Reporting from Taipei, Taiwan: At COMPUTEX 2025 in Taipei, Intel made a splash with the unveiling of new graphics processing units (GPUs) and AI accelerators aimed squarely at professionals and developers. These announcements coincide with a significant milestone: Intel’s 40th anniversary in Taiwan, a celebration of four decades of collaboration and innovation within Taiwan’s dynamic tech ecosystem around the x86 architecture. Intel emphasized the ongoing importance of x86 in accelerating innovation as high-performance AI applications demand more computational power.

Lip-Bu Tan, Intel CEO, underscored the significance of this anniversary, stating, “For the past 40 years, the power of our partnership with the Taiwan ecosystem has fueled innovation that has changed our world for the better. This week, we are renewing our commitment to our partners as we work to build a new Intel for the future. Together, we will create great products that delight our customers and capitalize on the exciting opportunities ahead.”

New Intel Arc Pro B-Series GPUs for AI and Workstation Performance
Expanding its professional graphics lineup, Intel introduced the Intel Arc Pro B60 and Intel Arc Pro B50 GPUs. Built on the advanced Xe2 architecture, these cards boast Intel Xe Matrix Extensions (XMX) AI cores and cutting-edge ray tracing units, delivering robust performance for creators, developers, and engineers.

“The Intel Arc Pro B-Series showcases Intel’s commitment in GPU technology and ecosystem partnerships,” said Vivian Lien, vice president and general manager of Client Graphics at Intel. “With Xe2 architecture’s advanced capabilities and a growing software ecosystem, the new Arc Pro GPUs deliver accessibility and scalability to small and medium-sized businesses that have been looking for targeted solutions.”

Specifically designed for demanding AI inference tasks and professional workstation applications, the Arc Pro B-Series GPUs offer AI-ready capabilities, with the B60 featuring 24 gigabytes (GB) of memory and the B50 offering 16GB. Their multi-GPU scalability provides a versatile solution for creators, AI developers, and various professionals.

These GPUs are optimized for AEC (architecture, engineering, construction) and inference workstations, ensuring stability and performance through extensive Independent Software Vendor (ISV) certifications and optimized software. Notably, the Intel Arc Pro B-Series GPUs are compatible with both consumer and professional drivers on Windows. For Linux users, they support a containerized software stack, simplifying AI deployments and promising progressive feature and optimization upgrades. By combining substantial memory capacity with crucial software compatibility, the Arc Pro B-Series presents creators and AI developers with scalable and cost-effective solutions.

Intel also previewed Project Battlematrix, a configurable workstation-class platform powered by Intel Xeon processors. This platform aims to streamline AI development by supporting up to eight Intel Arc Pro B60 24GB GPUs, enabling the training of medium-sized (up to 150 billion parameters) and accurate AI models with a total of up to 192GB of video random-access memory.

Sampling of the Intel Arc Pro B60 GPU from add-in board partners like ASRock, Gunnir, Lanner, Maxsun, Onix, Senao, and Sparkle will begin in June 2025. The Intel Arc Pro B50 GPU will be available through Intel-authorized resellers starting in July 2025.

Intel Gaudi 3 AI Accelerators Available for PCIe and Rack Scale Deployments
Further solidifying its AI strategy, Intel announced new deployment options for its Intel Gaudi 3 AI accelerators. Intel Gaudi 3 PCIe cards will support scalable AI inferencing within existing data center server environments. These cards offer the flexibility to run various AI models, from Llama 3.1 8B to full-scale Llama 4 Scout or Maverick, catering to businesses of all sizes. The Intel Gaudi 3 PCIe cards are slated for availability in the second half of 2025.

Intel Gaudi 3 rack scale system reference designs are engineered for both flexibility and scale, accommodating up to 64 accelerators per rack and providing a massive 8.2 terabytes (TB) of high-bandwidth memory. Their open, modular design aims to prevent vendor lock-in. Equipped with liquid cooling, these systems are designed for robust performance and efficient total cost of ownership (TCO).

The Intel Gaudi rack scale architecture is optimized for running large AI models and excels in real-time inferencing with low-latency performance. These configurations underscore Intel’s commitment to open, flexible, and secure AI infrastructure, supporting both custom and Open Compute Project (OCP) designs for cloud service providers (CSPs).

Intel AI Assistant Builder Now Publicly Available
Following its initial reveal at CES 2025, Intel AI Assistant Builder is now publicly available in beta on GitHub. This lightweight, open software framework enables developers to build and run custom AI agents locally on Intel-based AI PCs. With Acer and ASUS showcasing innovative solutions at COMPUTEX, Intel AI Assistant Builder empowers developers and partners to rapidly create and deploy tailored AI agents for their organizations and end customers.

Reporting from Taipei, Taiwan: Nvidia CEO Jensen Huang delivered a keynote address at COMPUTEX today, unveiling a range of new artificial intelligence technologies for both enterprise and consumer applications, and announcing a new Taiwan office. Huang introduced the new Taiwan office, named Nvidia Constellation, which will be located in Taipei.

During his nearly two-hour speech at the COMPUTEX in Taipei, Huang presented new AI data center technology, AI cloud and computing products for consumers, and AI software for robotic applications. A key announcement was a partnership with Foxconn (Hon Hai Precision Industry Co Ltd) to build an AI factory in Taiwan using Nvidia’s Blackwell GPU series.

Huang also highlighted NVLink Fusion, a platform designed to facilitate the development of AI infrastructure. Analysts noted that this platform, which allows Nvidia’s servers to work with chips from other companies, represents a significant advancement for AI technology. Other announcements included small-form personal AI supercomputer products, new AI server technology, new AI agent software, and AI software for robots.

These new technologies will be developed in collaboration with several Nvidia partners, including Acer, MSI, Gigabyte, Marvell, MediaTek, Qualcomm, and Fujitsu. Huang’s presentation emphasised Nvidia’s commitment to maintaining its leadership in AI development. The company’s value has surged in the past two years due to the rapid growth of the AI industry.

Despite some recent concerns about AI chip demand and new export restrictions on sales to China, markets generally expect Nvidia to maintain its strong position. Continued investments in AI infrastructure by major U.S. companies like Microsoft, Meta, Amazon, and Alphabet indicate sustained near-term demand for Nvidia’s products. Huang repeatedly emphasized Taiwan’s importance to Nvidia’s growth, citing the island’s central role in the global chipmaking and electronics industry. TSMC and Foxconn are Nvidia’s largest suppliers.

Nvidia also announced plans to sell technology that will connect chips to enhance chip-to-chip communication for AI development. The new version of its NVLink technology, called NVLink Fusion, will be available to other chip designers to help create powerful custom AI systems with multiple interconnected chips. Marvell Technology and MediaTek plan to use NVLink Fusion in their custom chip development efforts.

Nvidia developed NVLink years ago to enable high-speed data exchange between chips, as seen in its GB200, which combines two Blackwell GPUs with a Grace processor. Huang noted the shift in Nvidia’s focus, stating that his presentations once dedicated 90% of their time to graphics chips, but now the company has expanded from its origins as a video game graphics chip maker to become the leading producer of chips for the AI boom that began with the launch of ChatGPT in 2022.

Nvidia has also been designing CPUs for Microsoft’s Windows operating system, using technology from Arm Holdings. Last year at COMPUTEX, Huang’s presence generated significant excitement in Taiwan. During Nvidia’s annual developer conference in March, Huang detailed the company’s strategy to address the shift from building large AI models to running AI-based applications.

He introduced several new generations of AI chips, including the Blackwell Ultra (coming later this year) and the Rubin processors, followed by Feynman processors (set to arrive in 2028). Nvidia also launched a desktop version of its AI chips, called DGX Spark, for AI researchers. Huang announced that DGX Spark is now in full production and will be available in a “few weeks”. COMPUTEX, which is expected to host 1,400 exhibitors, is the first major gathering of computer and chip executives in Asia since U.S. President Donald Trump threatened to impose sweeping tariffs to push companies to increase production in the U.S.

Ahead of Computex, Acer has made a series of announcements detailing its new and upcoming AI-powered devices, including laptops, desktops, wearables, and e-mobility solutions. For gamers, Acer unveiled the Predator Triton 14 AI gaming laptop, featuring an Intel Core Ultra 9 processor and NVIDIA GeForce RTX 5070 GPU. Acer highlights its use of graphene thermal interface materials (TIM) in the CPU, which it says is an industry first for laptops, aimed at improving cooling. The Predator Helios Neo 14 AI, with similar specifications, was also announced, alongside the Predator Orion 3000 gaming desktop, equipped with an Intel Core Ultra 7 processor and the same NVIDIA GeForce RTX 5070 GPU found in the laptops.

The Predator Triton 14 AI is set to launch in EMEA in July, with a starting price of EUR 2,999. The Predator Helios Neo 14 AI will start at EUR 1,699, and the Predator Orion 3000 desktop will be available in EMEA for EUR 1,099. Acer’s more budget-friendly Swift lineup also received updates with the introduction of the Swift X 14 AI and Swift X 14 AI Creator laptops. Both models are powered by the NVIDIA GeForce RTX 5070 GPU, emphasizing their AI capabilities and graphics performance. The Swift 14 AI is a Copilot PC, utilizing AMD Ryzen AI 300 series processors, while the Swift X 14 AI Creator features Intel Core Ultra Processors (Series 2). Both laptops boast 3K OLED displays and haptic touchpads. These Swift AI laptops are slated for a July release in EMEA, with a starting price of EUR 1,799.

In the thin and light category, Acer showcased the Swift Edge and Swift Go AI Copilot+ laptops. The Swift Go AI, constructed with a magnesium-aluminum chassis, weighs just 0.99 kilograms and claims up to 21 hours of battery life. It will be offered in 14-inch and 16-inch versions, with the latter promising 16 hours of battery life. All three Swift laptops are powered by Intel Core Ultra 200V series processors and are “eligible to receive updates to Copilot+ PO experiences.”

The Swift Edge is expected to arrive in EMEA in June, starting at EUR 1,599. The Swift Go 16 AI will have a starting price of EUR 1,200, while the 14-inch version will start at EUR 1,199. The Aspire series also sees new AI Copilot+ laptops with the Aspire 14 AI and Aspire 16 AI, offering configurations with either Intel Core Ultra or AMD Ryzen processors and Intel Arc graphics.

Acer also introduced new Predator and Nitro gaming and smart monitors. Highlights include the Predator X27U F5, featuring a 500 Hz refresh rate, and the Predator X27 X, with 4K resolution. Both support AMD FreeSync Premium Pro technology. Additionally, the company announced the Acer Nitro GA321QK P and Nitro GA341CUR W0 Google TV smart monitors, and the Acer Nitro PG271K, a 4K portable monitor with a kickstand.

Expanding into the wearables market, Acer unveiled the Acer FreeSense ring, designed for 24/7 biometric tracking and “AI-driven wellness insights.” The company also announced the Acer AI Transbuds, capable of real-time, two-way voice translation for 15 major languages, even when used by only one person.

Acer is also entering the e-mobility sector, showcasing new e-scooters and e-bikes for urban commuting and outdoor use. The lineup includes the Acer eScooter Series 4 Select, Series 5 Select, Predator ES Storm, and Predator ES Thunder, all equipped with 400W to 500W motors, adaptive braking, and suspension systems. The Acer eMobility App will allow riders to track performance and manage security settings. The Predator eRanger, a fat-tire e-bike previously shown at the 2025 Taipei Cycle Show, was also on display.

Acer has not yet released pricing details for its wearables or e-mobility devices.