Last year, Dell killed off the XPS lineup. However, the brand quickly learned the error of its ways and resurrected the lineup this year. When the XPS 14 and 16 launched last January, the XPS 13, touted as the lineup’s thinnest laptop to date, made some interesting teases. Now, at COMPUTEX 2026, the new 13.4-inch notebook finally made an appearance.

Though the number in the name says “13,” the XPS 13 is defined by more than just its screen size. It’s the thinnest notebook in the entire XPS lineup, measuring at only 12.7 millimeters and weighing only 2.2 pounds. And yet it’s still remarkably durable with a premium CNC aluminum construction.

Inside, the XPS 13 can carry up to a Series 3 Intel Core Ultra 7 chipset. This is paired with up to 32GB of memory and up to 1TB of storage.

Meanwhile, the 13.4-inch display touts 2.5K resolution and an InfinityEdge touchscreen display. It has a 500nit typical brightness, up to 120Hz VRR, and Dolby Vision. You’re also getting a quad speaker setup with two main speakers and two tweeters, all of which are compatible with Dolby Atmos.

Despite its size, the notebook is fitted with 2 USB-C ports with DisplayPort 2.1 and power delivery. Finally, the XPS 13 comes with a 52Whr battery, providing up to 17 hours of charge and is compatible with the included 65W charger.

It will start at only US$ 599 for eligible students.

SEE ALSO: Dell launches reimagined, refined XPS lineup

“Useful AI has arrived,” NVIDIA CEO Jensen Huang proclaimed in his monumental keynote speech to kick off COMPUTEX 2026, essentially admitting to the state of AI prior to today. But don’t let the past mess with how you view the present. AI today, according to the graphics-turned-AI company, will now serve everyday users, rather than procuring abstract and intangible benefits. For us normal people, it all starts with the new RTX Spark.

Rarely does a new product claim to reinvent the entire industry. Confident in its own abilities, that’s just what the RTX Spark wants to do. But what is it exactly?

Think of it like Microsoft’s Copilot+ PCs but on AI-powered steroids. NVIDIA wants to bring the power of a supercomputer into the hands of regular consumers, just like how the personal computer (PC) brought gigantic computers to the comfort of everyone’s homes.

Co-engineered with Microsoft, RTX Spark will run autonomous AI agents the entire day. The agent is tied to you and your intent but will orchestrate all your devices under one roof.

Let’s get down to the nitty-gritty. The new series will leverage a Blackwell RTX GPU and a 20-core Grace CPU from MediaTek. It can also support up to 128GB of unified memory, delivering several of the most powerful (and most expensive) PCs available to consumers today. With the capabilities comparable to an RTX 5070, the series can supposedly play modern games with ease.

Inside, it will come with CUDA, the trove of information that powers AI technology today. Finally, despite how slim NVIDIA promises the series will be, the RTX Spark should come with a battery that lasts the entire day.

The first NVIDIA RTX Spark laptops will launch later this fall with six premium laptops for various brands. This, along with the 30 models launching soon after, will include Acer, ASUS, Dell, HP, Lenovo, and MSI. No word yet on how much these will cost, though.

At COMPUTEX 2026, GIGABYTE is demonstrating how AI infrastructure is built, deployed, and operated at scale — from full-stack to real-world deployment.

This done through a comprehensive lineup of systems, software, and real-world deployments, turning the company’s “Future Landing” vision into reality.

From rack-scale AI factories and modular data centers to physical AI workflows, clinical applications, and on-prem AI agents, GIGABYTE is demonstrating how AI infrastructure moves beyond planning and into production.

Rack-scale AI

At the foundation of GIGABYTE’s “Future Landing” is an expanded portfolio of rack-scale AI infrastructure. These are designed for next-gen AI factories.

Part of the broader rack-scale portfolio is the NVIDIA Vera Rubin NVL72. The company is also presenting the GAIFA (GIGABYTE AI Factory Accelerator), built in Taiwan where next-generation compute, high-speed networking, and NVIDIA’s software stack are integrated.

Scaling AI infrastructure

For accelerated deployment beyond conventional data center construction, GIGABYTE introduces GADU (GIGABYTE Accelerated Deployment Unit).

This modular AI infrastructure platform integrates high-density compute, advanced cooling, and power distribution into transportable and deployment-ready systems.

GADU enables organizations to expand AI capacity. That’s with significantly reduced deployment timelines, while maintaining flexibility across diverse operational environments.

Furthermore, for generative and agentic AI systems, GIGABYTE is extending hardware system options for scaling. There’s the AMD Instinct MI355X and MI350P. These are both built on the fourth-generation CDNA architecture.

Physical AI from Simulation to Action

To demonstrate how AI moves beyond simulation to real-world scenarios, GIGABYTE also presented a complete “real-to-sim-to-real” workflow, built on NVIDIA’s full-stack AI platforms.

Here’s a quick rundown of the workflow:

• High-fidelity digital twins built through NVIDIA OVX systems for large-scale Omniverse simulations
• Training advances through platform on NVIDIA HGX architecture, where synthetic data generation and reinforcement learning take place
• Once validated, trained models are deployed to edge systems, powered by NVIDIA Jetson for real-time robotic control

Clinical AI

At the GIGABYTE booth, the company is also showcasing a clinical AI ecosystem for real-time medical inference at the point of care.

This Clinical AI works alongside healthcare partners. It integrates BRIX mini PCs with AI-assisted diagnostic systems supporting bone marrow smear classification and real-time polyp detection during colonoscopy.