Axelera AI, the leading provider of purpose-built AI hardware acceleration technology, today announced Europa, an AI processor unit (AIPU) that sets a new performance/price standard for multi-user generative AI and computer vision applications. Europa's combination of processing power, energy and thermal efficiency, compact packaging, and multiple form factor options make it ideal for compute-intensive, multi-modal AI inference applications from the edge to enterprise servers. Designed to be cost-effective and accessible, Europa enables any organization to unlock the power of cutting-edge AI applications.

Next Gen Architecture For Real World AI Processing Power
The Europa AIPU features eight second-generation AI cores, each incorporating Axelera's advanced Digital In-Memory Compute (D-IMC) technology and large vector engines, delivering up to 629 TOPS at INT8 precision. Complementing the AI cores are two clusters of eight dedicated RISC-V vector processing cores for non-AI pre- and post-processing, achieving a peak performance of 4915 GOPS. An integrated H.264/H.265 decoder further offloads media tasks, enabling customers to reserve the host CPU entirely for business logic and application-level workloads. These cores, programmable through the Voyager SDK, will provide better total cost of ownership than having to offload to other hardware.

JEDEC Solid State Technology Association, the global leader in standards development for the microelectronics industry, today announced it is nearing completion of JESD328: LPDDR5/5X Small Outline Compression Attached Memory Module (SOCAMM2) Common Standard, an upcoming standard for low-profile LPDRAM modules developed specifically for data center AI applications. When published, JESD328 is designed to provide a memory platform that delivers modular, low-power, high-bandwidth memory for AI CPU servers and accelerated computing platforms. JESD328 will define a compact, serviceable LPDDR5X-based module, Small Outline CAMM2 (SOCAMM2).

JESD328 SOCAMM2 is set to introduce an optimized, compact form factor for servers with a mechanical outline designed for high-density data center chassis and board layouts, with scalability to provide a path to support the large module capacities required by AI training and inference servers.

SK hynix showcased groundbreaking memory solutions tailored for the AI era at Dell Technologies Forum 2025 held on September 17 in Seoul, Korea. Dell Technologies Forum is a series of regional tech conferences linked to the company's annual flagship event Dell Technologies World in Las Vegas. The forum held in Seoul is one of Korea's largest IT conferences, inviting key stakeholders. Held under the theme "Beyond the Limits of Imagination," this year's event attracted more than 4,000 attendees for technical sessions, product demonstrations and immersive showcases.

During the forum, SK hynix presented its product portfolio for AI infrastructure under the slogan "Guardians of the AI - Full-Stack AI Memory Provider." The company's booth, featuring distinctive characters, captivated visitors by highlighting its core technologies in an accessible manner.

Micron Technology is pushing the boundaries of laptop performance with higher speeds for Crucial LPCAMM2 memory, now reaching up to 8,533 megatransfers per second (MT/s). This latest enhancement brings even more performance to Crucial's standards-based compact memory module, purpose-built to allow easy upgrades, efficient energy use and AI-ready computing.

With the new speed and densities up to 64 GB, Crucial LPCAMM2 allows users to run complex simulations, real-time AI tasks and data-intensive workloads with ease. This launch marks a major step forward in mobile memory innovation, offering 1.5 times faster performance and lower power consumption than traditional DDR5 SODIMMs. The modules deliver all the benefits of Micron's leading LPDDR5X mobile memory in a form factor that is less than half the size of standard SODIMMs, enabling users to upgrade, repair and extend the life of their devices, reducing e-waste.

Beelink has officially released the EQR7 mini PC, which was previously limited to China, for global markets, and it pitches the mini PC as a quiet workstation PC for office use, although it should be more than capable of some light gaming workloads, as well. The EQR7 is only available with 32 GB of soldered LPDDR5 RAM in dual channel and a 1 TB 2280 M.2 SSD, but it can be had from Beelink directly with an AMD Ryzen 7 7735U for $339 or AMD Ryzen 5 7535U for $329. It is also available on Amazon, albeit at a slightly higher price. While RAM is soldered, the EQR7 has dual 2280 M.2 slots for expandable storage. Those AMD APUs mean that the EQR7 is available with up to an AMD Radeon 680M, which is more than capable of casual gaming, even if it is showing its age.

Despite the small form factor, measuring in at just 126 mm × 126 mm and 44.2 mm tall, the Beelink EQR7 features decent connectivity, with dual gigabit Ethernet ports, Wi-Fi 6, and Bluetooth 5.2. The physical ports include a front-mounted USB 3.2 Gen 2 type-A port, a USB 3.2 Gen 2 Type-C port, and a 3.5 mm combo audio jack. On the rear, there are dual HDMI 2.0 ports, dual USB 3.2 Gen 2 ports, and a solitary USB 2.0 port, alongside the aforementioned Ethernet ports and an AC power input. That AC power input also conveniently connects directly to the wall, with no power adapter necessary. There are no USB4 ports, so gamers can forget about connecting an eGPU for additional performance, but the EQR7 isn't really marketed as a gaming mini PC, to begin with. Beelink says that the MCS2.0 cooling system and the "silent" internal fan allow the EQR7 to operate at 32 dB, but it's unclear whether that is under load or at idle, so this claim should be taken with a grain of salt.

TrendForce's latest investigations reveal that the three major DRAM suppliers continue to allocate advanced process capacity primarily to high-end server DRAM and HBM, crowding out capacity for PC, mobile, and consumer applications. Combined with diverging demand across end products, this dynamic is keeping legacy-process DRAM price hikes pronounced, while newer-generation products are seeing more modest gains. Overall, conventional DRAM prices are expected to rise 8-13% QoQ in 4Q25, and when HBM is included, the increase could widen to 13-18%.

PC DRAM prices edge up slightly, U.S. CSPs consider early server DRAM procurement
Shipments of finished PCs are projected to decline with PC promotions and restocking momentum slowing in 4Q25, leading major OEMs to cut DRAM bit procurement. Meanwhile, as the three suppliers aggressively shift capacity toward server DDR5, the supply of both PC DDR5 and DDR4 remains constrained. As a result, PC DRAM prices are forecast to continue edging upward through the quarter.

SK Hynix is installing a new pilot production line at its M15X facility in Cheongju, bringing online a new 1b DRAM manufacturing line designed to address surging HBM demand, The Bell reports. Equipment installation is scheduled to be completed by November with the pilot line targeting an initial monthly output of 10,000 wafers. The move highlights different approaches among memory leaders, while Samsung pursues 1c DRAM technology for HBM4 applications, both SK Hynix and Micron have committed to 1b DRAM architecture, according to Aju News. This difference reflects varying technical roadmaps as companies race to capture as much as they can from the expanding AI memory markets.

On the other hand it seems that memory price adjustments have become a widespread practice. TrendForce reports citing SeDaily and Business Korea notes that SK Hynix has entered price negotiations with customers, however, no formal announcement has been made yet. The company is tweaking its prices in line with current market trends. This hints at a unified industry shift towards pricier memory across all big-name suppliers. Micron kicked things off in September first letting its channel partners know about 20-30% price hikes for DRAM. Samsung followed by informing major customers that Q4 contract prices for DRAM products, including LPDDR4X and LPDDR5/5X variants, would rise 15-30%. NAND pricing isn't immune either—eMMC and UFS contract prices are expected to climb 5-10%. On the same note, TrendForce indicates suppliers are ramping QLC SSD production, with capacity utilization projected to continue its growth through 2026. Inference AI workloads are expected to sustain demand into 2027, potentially creating enterprise SSD supply constraints by 2026.

According to Korean media outlet News Daily, Samsung Electronics has confirmed that it has raised contract prices for DRAM by as much as 15-30% and mobile NAND flash by roughly 5-10%, in response to a noticeably tighter supply/demand balance. Samsung reportedly notified major customers ahead of fourth-quarter contracting, following a market shift, where older-generation parts are being phased down while demand from hyperscale cloud operators and device makers racing to ship AI-enabled PCs and flagship phones has surged. Additionally, Samsung's rivals such as SK Hynix and Micron have signaled similar adjustments, with several suppliers notifying clients of double-digit increases and temporary limits on new quotes. For buyers, the result is immediate price pressure across server and mobile storage purchases, while for vendors, it is a tactical reprioritization of wafer capacity toward higher-margin, next-generation product lines.

Samsung's DRAM LPDDR4X, LPDDR5, and LPDDR5X families, as well as embedded NAND formats such as eMMC and UFS, are among those experiencing contract uplifts. Meanwhile, shortages in commodity DDR4 briefly pushed spot quotes sharply higher. Memory makers have redirected capacity to DDR5 and HBM, prioritizing HBM orders used by AI accelerators, tightening the supply of consumer-focused DRAM. Samsung, which retains roughly a third of global DRAM and NAND market share, is accelerating development of next-generation LPDDR6 even as it tweaks production schedules for legacy nodes. The pricing reset is expected to materially improve semiconductor division margins in the near term, cushioning against earlier cyclical weakness. For customers, it will force revisions to inventory planning and, likely, a renewed focus on design choices for price/performance.

During yesterday's Q&A with Intel and NVIDIA CEOs, NVIDIA CEO Jensen Huang confirmed that the current GB10 "Grace Blackwell" Superchip is the base design of the upcoming N1 laptop processor. "We also have a new Arm product that's called N1. And that product is—that processor is going to go into the DGX Spark and many other versions of products like that. And so we're super excited about the Arm road map, and this doesn't affect any of that," said Jensen Huang. This means that the GB10 Superchip is the blueprint for the design of N1, and we can already create a rough performance tier based on the previous information about the GB10.

At Hot Chips 2025, NVIDIA gave more details about GB10. The chip pairs a MediaTek-sourced Arm CPU die with a Blackwell GPU die in a 2.5D package built on TSMC's 3 nm process. The CPU side implements 20 Arm v9.2 cores split into two clusters of ten, each cluster backed by a 16 MB shared L3 (32 MB total) while each core retains private L2 storage. The memory subsystem is a unified LPDDR5X-9400 fabric on a 256-bit bus, supporting up to 128 GB and delivering roughly 301 GB/s of raw bandwidth to the package. For consumer laptops, we need to see if this memory capacity remains an option. High-speed I/O is concentrated on the CPU die, while NVMe storage and peripherals ride PCIe 5.0 lanes there. The package is rated at approximately 140 W TDP and exposes multi-display outputs (DisplayPort alt-mode plus HDMI 2.1a options) alongside security and virtualization features intended for professional workloads.

On September 17, the LattePanda Team officially launched the new LattePanda IOTA, a palm-sized x86 single-board computer powered by the Intel N150 quad-core processor. Compared to the previous-generation, LattePanda IOTA offers up to 8x the performance, and provides 2-3x faster CPU speeds and 10-30x better GPU capabilities than the Raspberry Pi 5. It is available with either 8 GB or 16 GB of LPDDR5 memory and 64 GB or 128 GB of eMMC storage, bringing powerful computing to a palm-sized form factor.

LattePanda IOTA maintains full compatibility with the classic LattePanda V1 footprint, including dimensions, interfaces, and mounting holes, making upgrades straightforward. The board features HDMI 2.1 and eDP display output, USB 3.2 Gen 2, Gigabit Ethernet, an M.2 expansion slot, and a GPIO header supporting UART, I2C, and USB 2.0. Additionally, it includes a built-in RP2040 coprocessor, broadening its range of potential applications. With its efficient x86 architecture, extensive I/O options, and long-term supply stability, LattePanda IOTA is well suited for prototyping, edge computing, smart instrumentation, handheld devices, industrial control, and embedded systems.

NVIDIA has reportedly discontinued its first-generation Small Outline Compression Attached Memory Modules, known as SOCAMM. According to ET News, NVIDIA is shifting focus to the second-generation SOCAMM2 standard, which is being revamped for higher memory speeds, resulting in significantly improved bandwidth and performance. The first-generation SOCAMM was developed independently by NVIDIA, outside of any official memory consortium like JEDEC. It featured a detachable module design, offering superior specifications compared to existing solutions, with 694 I/O ports (compared to LPCAMM's 644 and traditional DRAM's 260), direct LPDDR5X memory substrate integration, and a more compact form factor.

NVIDIA aims for SOCAMM2 to become a faster, more widely adopted standard, potentially with JEDEC's assistance to standardize the format. SOCAMM2 is expected to maintain the same 649 I/O ports exposed to the motherboard but will integrate faster LPDDR5X running at 9600 MT/s, compared to the first-generation SOCAMM's 8533 MT/s. The new SOCAMM2 might also use LPDDR6 and source it from multiple suppliers. NVIDIA currently ships SOCAMM in its "Blackwell Ultra" GB300 NVL72 system, providing 18 TB of LPDDR5X memory at 14.3 TB/s bandwidth. An upgrade to SOCAMM2 could increase that to 16 TB/s bandwidth as a simple drop-in replacement, stemming from 9600 MT/s LPDDR5X. Additionally, NVIDIA may be diversifying its supplier strategy. The first-gen SOCAMM was supplied by Micron, but now Samsung and SK Hynix might join, although no specific details or timeline are available.

BIOSTAR, a leading manufacturer of edge computing solutions, industrial motherboards, graphics cards, and storage devices, introduces the EdgeComp MS-NANX 8G/16G, a high-performance, fanless edge AI system accelerated by NVIDIA Jetson Orin NX platform with Super Mode support. Delivering up to 157 TOPS of AI performance in a rugged and compact form factor, the MS-NANX is engineered to accelerate real-time AI workloads across smart retail, industrial automation, transportation, smart city infrastructure, and modern agriculture.

Designed for system integrators and AI software developers, the EdgeComp MS-NANX 8G/16G offers scalable performance to meet a wide range of edge computing needs. Supporting both NVIDIA Jetson Orin NX 16 GB and 8 GB modules, it features a 1024-core NVIDIA Ampere Architecture GPU with 32 Tensor Cores, paired with up to an 8-core Arm Cortex-A78AE CPU and 128-bit LPDDR5 memory delivering up to 102.4 GB/s bandwidth. This hardware platform is backed by a comprehensive range of expansion options, including M.2 slots for NVMe SSDs, Wi-Fi and Bluetooth modules, and 4G or 5G connectivity via Nano SIM slot. Extensive I/O support includes two Ethernet ports (1 Gigabit Ethernet and 2.5 Gigabit Ethernet with optional Power over Ethernet), four USB 3.2 Type-A ports, one HDMI 2.1 port, one Micro USB port for OS flashing, one RS232 port with CAN interface, one RS232/422/485 combo port, and a 3.5 mm audio jack with microphone and line-out support. The system is built for reliable integration across industrial, commercial, and public sector environments.

DIGIERA today unveiled HoloMax, the world's first glasses-free 3D laptop-handheld hybrid, at IFA 2025. HoloMax pairs an immersive 10.95" 2.5K, 120 Hz autostereoscopic display with real-time eye-tracking and a high-performance compute platform to convert standard 2D titles into natural stereoscopic 3D with a single button—no headsets required.

HoloMax's autostereoscopic 3D pipeline with eye-tracking keeps visuals stable across viewing angles and motion, delivering instant depth for games, video, and creative previews. The device is engineered for demanding experiences while remaining portable: it combines console-class compute with flexible form factors—detachable magnetic controllers, an adjustable stand, and a magnetic snap-on keyboard to switch between handheld and desktop workflows in seconds.

MeLE has launched the Cyber X1, a compact fanless mini PC powered by Intel's N150 processor from the Twin Lake lineup. The system measures 131 x 81 x 24 mm and weighs 288 grams making it barely larger than a smartphone. The Cyber X1 features 16 GB of soldered LPDDR5-4800 memory and includes a single M.2 2280 slot for user-replaceable PCIe 3.0 NVMe storage, with the initial configuration offering 512 GB capacity. MeLE has configured the Intel N150 to run at 10 W TDP, higher than the typical 6 W base power rating, enabled by an enhanced passive cooling solution featuring hundreds of thermal-conductive plastic spikes on the top surface.

Connectivity options include dual HDMI 2.0 outputs supporting up to three displays total, one USB 3.2 Gen 2 Type-C port with 10 Gbps data transfer and video output, two additional USB 3.2 Type-A ports, one USB 2.0 Type-A port, Gigabit Ethernet, 3.5 mm audio jack, and a microSD card reader capable of 60 MB/s speeds. Wireless connectivity offers WiFi 5 and Bluetooth 5.1 support. The design builds upon MeLE's earlier Quieter 4C model but incorporates improved thermal management allowing higher sustained performance. While pricing hasn't been officially announced globally, European listings show €299.99, approximately $349. An Amazon product page suggests North American availability is imminent, though no specific launch date has been confirmed.

Acer today announced a suite of commercial products at IFA 2025, designed to empower the modern workforce. The lineup includes the Acer TravelMate X14 AI Copilot+ PC and three new additions to the eco-friendly Vero projector series: the Acer Vero XL2320p, Vero XL2521, and Vero SL2520n.

"As AI is becoming increasingly important in modern business workflows, small and medium businesses need devices that integrate intelligent tools, robust security, and mobility," said James Lin, General Manager, Notebooks, Acer Inc. "The TravelMate X 14 AI was designed to deliver just that, powered by the latest Intel Core Ultra processors and Copilot+ PC experiences, offering professionals a lightweight yet durable solution crafted to accelerate productivity and collaboration from anywhere."

Acer today announced a broad expansion of its consumer product portfolio during its next@acer global press conference at IFA. It included various additions to its Iconia X and Iconia A series tablets, a pair of elegant monitors from amadana, the high-performance Acer CE270U Z OLED display, and the Acer Connect T36 Wi-Fi 7 Dual-Band Mesh Router - all designed to elevate how people work, learn, and stay connected.

Acer Iconia X Series Tablets
The Iconia X series tablets are crafted to support both productivity and entertainment, powered by Android, vivid displays, and AI-enabled features that enhance everyday experiences.

A fresh leak has shed light on AMD's next-gen Medusa Halo APU that is set to launch in 2027 as the company's top-of-the-line chip (dismissing previous rumors about AMD cancelling Medusa Halo APU). Moore's Law is Dead has shared information suggesting Medusa Halo will pack Zen 6 CPU chiplets made using TSMC's cutting-edge N2P process, with the I/O die will be made using the N3P process. The base Medusa Halo APUs have 12 Zen 6 cores alongside 2 power-efficient Zen 6 LP cores, while high-end versions might include an extra 12-core Zen 6 CCD. This means AMD will provide Medusa Halo APUs with 24 CPU cores (and possible even up to 26 cores).

For graphics, the Medusa Halo APUs will feature 48 compute units (CUs) based on RDNA 5 architecture alongside 20 MB of L2 cache. This represents a significant jump from the 40 CUs in today's Strix Halo APU and should deliver graphics power on par with an NVIDIA GeForce RTX 5070 Ti, a big step forward for built-in GPUs. Memory support is in line for a big boost, with Medusa Halo set to offer either a 384-bit LPDDR6 or 256-bit LPDDR5X memory controller. This will provide high bandwidth to match its upgraded GPU needs.

Corsair, a world leader in advanced performance PCs for AI developers, engineers, and professionals, today announced the release of the Corsair AI WORKSTATION 300 desktop PC for Europe, an AI-optimized PC engineered to deliver a powerful AI workstation solution in a compact 4.4-liter form factor.

The Corsair AI WORKSTATION 300 is powered by AMD Ryzen AI Max 300 Series processors, offering configurations up to the AMD Ryzen AI Max+ 395 with 96 GB of Variable Graphics Memory, powerful RDNA 3.5 graphics technology with 40 compute units, and features cutting-edge XDNA 2 NPU architecture delivering up to 50 TOPS of AI acceleration. Thanks to AMD's "Strix Halo" memory capabilities, the system is specifically optimized for advanced client-side AI workflows such as LM Studio and Stable Diffusion. With up to 128 GB of unified LPDDR5X memory (96 GB can be dynamically allocated as VRAM), it supports large-scale model inference workloads that standard desktop GPUs simply can't handle.

TrendForce's latest findings reveal the DDR4 market is set to remain in a persistent state of undersupply and strong price growth through 2H25. Rigid server orders are crowding out supply for the PC and consumer markets, forcing PC OEMs to accelerate DDR5 adoption.

Manufacturers of consumer electronics, meanwhile, are struggling with high prices and material shortages. The tight DRAM supply-demand balance has also pushed up mobile DRAM contract prices, with 3Q25 LPPR4X posting the largest single-quarter increase in nearly a decade.

Innodisk, a leading provider of industrial-grade memory solutions, officially releases its new DDR5 CAMM2 and LPDDR5X CAMM2 (LPCAMM2) memory modules. With 60% less space and blazing speeds up to 6400 MT/s and 8533 MT/s, respectively, Innodisk's CAMM2 series marks a strategic leap into the rugged device and compact system market.

Innodisk's new CAMM2 series features a distinctive design. DDR5 CAMM2 and LPDDR5X CAMM2 consolidate dual channels on a single module, effectively combining the capacity and performance of two separate SODIMMs into a compact footprint. Despite handling a greater signal load, the modules feature a more simplified PCB routing that enhances signal integrity, resulting in cleaner and more reliable data transmission.

Samsung Electronics today reported financial results for the second quarter ended June 30, 2025. The Company posted KRW 74.6 trillion in consolidated revenue, a decrease of 5.8% compared to the previous quarter. Operating profit decreased to KRW 4.7 trillion.

The Device Solutions (DS) Division reported an increase in revenue on the back of expanded sales of high density, high-performance memory products, but inventory value adjustments in memory and one-off costs related to the impacts of export restrictions related to China in non-memory had an adverse effect on profit. In the Device eXperience (DX) Division, operating profit declined quarter-on-quarter due to a sequential decline in sales volume following the launch of new smartphone models in the first quarter. Looking ahead to H2, the DS Division plans to proactively meet the growing demand for high-value-added and AI-driven products and continue to strengthen competitiveness in advanced semiconductors. The DX Division will seek to minimize the impact of uncertainties stemming from tariff policies that are likely to persist.

New leaks about the upcoming ASUS ROG Xbox Ally point to an earlier-than-expected pre-order date. According to Dealabs and @billbil_kun on X, the Xbox-branded gaming handheld will be available for pre-order on August 20, aligning with the opening day of Gamescom 2025. Xbox Wire notes that there is a Gamescom presentation "featuring the ROG Xbox Ally handhelds," scheduled for August 20 at 3 PM CEST. The new gaming handheld will be available in two versions. The eight-core AMD Ryzen AI Z2 Extreme will power the Xbox Ally X, and the quad-core AMD Ryzen Z2 A will be the core of the vanilla Xbox Ally. The Xbox Ally also comes with less RAM and storage than the Xbox Ally X, with the latter featuring 24 GB of LPDDR5x-8000 RAM and a 1 TB SSD, while the former has just 16 GB of LPDDR5x-6400 and a 512 GB SSD.

With the exception of the single USB4 port on the Xbox Ally X not making it to the Xbox Ally, and a larger 80 Wh battery in the Ally X, the two devices are otherwise more or less identical. Previous leaks have indicated that the expected price of the Xbox Ally X will fall around the €899 mark, while the less powerful Xbox Ally will cost around €599 in Europe. It's unclear how these prices will translate to the US, but the current All-Digital Xbox Series X costs €549 in the EU and $549 in the US, so it's likely that the US pricing for the ASUS ROG Xbox Ally and Xbox Ally X will also be a 1:1 conversion rather than following regional pricing.

Following yesterday's launch of the AMD Ryzen AI Max+ 395-powered Ayaneo Next 2, Corsair has announced yet another small-form-factor PC powered by AMD's Strix Halo APU. The Corsair AI Workstation 300, however, is a small-form-factor desktop PC measuring in at just 96.5 × 188.4 × 247.5 mm and launching at a pricey $1,599.99. That starting price only gets you into the eight-core AMD Ryzen AI Max 385, though, with the 16-core Ryzen AI Max+ 395 coming in at a staggering $1,999.99. The system offers up to 128 GB of RAM and 4 TB of NVMe SSD storage. The fully kitted-out system with the AI Max+ 395 APU, 128 GB RAM, and 4 TB SSD comes in at $2,299.99 from Corsair directly. The whole system is also powered by an internal 350 W flex ATX PSU, which should give the APU plenty of power for intensive tasks without having to have an external power brick.

Corsair's model is based on the same chassis as a design by Sixunited, whose Strix Halo Mini ITX motherboards were recently featured here as well. Just like those aforementioned motherboard designs, the Corsair AI Workstation 300 is built around LPDDR5x-8000 RAM, and even the starting configuration features 64 GB of RAM, communicating the workstation's intended workload as an AI workstation rather than a compact gaming system. The AI Workstation 300 ships with Corsair's AI Software Suite, which includes Amuse AI, an image generator based on Stable Diffusion and AMD Super Resolution, and Jan AI, an offline, open-source AI assistant based on Llama 3 and Mistral. Given the limitations of the small chassis, the AI Workstation 300 features a reasonably healthy IO selection, with an SD 4.0 card reader, 1× USB4 port, 2× USB 3.2 Gen 2 Type-A ports, and a 3.5 mm combo jack on the front. The rear has 1× HDMI 2.1, 1× DP 1.4, 1× USB4, 1× USB 3.2 Gen 2 Type-A, 1× 2.5 GbE, and 1× 3.5 mm headphone/microphone combo jack.

Edge computing leader AAEON (Stock Code: 6579), today announced another breakthrough with the release of the PICO-MTU4-SEMI, a compact Mini PC powered by the Intel Core Ultra 5 Processor 125U.

Measuring just 108 mm x 95 mm x 43 mm, the PICO-MTU4-SEMI's size is particularly notable, and not only because it is the smallest Mini PC equipped with Intel Core Ultra CPUs released to date. Despite the PC's high-performance processor base, the PICO-MTU4-SEMI is the company's first Intel Core-powered PICO-SEMI PC with fanless heat dissipation, allowing the system to be integrated into larger setups such as industrial robots and AGVs while still providing a broad selection of interfaces for peripheral devices such as cameras and sensors.

The Lenovo Legion Go 2 was officially revealed at CES 2025 earlier this year, confirming that the new device would be launching later this year with the AMD Ryzen Z2 Extreme. Now, however, it seems as though prototype units of the Legion Go 2 have started leaking in China, with one Chinese content creator on Bilibili managing to publish a video review of the device on the social media platform. The reviewer gives us a look at the internals and specifications of the device, confirming that it will ship with a 144 Hz OLED display with a resolution of 1,920 × 1,200, a peak brightness of 500 nits, and 97% coverage of the DCI-P3 color space.

The review unit in question was apparently sold on a Chinese secondhand marketplace called GoFish after the production facility where it was produced shut down—according to one Redditor's claims, at least—and it apparently had the regular AMD Ryzen Z2 SoC, not the Z2 Extreme that is expected to power devices like the ASUS ROG Xbox Ally gaming handheld when that eventually launches. As expected, though, the Legion Go 2 features 32 GB of LPDDR5x-7500 memory and an AMD Radeon 780M iGPU. The battery capacity also seems to have been given a substantial upgrade to 74 Wh, compared to the 49.2 Wh in the original Legion Go. Judging by the stickers on the back of this particular pre-production unit, it was manufactured in November 2024, so there have likely been significant changes to the design and internals since then, and the information presented in the Bilibili video can be taken with a healthy helping of salt.