computing systems and graphics division (AXG), networking and edge division (NEX), Intel foundry services (IFS), and Mobileye. Since announcing its development route for the next five years last year and showing its determination to "start from scratch", Intel has made big moves: first investing $20 billion in the construction of two new fabs in Ohio, USA, and then acquiring the world's ninth largest wafer foundry, Gaota Semiconductor, for $5.4 billion. As a result, today's Intel 2022 Investor Conference has become particularly interesting.
During the conference, Pat Kissinger highlighted the foundry business, accelerated computing systems and graphics division, and Mobileye, and listed these as Intel's emerging businesses, and it is expected that by 2026, nearly half of its revenue growth will come from these emerging businesses, achieving annual revenue growth of 10% to 12%.
Xeon processor product roadmap reveals mysteries
In terms of data center and artificial intelligence, Intel Data Center and Artificial Intelligence has announced the next generation of Intel Xeon product roadmap to be released between 2022 and 2024. In the first quarter of 2022, Intel will deliver the Sapphire Rapids processor manufactured on the Intel 7 process and bring to market Intel's most feature-rich Xeon processor to date. Sapphire Rapids is known to dramatically improve performance for a wide range of workloads, achieving up to 30x performance gains in AI alone.
Emerald Rapids, a next-generation Xeon processor powered by Intel's 7-process process, will also be available in 2023, further enhancing the memory and security advantages of existing platforms. By 2024, Intel will introduce a new energy-efficient core Xeon processor, Sierra Forest, which is based on Intel 3 process and offers high-density and ultra-energy-efficient performance. In the same year, the Granite Rapids processor will also be available, and the process process will be upgraded from Intel 4 to Intel 3 to further strengthen Intel's leadership in the industry.
Sandra Rivera, Executive Vice President and General Manager of Intel's Data Center and Artificial Intelligence Group, said, "Today, we clearly demonstrated the industry-leading Intel Xeon product roadmap that will continue to drive our growth in 2024 and beyond, and comprehensively strengthen Intel's leadership position in the industry. Based on the diverse needs and time planning of our customers, we have worked with our customers to create a rich product portfolio, which will also drive further innovation through our customers' developer ecosystems. ”
In addition, Intel has unveiled its new architectural strategy. In future generations of Xeon processors, there will be a dual-track product roadmap based on both the P-core and the energy-efficient core, and the two optimized platforms will be integrated into a common, industry-defining platform. This new architectural strategy will maximize the performance per watt and segmentation capabilities of the product, thereby comprehensively enhancing Intel's competitiveness in the industry.
Client computing business has become the main driving force for growth
Intel said that as PC products become more important, the client computing division will continue to be the main driver of Intel's growth. In 2021, Intel's global PC shipments exceeded 340 million, up 27% from 2019. Intel predicts that the market demand for PCs will continue to grow, mainly from the growing demand of the existing customer base, as well as the increase in PC ownership and penetration rates worldwide.
At the Investor Conference, Intel Client Computing Division (CCG) outlined the client offerings to be launched in the coming years and updated the client computing division's roadmap. Among them, Raptor Lake, which uses Intel 7 process technology and high-performance hybrid architecture, is expected to be officially launched in the second half of 2022, which will bring two-digit performance improvements and stronger overclocking functions compared with the previous generation of Alder Lake. The Raptor Lake configuration will be upgraded to 24 cores and 32 threads, and will be compatible with the sockets of the AlderLake system.
Meteor Lake, which uses intel's 4 process, will be available in 2023, and Arrow Lake, the first product built using Intel's 20A process, will be officially available in 2024. It is understood that these two products combine artificial intelligence and graphics graphics technology that brings independent graphics performance, or will help Intel achieve a huge breakthrough in XPU.
Due to the surge in revenue from the Networking and Edge business, Intel also established the Networking and Edge Division (NEX) in 2021, and its products include infrastructure processing units (IPUs). Pat Kissinger believes that Intel will be able to provide customers with a way to program networks in this area, lead the transformation of mobile networks, and support more new technologies and workloads in the EDGE of ARTIFICIAL intelligence.
The Accelerated Computing Systems and Graphics business is expected to contribute nearly $10 billion in revenue
In terms of Accelerated Computing Systems and Graphics, Intel said that three sub-divisions of the Accelerated Computing Systems and Graphics Division (AXG) are shipping products as planned and are expected to generate more than $1 billion in revenue for the company in fiscal 2022. As Intel's primary growth engine, Pat Kissinger expects the Accelerated Computing Systems and Graphics division to generate more than $10 billion in revenue in 2026.
First of all, in terms of visual computing product roadmap and strategic planning, Intel updated the Time Node and Roadmap of Iris Graphics, and said that the Accelerated Computing Systems and Graphics Division is expected to ship more than 4 million independent GPUs in 2022, and OEMs will release laptops equipped with Intel Iris Graphics (codenamed Alchemy) in the first quarter of 2022. Meanwhile, Intel will ship discrete graphics for desktops in the second quarter of 2022 and discrete graphics for workstations in the third quarter. In addition, Celestial, for the super enthusiast market, has officially begun its architecture research and development work. At the same time, users can also directly read Intel Iris graphics information through the endgame project services, and get a long-time access and low latency computing experience. It is understood that the Endgame project will be officially launched this year.
Secondly, Intel also updated the supercomputing product roadmap and strategic plan, and said that more than 85% of the world's supercomputers are currently using Intel Xeon processors, on this basis, the Acceleration Computing Systems and Graphics business unit will further achieve higher computing power and memory bandwidth, and deliver industry-leading CPU and GPU product roadmaps to empower high-performance computing (HPC) and AI workloads. Intel will get more than 35 HPC-AI designs from leading OEMs and cloud service providers (CSPs). In addition, the Accelerated Computing Systems & Graphics Division has developed a technical route to Z-level computing (1,021 floating-point operations per second) by 2027.
Among them, Intel introduced a variety of its products to help it achieve its goal of achieving Z-level computing by 2027. First, the upcoming Sapphire Rapids has built-in high-bandwidth memory, which can provide up to 4x the memory bandwidth for applications, achieving a performance improvement of up to 2.8 times compared to the 3rd generation Intel Xeon Scalable processors. In addition, in the same computational fluid dynamics application scenario, Sapphire Rapids with built-in high-bandwidth memory performs 2.8 times better than comparable solutions.
Second, the Ponte Vecchio GPU, which will be delivered this year, meets industry-leading performance standards and demonstrates 2.6 times better performance than market-leading solutions. Third, Intel's first hardware-equipped AV1 encoder GPU, The Arcic Sound-M, achieved a 30% increase in bandwidth. In addition, it is equipped with an industry-leading open source media solution that enables high-quality transcoding, stream density, and cloud gaming for supercomputers in the media and analytics space. Arctic Sound-M is now customer-facing and is expected to ship in mid-2022. Fourth, Falcon Shores, which is scheduled to be available in 2024, is a new architecture that integrates x86 and Xe graphics cards in the same socket, and will achieve more than 5x performance improvements per watt of performance, compute density, memory capacity and bandwidth.
Foundry services will focus on automotive chips
The automotive industry is undergoing a revolutionary shift and showing considerable growth, with the automotive semiconductor industry doubling revenue and expected to reach $115 billion by 2030. However, the current incomplete supply chain and traditional process technology may not be able to meet the growing demand for automotive semiconductors or support the transition to more compute-intensive applications. Therefore, at this investor conference, Intel said that Intel Foundry Services (IFS) is assembling a dedicated automotive team with the intention of providing complete solutions for automakers.
First, Intel Foundry Services (IFS) will develop a high-performance, open automotive computing platform that helps automotive OEMs build next-generation experiences and solutions. Based on chiplet-based building blocks and Intel's advanced packaging technology, the platform provides optimized solutions for multiple technology nodes, algorithms, software, and applications to meet the computing needs of the next generation of autonomous vehicles.
Second, Intel Foundry Services (IFS) will optimize advanced processes and technologies for the specific needs of microcontrollers and automobiles, and combine them with advanced packaging technologies to help customers design different kinds of automotive semiconductor products. In this business, Intel will work with Mobileye. Intel CEO Pat Kissinger stressed that Mobileye has developed a comprehensive development strategy that will launch a hybrid system with full-featured vision and lidar, for which Mobileye has collected 2.5 billion kilometers of road data. At the same time, Intel Foundry Services (IFS) will also provide design services and access to Intel IP for automakers, enabling them to leverage Intel's technical expertise to build products.
In addition, Pat Kissinger once again responded to why Intel wanted to enter the foundry business. In his opinion, IDM makes IFS better, and IFS makes IDM better. Its IDM model provides billions of dollars in R&D funding in the foundry field, while IFS can help Intel build better PDK and EDA tools, richer IP, and enable semiconductor industry and foundry customers to benchmark Intel every day.