Double the Speed with New PCIe 6.0 SSD Technology

 Micron and Astera Labs joined forces at DesignCon 2025 to unveil the world's fastest PCIe 6.0 SSD. Although the product is not yet market-ready, it’s worthwhile to examine the technology’s basics and its future prospects.

Last August, Micron introduced the first prototype of a PCIe 6.0 SSD, and now a new version has been released with further improved read speeds. In the latest demonstration, the test setup featured two Micron PCIe 6.0 SSDs and an Nvidia H100 GPU connected through an Astera Labs Scorpio P-Series Fabric Switch. As a result, sequential read speeds exceeded 27 GB/s—double the performance of the fastest PCIe 5.0 drives available today, which offer a maximum bandwidth of 14–15 GB/s.

While these results seem impressive, it is important to note that they were achieved under controlled conditions. In real-world use, additional factors—such as temperature, power consumption, and system integration—may affect overall performance.

Another key aspect is the technology used during testing. The new Scorpio P-Series Fabric Switch plays a crucial role in the PCIe 6.0 solution. It can handle up to 64 PCIe 6.0 lanes simultaneously, which helps boost the system’s overall bandwidth and maintain signal integrity. This technology not only improves communication between storage devices but also optimizes data transfer among the CPU, GPU, and other components.

Another significant factor is Nvidia’s Magnum IO GPUDirect technology, which allows direct memory access between the GPU and storage devices, reducing the latency caused by the CPU. This method is essential for achieving high-speed data transfer, especially in AI and high-performance computing (HPC) systems.

Currently, these SSDs are not available for purchase, and several challenges remain before they can be widely deployed. For example, proper thermal management is crucial because increased data throughput generates significant heat. 

Share this post
Artificial intelligence, space, and humanity
Elon Musk, founder and CEO of SpaceX, Tesla, Neuralink, and xAI, shared his thoughts on the possible directions of the future in a recent interview, with a particular focus on artificial intelligence, space exploration, and the evolution of humanity.
 Answers while You train, Meta and Oakley's new AI smart glasses are here
Meta's latest product announcement makes it clear: the company firmly believes in the proliferation of wearable technology. Created in collaboration with Oakley, Oakley Meta HSTN smart glasses are aimed at users with an active lifestyle and offer features that combine everyday usability with the possibilities of artificial intelligence.
The new AI chip that is revolutionizing medicine and telecommunications makes decisions in nanoseconds
As more and more devices connect to the internet and demand grows for instant, high-bandwidth applications such as cloud-based gaming, video calls, and smart homes, the efficient operation of wireless networks is becoming an increasingly serious challenge. The problem is further exacerbated by the fact that the wireless spectrum—the available frequency band—is limited. In their search for a solution, engineers are increasingly turning to artificial intelligence, but current systems are often slow and energy-intensive. A new development that brings data transmission and processing up to the speed of light could change this situation.
Trump Mobile T1: The Chinese are already in Trump Tower
In June 2025, the Trump family announced a new mobile communications brand called Trump Mobile, whose flagship product is a smartphone called the T1. The device and its accompanying mobile package were unveiled at Trump Tower, timed to coincide with the 10th anniversary of the former president's first presidential campaign. The goal of the project, at least on a communications level, is to encourage American manufacturing and job creation, as well as to boost the domestic technology sector.
Tesla's new FSD chip delivers five times more computing power
Tesla's next-generation FSD (Full Self-Driving) chip, the AI5/HW5, represents a significant step forward in the development of the company's self-driving technology. Although the previous model, known as HW4, already offered powerful performance on its own, the new chip's 2000-2500 TOPS (tera operations per second) computing power effectively sets a new standard for in-vehicle artificial intelligence applications. This fivefold increase is not just a technical feat, but also allows the system to work with more complex, finely tuned algorithms that require less human intervention – algorithms that bring us closer to the possibility of completely unsupervised driving.
New DRAM for next-generation AI GPUs
NVIDIA has commissioned Samsung Electronics, SK hynix, and Micron to develop SoCEM. This is a new memory module standard invented by NVIDIA, and Micron is reportedly the first to receive a license for mass production. This puts the American memory manufacturer ahead of its South Korean rivals.

Linux distribution updates released in the last few days