Meta has announced the next generation of its Meta Training and Inference Accelerator (MTIA) chip, which is designed to train and infer AI models at scale. The newest MTIA chip is a second-generation design of Meta's custom silicon for AI, and it is being built on TSMC's 5 nm technology. Running at the frequency of 1.35 GHz, the new chip is getting a boost to 90 Watts of TDP per package compared to just 25 Watts for the first-generation design. Basic Linear Algebra Subprograms (BLAS) processing is where the chip shines, and it includes matrix multiplication and vector/SIMD processing. At GEMM matrix processing, each chip can process 708 TeraFLOPS at INT8 (presumably meant FP8 in the spec) with sparsity, 354 TeraFLOPS without, 354 TeraFLOPS at FP16/BF16 with sparsity, and 177 TeraFLOPS without.

Classical vector and processing is a bit slower at 11.06 TeraFLOPS at INT8 (FP8), 5.53 TeraFLOPS at FP16/BF16, and 2.76 TFLOPS single-precision FP32. The MTIA chip is specifically designed to run AI training and inference on Meta's PyTorch AI framework, with an open-source Triton backend that produces compiler code for optimal performance. Meta uses this for all its Llama models, and with Llama3 just around the corner, it could be trained on these chips. To package it into a system, Meta puts two of these chips onto a board and pairs them with 128 GB of LPDDR5 memory. The board is connected via PCIe Gen 5 to a system where 12 boards are stacked densely. This process is repeated six times in a single rack for 72 boards and 144 chips in a single rack for a total of 101.95 PetaFLOPS, assuming linear scaling at INT8 (FP8) precision. Of course, linear scaling is not quite possible in scale-out systems, which could bring it down to under 100 PetaFLOPS per rack.Below, you can see images of the chip floorplan, specifications compared to the prior version, as well as the system.

During the Vision 2024 event, Intel announced its latest Gaudi 3 AI accelerator, promising significant improvements over its predecessor. Intel claims the Gaudi 3 offers up to 70% improvement in training performance, 50% better inference, and 40% better efficiency than Nvidia's H100 processors. The new AI accelerator is presented as a PCIe Gen 5 dual-slot add-in card with a 600 W TDP or an OAM module with 900 W. The PCIe card has the same peak 1,835 TeraFLOPS of FP8 performance as the OAM module despite a 300 W lower TDP. The PCIe version works as a group of four per system, while the OAM HL-325L modules can be run in an eight-accelerator configuration per server. This likely will result in a lower sustained performance, given the lower TDP, but it confirms that the same silicon is used, just finetuned with a lower frequency. Built on TSMC's N5 5 nm node, the AI accelerator features 64 Tensor Cores, delivering double the FP8 and quadruple FP16 performance over the previous generation Gaudi 2.

The Gaudi 3 AI chip comes with 128 GB of HBM2E with 3.7 TB/s of bandwidth and 24 200 Gbps Ethernet NICs, with dual 400 Gbps NICs used for scale-out. All of that is laid out on 10 tiles that make up the Gaudi 3 accelerator, which you can see pictured below. There is 96 MB of SRAM split between two compute tiles, which acts as a low-level cache that bridges data communication between Tensor Cores and HBM memory. Intel also announced support for the new performance-boosting standardized MXFP4 data format and is developing an AI NIC ASIC for Ultra Ethernet Consortium-compliant networking. The Gaudi 3 supports clusters of up to 8192 cards, coming from 1024 nodes comprised of systems with eight accelerators. It is on track for volume production in Q3, offering a cost-effective alternative to NVIDIA accelerators with the additional promise of a more open ecosystem. More information and a deeper dive can be found in the Gaudi 3 Whitepaper.

Samsung Electronics America, the leader in advanced memory technology, today introduced the latest iteration of its EVO Select and EVO Plus microSD cards, designed for mobile devices and handheld gaming consoles. The Samsung EVO Select and EVO Plus microSD cards boast increased transfer speeds of up to 160 MB/s, a 23% increase over their predecessors. They offer a turnkey way for everyday users and gamers to add storage space to their devices for content like games, files, photos, and videos.

"With the mobile and handheld gaming market on the rise, we sought to make data transfers on those devices even faster with these updated EVO Select and EVO Plus lines," said Jose Hernandez, Senior Director of Memory Product Marketing at Samsung. "We also heard the need for more space options like a 1 TB memory card for digital files like games, video footage from phones, cameras and drones, photos, and creative designs. With this addition to the lineup, you can be confident you'll have plenty of space, fast speeds and the durability needed to last for years to come."

Earlier teardowns of 13-inch Apple M3 MacBook Air models revealed a pleasing storage performance upgrade. Popular American e-commerce watchdog, iFixit, has joined in on the fun—CEO Kyle Wiens has confirmed that the 15-inch M3 MacBook Air 256 GB base model also houses two 128 GB flash storage chips. A provisional repairability score of 5 out of 10 has been awarded—this verdict could change once iFixit staffers finish up in-depth investigations into Apple's latest thirteen and fifteen inch ultraslim notebooks. A revised figure might appear online once the site publishes its full how-to-guides.

iFixit's video teardown demonstrates that not much has changed when comparing the new models to Apple's M2 MacBook Air family of products. Tinkerers will face the usual obstacles, mainly dreaded pentalobe screw designs. The team discovered plastic pull-tabs during the removal of M3 MacBook Air batteries—a pleasing alternative to older (headache inducing) adhesive-fastened methods of securing power cells in place. The iFixit team had to deal with many fiddly screws and brackets during excavation efforts—they noted that Apple's interior design does not include any labelling, and the screws are not numbered. Framework's Laptop 16 was cited as a shining example of doing things correctly.

ASUS today announced BIOS updates that enable support for up to 256 GB of memory on its Intel 700 and 600 series motherboards that use DDR5 modules. Such models with four DIMM slots can now support up to 256 GB of memory, while such models with two DIMM slots can now support up to 128 GB. These enhancements significantly improve multitasking potential, ensuring smooth and seamless computing. AMD AM5 motherboards from ASUS do not require a BIOS update to enable support for up to 256 GB of DDR5 memory modules.

The BIOS updates can be accessed on the ASUS support pages for the models listed below.

CTL, a global cloud computing solution leader for education and enterprise, will demonstrate the world's first Chromebook supporting 5G RedCap connectivity powered by, which was the world's first announced commercial release 17 5G RedCap modem RF system. CTL is preparing its Chromebook product lines for the future, where 5G will replace LTE connectivity. With more schools and enterprises standing up private cellular networks for anytime, anywhere connectivity, CTL is at the forefront of bringing built-in 5G connectivity to Chromebook devices.

Erik Stromquist, CTL's CEO, said, "CTL is committed to developing and introducing affordable solutions for the education and enterprise markets. We believe there is a big opportunity to address the digital divide by providing more advanced and affordable private network solutions to extend internet access to every user regardless of location. We are thrilled to collaborate with Qualcomm Technologies in demonstrating 5G next-gen possibilities at this year's Mobile World Congress."

An exclusive report provides an initial tease of MSI's relaunch of MPOWER—a beloved product line of high performance yet wallet friendly motherboards. Wccftech published their Z790MPOWER model coverage only a few hours ago. MSI's final batch of MPOWER-branded boards landed back in 2017, with Z170 and Z270 chipsets (on Intel Socket 1151). Here is Wccftech's statement on the matter: "MSI is marking the return of the MPOWER series with a new and cost-effective Z790 product, the Z790MPOWER. This motherboard may look like a very mainstream design but it has something that only a few high-end motherboards can do and that is support for the best DDR5 memory out there."

They moved onto showcasing the board's feature set: "Starting with the details, the MSI Z790MPOWER motherboard features the LGA 1700/1800 socket & supports 12th, 13th, and 14th Gen CPUs from Intel. It is powered by a 15-Phase VRM design which is provided power through dual 8-pin connectors. There are large heatsinks over the VRMs and the Mini-ATX with a silver and black finish looks great." The usual bits of overclocking terminology adorn Z790MPOWER's various heatsinks—including "Overclock, Frequency, MHz, Voltages and Clock."

Chinese company Sipeed has introduced the Lichee Pocket 4A, one of the first handheld gaming devices based on the RISC-V open-source instruction set architecture (ISA). Sipeed positions the device as a retro gaming platform capable of running simple titles via software rendering or GPU acceleration. At its core is Alibaba's T-Head TH1520 processor featuring four 2.50 GHz Xuantie C910 RISC-V general-purpose CPU cores and an unnamed Imagination GPU. The chip was originally aimed at laptop designs. Memory options include 8 GB or 16 GB LPDDR4X RAM and 32 GB or 128 GB of storage. The Lichee Pocket 4A has a 7-inch 1280x800 LCD touchscreen, Wi-Fi/Bluetooth connectivity, and an array of wired ports like USB and Ethernet. It weighs under 500 grams. The device can run Android or Linux distributions like Debian, Ubuntu, and others.

As an early RISC-V gaming entrant, performance expectations should be modest—the focus is retro gaming and small indie titles, not modern AAA games. Specific gaming capabilities remain to be fully tested. However, the release helps showcase RISC-V's potential for consumer electronics and competitive positioning against proprietary ISAs like ARM. Pricing is still undefined, but another Sipeed handheld console retails for around $250 currently. Reception from enthusiasts and developers will demonstrate whether there's a viable market for RISC-V gaming devices. Success could encourage additional hardware experimentation efforts across emerging open architectures. With a 6000 mAh battery, battery life should be decent. Other specifications can be seen in the table below, and the pre-order link is here.

GIGABYTE Technology, a leading global powerhouse in motherboards, graphics cards, and cutting-edge hardware solutions, proudly announces that GIGABYTE DDR5 motherboards deliver breakthrough support for a memory capacity of single DIMM 64 GB with the latest Kingston FURY Renegade DDR5 memory. The 4DIMM model boosts the memory capacity up to 256 GB, while the 2DIMM model has a maximum capacity of up to 128 GB.⁠

With innovative layout design and top-notch components, GIGABYTE motherboards have played a leading role in memory performance. From the Intel 700 and 600 series to the AMD 600 series, GIGABYTE DDR5 motherboards will support a memory capacity of single DIMM 64 GB through BIOS update. The updated BIOS and memory support list will be available on the GIGABYTE official website, coinciding with the launch of the latest Kingston FURY Renegade DDR5 memory. Please stay tuned to the GIGABYTE official website for further information.

GIGABYTE Technology: Giga Computing, a subsidiary of GIGABYTE and an industry leader in high-performance servers and IT infrastructure, is thrilled to present a cutting-edge series of servers optimized for the newly launched Intel Xeon E-2400 processors. These servers deliver essential computing power, ensuring a dependable workload for a wide range of enterprise and edge computing applications, all while maintaining an impressive price-to-performance ratio.

"We are thrilled to unveil our latest server product line, which is engineered to deliver unparalleled performance and reliability," said Vincent Wang, sales VP at Giga Computing. "By leveraging the power of the new Intel Xeon E processors, our servers empower businesses to elevate their computational capabilities, enabling them to achieve greater efficiency and productivity. Whether it's for enterprise applications or edge computing tasks, GIGABYTE servers are the cornerstone of innovation in the digital landscape." ⁠

ADATA Industrial, the world's leading brand for industrial-grade embedded storage, officially released today its SATA 31D series of industrial-grade solid-state drives including 2.5-inch ISSS31D, M.2 2280 IM2S31D8, and M.2 2242 IM2S31D4, designed for retail terminals and embedded systems.

The SATA 31D series of industrial-grade SSDs utilize 112-layer 3D TLC flash memory developed by WDC, boasting a P/E Cycle of 3,000 which is comparable to MLC. The 31D series also offers a variety of ultra-thin and compact mainstream specifications such as SATA 2.5-inch, M.2 2280, M.2 2242, and various capacity options from 128 GB to 2 TB. 112-layer 3D NAND (BiCS5) 31D series SSDs all support thermal throttling technology, which reduces SSD transmission performance in stages to effectively mitigate the risk of data damage due to overheating. Furthermore, a LDPC ECC error correction mechanism and End-to-End Data Protection technology ensure reliable data transfer and improve data integrity. The 31D series is eminently suitable for POS systems, information kiosks, digital signage, and embedded equipment.

QNAP Systems, Inc., a leading computing, and storage solutions innovator, today unveiled the new high-capacity TS-hx77AXU-RP ZFS NAS series, including 12-bay TS-h1277AXU-RP and 16-bay TS-h1677AXU-RP rackmount models powered by AMD Ryzen 7000 Series processors based on the cutting-edge AMD Socket AM5 platform. By integrating robust hardware and diverse I/O including DDR5 RAM, M.2 PCIe Gen 5, PCIe Gen 4 slots, and redundant power supplies, the TS-hx77AXU-RP series unleashes enterprise-level performance and delivers ultra-high bandwidth, futureproof expandability, and trusted reliability for performance-demanding Tier 2 storage, virtualization, 4K video editing, and PB-level storage applications.

"The whole new TS-hx77AXU-RP series ZFS NAS provides enterprises with superb performance and large storage capacity to tackle business-critical workloads and storage-demanding applications. The AMD Ryzen 7000 Series multi-core processors further unlocks the key performance of DDR5 and M.2 PCIe Gen 5," said Alex Shih, Product Manager of QNAP, adding "Paired with the ZFS file system that's most suited for business applications, the TS-hx77AXU-RP series stands out as the top choice for storage solutions that require uncompromising data integrity."

GIGABYTE Technology: Giga Computing, a subsidiary of GIGABYTE and an industry leader in high-performance servers, and IT infrastructure, today announced the GIGABYTE G383-R80 for the AMD Instinct MI300A APU and two GIGABYTE G593 series servers for the AMD Instinct MI300X GPU and AMD EPYC 9004 Series processor. As a testament to the performance of AMD Instinct MI300 Series family of products, the El Capitan supercomputer at Lawrence Livermore National Laboratory uses the MI300A APU to power exascale computing. And these new GIGABYTE servers are the ideal platform to propel discoveries in HPC & AI at exascale.⁠

Marrying of a CPU & GPU: G383-R80
For incredible advancements in HPC there is the GIGABYTE G383-R80 that houses four LGA6096 sockets for MI300A APUs. This chip integrates a CPU that has twenty-four AMD Zen 4 cores with a powerful GPU built with AMD CDNA 3 GPU cores. And the chiplet design shares 128 GB of unified HBM3 memory for impressive performance for large AI models. The G383 server has lots of expansion slots for networking, storage, or other accelerators, with a total of twelve PCIe Gen 5 slots. And in the front of the chassis are eight 2.5" Gen 5 NVMe bays to handle heavy workloads such as real-time big data analytics and latency-sensitive workloads in finance and telecom. ⁠

ADLINK Technology Inc., a global leader in edge computing, a global supplier of industrial PCs and motherboards, and a Titanium member of the Intel Partner Alliance, announces the launch of the new IMB-M47 industrial ATX motherboard for 12th and 13th Gen Intel Core i9/i7/i5/i3 processors. The IMB-M47 ATX motherboard delivers high-performance computing power, with multiple I/O and expansion ports, including three simultaneous independent displays, USB 3.2 Gen 2x2 (20 Gb/s), up to 128 GB DDR5 4800 MHz, 3x 2.5 GbE, multi-M.2 Key M, TPM 2.0, and PCIe 5.0 high-performance add-on cards for complex Edge AIoT processing tasks in industrial automation, machine vision, factory automation, and logistics.

The IMB-M47 industrial ATX motherboard supports 12th and 13th Gen Intel Core processors that utilize a high-performance hybrid architecture with up to sixteen E-cores (Efficient-cores) for multi-threaded background task management and multitasking, and up to eight P-cores (Performance-cores) for demanding single-threaded work such as graphics and video rendering. The 13th Gen Intel Core processors speed up edge applications that benefit from faster memory performance with faster DDR5 memory.

Advantech has skillfully incorporated 13th Gen Intel Core processors across a wide array of products, setting new benchmarks for performance and capability. These enhanced solutions harness cutting-edge CPU architectures in their latest designs, delivering a noticeable performance boost over the previous processor generation. This upgrade significantly enhances application efficiency in smart factories, computer vision, intelligent retail, healthcare, and edge AI applications. At Advantech, we are unwavering in our commitment to providing long-term support, ensuring that customers can rely on our products for uninterrupted operation.

Advantech's new solutions offer Intel processors with up to 24 cores and 32 threads, up to 128 GB of DDR5 memory, and a PCIe 4.0/5.0 x16 slot for efficient multitasking. Experience optimized productivity with Advantech's Embedded Design-In Services and quality solutions featuring industrial-grade components. Advantech's embedded OS and value-added software options include Windows, Ubuntu, and Yocto BSP, along with the Edge AI Suite for AI performance evaluation and DeviceOn for device management.

Today, Google announced its next generation of Chromebook devices, called the Chromebook Plus, said to improve upon the legacy set by Chromebooks over a decade ago. Starting at an enticing price point of $399, this new breed of Chromebooks integrates powerful AI capabilities and a range of built-in Google apps. Notably, it features tools like the Google Photos Magic Eraser and web-based Adobe Photoshop, positioning itself as a dynamic tool for productivity and creative exploration. In collaboration with hardware manufacturers such as Acer, ASUS, HP, and Lenovo, Google is launching a lineup of eight Chromebook Plus devices on the launch date, with more possibly coming in the future.

Each model boasts improved hardware configurations over the regular Chromebook, including processors like the Intel Core i3 12th Gen or the AMD Ryzen 3 7000 series, a minimum of 8 GB RAM, and 128 GB storage. Users are also in for a visual treat with a 1080p IPS display, ensuring crisp visuals for entertainment and work. And for the modern remote workforce, video conferencing gets a substantial upgrade. Every Chromebook Plus comes equipped with a 1080p camera and utilizes AI enhancements to elevate video call clarity, with compatibility spanning various platforms, including Google Meet, Zoom, and Microsoft Teams. Set to be available from October 8, 2023, in the US and October 9 in Canada and Europe, the Chromebook Plus is positioning itself as the go-to device for many users. On the other hand, the AI features are slated for arrival in 2024, when companies ensure their software is compatible.Below you can see the upcoming models.

Apple today announced iPhone 15 and iPhone 15 Plus, featuring an industry-first color-infused back glass with a stunning, textured matte finish, and a new contoured edge on the aluminium enclosure. Both models feature the Dynamic Island, and an advanced camera system designed to help users take fantastic photos of everyday moments in their lives. A powerful 48MP Main camera enables super-high-resolution photos and a new 2x Telephoto option to give users a total of three optical zoom levels—like having a third camera. The iPhone 15 lineup also introduces the next generation of portraits, making it easier to capture portraits with great detail and low-light performance. Building on Apple's innovative satellite infrastructure, Roadside Assistance via satellite can connect users to AAA if they have car trouble while off the grid. With A16 Bionic for powerful, proven performance; a USB‑C connector; Precision Finding for Find My friends; and industry-leading durability features, iPhone 15 and iPhone 15 Plus represent a huge leap forward.

iPhone 15 and iPhone 15 Plus will be available in five stunning new colors: pink, yellow, green, blue, and black. Pre-orders begin Friday, September 15, with availability beginning Friday, September 22.

Apple today debuted iPhone 15 Pro and iPhone 15 Pro Max, designed with aerospace-grade titanium that's strong yet lightweight to deliver Apple's lightest Pro models ever. The new design also features contoured edges and a customizable Action button, allowing users to personalize their iPhone experience. Powerful camera upgrades enable the equivalent of seven pro lenses with incredible image quality, including a more advanced 48MP Main camera system that now supports the new super-high-resolution 24MP default, the next generation of portraits with Focus and Depth Control, improvements to Night mode and Smart HDR, and an all-new 5x Telephoto camera exclusively on iPhone 15 Pro Max. A17 Pro unlocks next-level gaming experiences and pro performance. The new USB‑C connector is supercharged with USB 3 speeds—up to 20x faster than USB 2—and together with new video formats, enables powerful pro workflows that were not possible before. And with the addition of Roadside Assistance via satellite, the iPhone 15 Pro lineup builds on Apple's innovative satellite infrastructure to connect users to help if they have car trouble while off the grid.

iPhone 15 Pro and iPhone 15 Pro Max will be available in four stunning new finishes, including black titanium, white titanium, blue titanium, and natural titanium. Pre-orders begin Friday, September 15, with availability beginning Friday, September 22.

Lexar, a leading global brand of flash memory solutions, is unveiling a variety of new products that are estimated to be launched in Q4 2023. Designed for photography professionals and PC enthusiasts to accelerate their workflow, the expanded portfolio includes a new microSD card, internal SSDs, DRAM, and portable SSDs.

The new Professional GOLD micro SDXC UHS-II card, available in 128 GB and 256 GB, leverages the high-speed performance and a Video Speed Class 60 (V60) rating, allowing users to quickly capture, play back, and transfer high-definition multimedia files, including 4K video, with a read transfer speed up to 280 MB/s. This card is also great for use with gaming devices, sports camcorders, tablets, or smartphones, and drones.

collaboration with diverse industries and support various applications
Samsung Electronics, a world leader in advanced memory technology, today announced that it has developed the industry's first and highest-capacity 32-gigabit (Gb) DDR5 DRAM using 12 nanometer (nm)-class process technology. This achievement comes after Samsung began mass production of its 12 nm-class 16Gb DDR5 DRAM in May 2023. It solidifies Samsung's leadership in next-generation DRAM technology and signals the next chapter of high-capacity memory.

"With our 12 nm-class 32Gb DRAM, we have secured a solution that will enable DRAM modules of up to 1-terabyte (TB), allowing us to be ideally positioned to serve the growing need for high-capacity DRAM in the era of AI (Artificial Intelligence) and big data," said SangJoon Hwang, Executive Vice President of DRAM Product & Technology at Samsung Electronics. "We will continue to develop DRAM solutions through differentiated process and design technologies to break the boundaries of memory technology."

Kingston Digital, Inc., the Flash memory affiliate of Kingston Technology Company, Inc., a world leader in memory products and technology solutions, today launched the Kingston IronKey Keypad 200C, a hardware-encrypted USB Type-C drive that ensures both security and convenience at your fingertips. Now, users can achieve seamless data protection and effortless compatibility without the use of adapters when using USB-C equipped devices.

KP200C offers FIPS 140-3 Level 3 (Pending) certified military-grade security in an OS-independent alphanumeric keypad for easy-to-use PIN access. It incorporates XTS-AES 256-bit hardware-based encryption with Brute Force password attack protection and BadUSB protection with digitally signed firmware ensuring an unparalleled level of safeguarding for your valuable data. With the Multi-PIN option (Admin/User), the keypad can be used to set up an easy-to-remember alphanumeric PIN for the Admin or User accounts. Admin can restore a User PIN to provide access to the drive if the User PIN is forgotten.

V-COLOR Technology Inc, a leading memory manufacturer is proud to present the revolutionary DDR5 OC R-DIMM Workstation Memory in configurations 16 GB, 32 GB, and 64 GB. Designed for use with the latest unlocked Intel Xeon W-2400X and W-3400X series processors on respective Intel W790 chipset-based motherboards supporting quad-channel and octo-channel memory

Extreme Overclocking: Experience unrivaled power with OC R-DIMM Workstation Memory. Engineered for extreme overclocking capabilities, this memory module dares to push the limits and achieve unparalleled speeds. With v-color technology, effortlessly unlock the full potential of workstations and conquer even the most demanding tasks. Intel XMP 3.0 Certified: Seamlessly compatible with Intel XMP 3.0, OC R-DIMM Workstation Memory ensures a hassle-free and optimized experience. Effortlessly enhance system's performance and achieve the perfect balance between speed and stability.

When Lawrence Livermore National Laboratory (LLNL) announced the creation of a two-ExaFLOP supercomputer named El Capitan, we heard that AMD would power it with its Instinct MI300 accelerator. Today, LNLL published a Tweet that states, "We've begun receiving & installing components for El Capitan, @NNSANews' first #exascale #supercomputer. While we're still a ways from deploying it for national security purposes in 2024, it's exciting to see years of work becoming reality." As published images show, HPE racks filled with AMD Instinct MI300 are showing up now at LNLL's facility, and the supercomputer is expected to go operational in 2024. This could mean that November 2023 TOP500 list update wouldn't feature El Capitan, as system enablement would be very hard to achieve in four months until then.

The El Capitan supercomputer is expected to run on AMD Instinct MI300A accelerator, which features 24 Zen4 cores, CDNA3 architecture, and 128 GB of HBM3 memory. All paired together in a four-accelerator configuration goes inside each node from HPE, also getting water cooling treatment. While we don't have many further details on the memory and storage of El Capitan, we know that the system will exceed two ExFLOPS at peak and will consume close to 40 MW of power.

ASRock Industrial is introducing new choices in industrial motherboards powered by 13th Gen Intel Core Processors (Raptor Lake-S) with up to 24 cores and 32 threads that boost computer-intensive edge performance. They come equipped with Intel W680, Q670, and H610 chipsets, and offer support for up to DDR5-5600 memory modules and PCIe Gen 5, allowing expanded possibilities and seamless integration within industry-specific applications.

By harnessing the power of the 13th Gen Intel Core Processors, they leap up to 1.04x/1.34x/1.25x faster in single-thread, multi-thread, and CPU image classification inference performance, respectively, compared to the preceding 12th Gen Intel Core processors. The new 13th Gen Intel CPU motherboards with DDR5 support, available in Mini-ITX, Micro-ATX, and ATX form factors, have been specifically designed to cater to the unique requirements of the Edge AIoT applications in commerce, automation, robot, entertainment, and security industries.

TrendForce reports that explosive growth in generative AI applications like chatbots has spurred significant expansion in AI server development in 2023. Major CSPs including Microsoft, Google, AWS, as well as Chinese enterprises like Baidu and ByteDance, have invested heavily in high-end AI servers to continuously train and optimize their AI models. This reliance on high-end AI servers necessitates the use of high-end AI chips, which in turn will not only drive up demand for HBM during 2023~2024, but is also expected to boost growth in advanced packaging capacity by 30~40% in 2024.

TrendForce highlights that to augment the computational efficiency of AI servers and enhance memory transmission bandwidth, leading AI chip makers such as Nvidia, AMD, and Intel have opted to incorporate HBM. Presently, Nvidia's A100 and H100 chips each boast up to 80 GB of HBM2e and HBM3. In its latest integrated CPU and GPU, the Grace Hopper Superchip, Nvidia expanded a single chip's HBM capacity by 20%, hitting a mark of 96 GB. AMD's MI300 also uses HBM3, with the MI300A capacity remaining at 128 GB like its predecessor, while the more advanced MI300X has ramped up to 192 GB, marking a 50% increase. Google is expected to broaden its partnership with Broadcom in late 2023 to produce the AISC AI accelerator chip TPU, which will also incorporate HBM memory, in order to extend AI infrastructure.