Tech giants unveil revolutionary AI products

TAIPEI — Artificial intelligence (AI) is no longer within the realm of science fiction. Humanity has reached a point of advanced computing with AI as a revolutionary tool after many years of building and improving technology. This is the focal point of Taiwan's biggest annual technology trade show, Computex 2024, with the theme "Connecting AI."

Rick Tsai, MediaTek Inc. vice chairman and chief executive, said, "AI will soon be everywhere, from devices as small as smartphones, smart electronic household appliances and offices."

He said his company wants to be "a major player in enabling ubiquitous AI for everyone and is also going for hybrid computing from edge to the cloud."

Tsai added that ubiquitous AI has four pillars: High-performance computing, AI accelerators in the cloud, moving from cloud to edge, and a strong ecosystem.

High-performance computing involves a huge computing power, specifically graphics processing units (GPUs) and neural processing units (NPUs), since AI will be present in almost all devices. AI accelerators in the cloud are also important to make it plausible on the edge.

Moving from cloud to edge will enable users to fully integrate AI without compromising privacy. Lastly, a strong ecosystem involves collaborating with other companies to bring this to customers.

During Computex 2024, key industry players revealed their plans to make AI computing more accessible, efficient and reliable because it will be everywhere soon.

AMD: AI in modern computing

Lisa Su, Advanced Micro Devices Inc. (AMD) chief executive, emphasized the importance of AI in modern computing. To keep up with this, AMD will be releasing a series of computer processing unit (CPU) cores that are optimized for AI.

"AMD is focused on delivering a broad portfolio of high-performance and energy-efficient computer engines for AI training and inference, including CPUs, GPUs and NPUs, enabling an open, proven and a developer-friendly ecosystem that ensures all leading AI frameworks, libraries and models are fully enabled on AMD hardware," said Su. "This goes with co-innovating with its partners as well. "

One of its new products under the Zen 5 core architecture is the Ryzen 9 9950X, the highest-performance and most energy-efficient core that is perfect for gaming personal computers (PCs), supercomputers and data centers, according to AMD. The 12-core, 8-core and 6-core versions will be released in July.

While Strix is its next-generation processor for ultrathin and premium notebooks that combines the Zen 5 CPU, faster RDNA 3.5 graphics and XDNA2 NPU. These features allow faster graphics and better performance. It is the "best technology in one chip," Su said.

The XDNA2 NPU is specially crafted to balance accuracy and performance — something that developers frequently compromise for the other. It supports block FP16, which helps AI generate better images.

The XDNA2 NPU core enables leadership performance and capabilities for AI PCs.

"AI is our number one priority, and we're at the beginning of an incredibly exciting time for the industry as AI transforms virtually every business, improves our quality of life, and reshapes every part of the computing market," Su said.

AMD has also taken AI computing to the edge, where its technology has been assisting social sectors such as Illumina health care, Subaru in automotive, and Canon and Hitachi industrial. AMD's Versal AI Edge Gen 3 series is perfectly suited for these functions, as it integrates preprocessing, inference and post-processing in a single chip.

When it comes to data centers, AMD also has a full portfolio of high-performance CPUs and GPUs to address server workloads. AMD's EPYC is the processor of choice for cloud computing and powering internal workloads for large hyperscalers. This not only enables new AI capabilities, but it can also reduce rack space by 80 percent and consume 65 percent less energy.

AMD's upcoming fifth-generation EPYC processor is named Turin, which features 192 cores, 384 threads and 13 different chiplets. With Turin, AI inference performance is possible with less hardware.

In terms of data center GPUs, the MI300X provides out-of-the-box support for most common AI models and delivers significantly better inferencing performance.

Intel: AI in every device, company

Patrick Gelsinger, Intel chief executive, explained how AI will soon be in every device and company, which is why the world needs more flexible and resilient supply chains. This has led to the company creating the Intel Foundry.

Patrick Gelsinger, Intel chief executive, shows some of the products that Intel plans to release to enable artificial intelligence computing for everyone.

When it comes to AI powering data centers, Intel has been the key in enabling cloud data centers with 130 million Xeons. However, there have been more demands for open ecosystems, more computing and performance, energy efficiency, and greater server capacity.

To address this, Intel will be launching the Xeon 6 with E-cores.

"We see this as an essential upgrade for the modern data center. A high core count, high density [and] exceptional performance per watt," Gelsinger said.

This new Xeon will allow companies to condense and attack workloads with performance and efficiency, ultimately reducing the size of servers.

Later this year, Intel also plans to bring out the second generation of the Xeon 6 with 288 cores. In the third quarter of the year, it plans to release the Xeon's P-core version.

On the other hand, Intel Gaudi 2 is a cost-effective, high-performance, and generative AI training and inference solution. Gelsinger said Gaudi delivers incredible performance, which will still be improved with Gaudi 3 with a faster time, faster inference and better performance per dollar.

"Together with our partners, we're thrilled with the momentum and the mass market opportunities that Gaudi 3 is bringing forward to our customers because they want choice, [total cost of ownership], alternatives and performance. Gaudi 3 and our partner ecosystem are delivering exactly that," Gelsinger said.

Intel has also announced the launching of the Intel Cyber Edge software platform, making it easy for enterprises to deploy, secure and manage AI applications at the edge. Intel has 990,000 edge deployments and 200 million CPUs at the edge. Therefore, this technology will be exceptionally useful.

In terms of AI PCs, Intel is at the forefront of creating this with Lunar Lake. Gelsinger said, "Lunar Lake will power the largest number of next-generation AI PCs with over 80 designs and 20 original equipment manufacturers that will start shipping in volume." Other variations include the Meteor Lake, Lunar Lake for mobile, Aero Lake and Panther Lake.

Supermicro: Green computing at the forefront

Charles Liang, Super Micro Computer Inc. (Supermicro) co-founder, chairman, president and chief executive, said generative AI brings tremendous benefits to the world. Supermicro has been investing in this, as it is the first company to bring optimized GPU platforms into the market along with Nvidia.

Charles Liang (right), Supermicro co-founder, chairman, president and chief executive, with Jensen Huang, Nvidia co-founder, president and chief executive, advocate for green computing in a world of artificial intelligence.

Supermicro also aims to bring green computing to the forefront of AI solutions, with a big bonus, too.

"Together with our partners, we are on a mission to build a more sustainable data center infrastructure that addresses the rising need for more compute and storage in today's world, especially when it comes to generative AI training," Liang said.

In line with this, Supermicro has launched its newly optimized X14 and H14 portfolio, powered by Intel Xeon and Turin. Its new Direct Liquid Cooling (DLC) can use room-temperature water to provide optimal cooling for servers at a lower cost and with a smaller environmental impact.

"These latest high-performance chips provide better performance for one but at the same cost," Liang said.

Additionally, DLC computing could help save companies for as much as $60 million.

"$60 million may not be a really big deal and motivation for some big corporation, but through liquid cooling, we can reduce the huge amount of [carbon] emissions and make our planet greener," Liang said.

Microsoft: The power of Copilot Plus

Pavan Davuluri, corporate vice president of Windows Devices at Microsoft, showed a glimpse of the Copilot Plus PCs, wherein one device is powerful enough to keep up with users and efficient enough to be always running.

Some of its features are Recall, which helps instantly find anything in the PC through semantic index content in the background; live captions that translate audio in real-time, and Co-creator, which generates high-quality images by drawing and Paint.

"To realize the full power of AI in a PC, we [reengineered] the entire system from the chip through every layer of Windows. These are the fastest, most performant [and] most intelligent PCs," Davuluri said.

HP: Improving productivity, flexibility

Enrique Lores, Hewlett-Packard Co. (HP) president and chief executive, said there is "a significant tension between companies to improve productivity and flexibility, and AI can help close this gap."

With this, HP aims to deliver more AI PCs into the market that are carefully integrated with the new chips and processors to increase productivity for everyone.

Lores showed the next generation OmniBook, which is integrated with the latest Ryzen AI 300 series and will have 50 tera operations per second (TOPS). Generative AI models such as Stable Diffusion XL Turbo will also work well on the new OmniBook.

"Just imagine the productivity this is going to provide to product managers and creatives that are going to start creating their designs. [Just] with this solution, [it] will really be able to accelerate that work," AMD's Su said.

Lenovo: Laptops with AI processors

Luca Rossi, Lenovo Intelligent Devices Group president, said "the brand has been catering to AI computing, such as the ThinkStation P620 and the Lenovo AI Engine Plus software in its Legion gaming laptops."

These devices are used for a myriad of applications, such as video editing, 3-dimensional (3D) rendering and gaming. Rossi has revealed that later this year, Lenovo will release AI laptops with AI processors in its various laptop models, such as Yoga, ThinkPad and ThinkBook.

Lenovo has also engineered Creator Zone, a software that is tailor-made for creators to boost creativity and productivity. Lenovo AI can also interpret user data to ensure a more personalized output.

Aside from that, Lenovo will also have Smart Connect, which unifies AI PCs, tablets, smartphones and other related devices into the Lenovo ecosystem.

Asus: Pushing the boundaries of AI

Jonney Shih, Asus chairman, said AI will be "one of the most disruptive innovations of mankind, and AI PCs will play a critical role in the new distributed hybrid AI ecosystem."

"The ubiquitous AI era is the mega paradigm shift that we have long envisioned at Asus, and I'm so overjoyed that it's finally becoming a reality. The world will be full of AI brains that come in different forms and sizes," Shih said.

Asus has bared its range of cutting-edge AI PCs across its different brands, such as ZenBook, ProArt, Vivobook, Asus Tuf and Republic of Gamers laptops. These new lineups are equipped with powerful NPUs with 50 TOPS.

Asus has also shown StoryCube, an AI-powered digital asset management application designed to provide users with a seamless and efficient file-organizing experience. It is capable of recognizing faces and detecting and sorting various files and media.

"We are pushing the boundaries of AI PC innovation and delivering truly groundbreaking AI [experiences] to users," Shih said.

Stability.ai: Next frontier of genAI

Christian Laforte, Stability.ai chief technology officer and co-chief executive, works with AMD's MI300X to deliver the Stable Diffusion Open Access AI models, which allow many developers, enthusiasts and researchers to create and edit images using applications, such as Stable Assistant and Stable Artisan.

"We're involved in language, coding, music, speech and 3D. [Combining] all of those together, we aim to soon boost at least 10 times the productivity of filmmaking and video game creations," Laforte said.

Laforte also announced that the Stable Diffusion 3 medium model will be released on June 12 this year for everyone to download. This version boasts unprecedented visual quality and will help developers and researchers discover the next frontier of generative AI.

"A lot of work went into this, and we're really excited to see what the community will end up doing with this," Laforte said.