Finance Association News, December 1st (Reporter Fu Jing) "The rapid deployment of AI applications is continuously increasing the requirements for semiconductor performance, power consumption, integration, and reliability, which has driven the demand for system-level semiconductor manufacturing. Therefore, Intel is actively investing, and Intel 18A will be in mass production in 2025, with the next-generation AI PC processor Panther Lake and the next-generation datacenter processor Clearwater Forest based on Intel 18A also set to be released next year." Recently, Wang Rui, Senior Vice President of Intel and Chairman of Intel China, stated at the Intel New Quality Productivity Technology Ecology Conference.

Chips play a core role in the era of large models. According to reports from the Finance Association, on the application side, chip manufacturers have introduced AI PC processors that support local AI large model operations, with terminal manufacturers such as HP, Lenovo, and Dell responding actively. The AI PC ecosystem is increasingly vibrant, with global AI PC shipments accounting for 20% of total PC shipments in Q3 of this year, and many new AI PC products are set to be launched this year. In the datacenter domain, while GPUs dominate, CPUs are also targeting the growing large model inference market, with manufacturers focusing on increasing server chip production capacity. At the edge, by configuring CPUs, GPUs, and NPUs, edge AI processing capabilities have also driven the practical deployment of AI applications in sectors such as retail, pharmaceuticals, logistics, agriculture, transportation, and manufacturing.

"To address the challenges posed by the surge in AI workloads on computing power, systems, deployment costs, and more, Intel has introduced a full-stack AI product portfolio including Core Ultra processors, Xeon Scalable processors, Gaudi deep learning accelerators, Intel GPUs, and more." Speaking about the products, Wang Rui said.

Application side: Cumulative shipments of AI PCs to exceed 0.1 billion units by the end of next year.

The latest from Finance Association reporters reveals that Intel has set a target to exceed 0.1 billion cumulative shipments of AI PCs by the end of next year. As of the end of October, AI PCs based on Core Ultra processors have shipped 20 million units, with an expected increase to 40 million units by the end of 2024.

Since Intel first introduced the concept of AI PCs in the industry in September last year, AI PCs have continued to attract market attention and have stimulated a wave of PC replacements. Currently, looking at specific global shipments, research institution Canalys' latest data shows that in the third quarter of 2024, global AI PC shipments reached 13.3 million units, accounting for 20% of the total PC shipments for that quarter.

From an industry chain perspective, the most investment-worthy areas of AI PCs are in the upstream chip manufacturing and midstream whole machine manufacturing sectors, especially in CPU, GPU, NPU, storage, and other chip manufacturing. Currently, manufacturers such as Intel, AMD, Qualcomm, among others, have successively released AI PC processors that support local AI large model operations.

In September this year, Intel released the Core Ultra 200V series processor, codenamed Lunar Lake, increasing computing power to 120 TOPS. It is understood that multiple new AI PCs will be launched this year, such as: HP will launch the NextGen AI PC featuring the Lunar Lake platform at the end of the year, and Lenovo's previously launched AI PC Element series, Aura series, using Intel chips. The ThinkPad X1 Carbon Aura AI Element Edition is planned to be launched in early December.

Reported by Caixin Media.

According to reports by Caixin Media, Intel introduced AI features into the desktop field for the first time by releasing the Core Ultra 200S processor this year. At the same time, they will also explore opportunities in the all-in-one PC and mini PC market. Based on AI PCs, the efficiency, processing power, and level of intelligence of traditional applications will be improved. In the gaming sector, it will enable new gameplay, stimulating the replacement of gaming hardware. In order to create more AI PC application scenarios, Intel has collaborated with more than 40 local ISVs (software solution providers) to expand the software ecosystem.

During the conference, a Microsoft executive told Caixin Media reporters that if the hardware specifications reach at least 40 TOPS of NPU peak performance, a minimum of 16GB of memory, and at least 256GB of storage space, Windows 11 laptops can receive Windows 11 AI + PC certification. In December this year, features such as one-click HD enhancement for low-resolution images, one-click removal of excess elements from images, and Bing screenshot search will be pushed to Windows Insider for user access.

On the server side: Adding capacity and exploring 'oil-cooling' innovation.

Currently, Intel is also ramping up server chips, with Xeon 6 performance cores and efficiency core processors aimed at data centers already launched.

On October 28, Intel announced plans to expand its Chengdu packaging and testing base. Caixin Media reporters recently learned that the focus of this expansion is not only on establishing a one-stop customer solution center, but also on adding server chip capacity. Wang Rui emphasized that the Chengdu base will cover various products from client to server chips, extensively meeting the demands of the Chinese market and significantly reducing response times for customers.

It is reported by Caixin Media that currently more than 70% of GPU-accelerated servers have chosen Intel Xeon CPUs as their 'leaders'. Based on Xeon 6 performance cores, new servers from terminal manufacturers like Inspur Electronic Information Industry, Superfusion, Unisplendour Corporation, New H3C under Tsinghua Unigroup, ZTE, Lenovo, and other manufacturers will be successively launched from September this year to the first quarter of next year.

Datacenter computing power demand and energy consumption challenges continue to rise, making liquid cooling a key focus in the industry. Previously, Intel has cooperated with numerous OEM manufacturers, server manufacturers, cloud providers, software developers, and operating system providers to accelerate green transformation of datacenters. Financial Union reporters have recently learned that in the field of immersion liquid cooling, the 'oil cooling' innovation has also begun: Inspur Electronic Information Industry and Intel have released an oil-based immersion solution, using environmentally friendly cooling liquid, based on Intel's G-Flow immersion liquid cooling datacenter solution, providing heat dissipation for kilowatt-level chips. In terms of G-Flow, Intel has collaborated with GreenCloud Technologies and Li-Tech in the design phase to optimize flow channels and radiator designs in systems and cabinets, utilizing liquid high potential difference to generate gravitational effects for high-performance heat dissipation.

The immersion liquid cooling solution based on G-Flow uses liquid cooling oil. Financial Union reporters photo.

"In our solution, we use a synthetic coolant specifically designed for data centers, with a flash point above 200 degrees. In fact, the CPU, which has the highest temperature in electronic components, generally does not exceed 115 degrees, so it is relatively safe. In the past, liquid cooling used fluorinated liquids, but with increasing environmental awareness, more and more manufacturers may explore the use of 'oil cooling'. Inspur Electronic Information Industry has conducted extensive research in 'oil cooling'. Currently, this solution is still in the early stages of promotion, with hopes to scale it for mass production in the future," an Intel spokesperson told Financial Union reporters.