According to TrendForce statistics, the global DRAM module market generated $21.2 billion in revenue in 2025, up a strong 59% YoY. Rising prices and concentrated shipments from DRAM module makers in the second half of the year helped drive growth. This marked a significant acceleration from the 7% growth recorded in 2024.

DRAM market conditions changed dramatically between the first and second halves of 2025. In early 2025, some module makers increased their inventories of DDR4 chips in Q2, spurred by tariffs and the end-of-life (EOL) plans of DRAM suppliers for DDR4. They also supplied PC OEMs and other consumer electronics customers that had pulled forward procurement in response to tariffs.

Kopite7Kimi, a leaker with a strong track record on NVIDIA hardware, says the company's next-generation GeForce RTX 60-series based on the Rubin architecture is a 2028 product. "Unfortunately, GR20x is a product from 2028. That means they have postponed again," the leaker posted on X. Their earlier estimate was the second half of 2027, so this is another "delay," if you can even call it that. NVIDIA hasn't said anything official about RTX 60-series timing, so treat it as a rumor for now.

Interestingly, this new leak seems to align with what prominent leaker Kepler_L2 claimed a few days ago: that Rubin gaming GPUs would only arrive in 2028. It also goes against leaker Moore's Law Is Dead, who said about a week ago that an NVIDIA source told them gaming Rubin was greenlit for the first half of 2027. Two leakers against one doesn't settle anything, and we don't know who is right yet, but the 2028 timing does seem more likely given the current industry conditions. Of course, these plans can change before 2028, as the memory supply situation is currently very volatile, and hardware roadmaps are changing almost on a weekly basis.

Acer Chairman and CEO Jason Chen says PC prices could rise by another 5% to 20% in the fourth quarter of 2026, according to Taiwan's Economic Daily News. Chen reportedly said that DDR4 and DDR5 memory are no longer in short supply and that CPU supply is mostly fine, with only some low-end and mid-range chips still tight. Prices haven't followed this change, though. Memory and SSD costs are still high, so systems with more RAM or storage will probably see bigger increases. Chen expects the average selling price of PCs to peak around mid-2027, with a possible drop in the second half of that year.

Even if component costs fall earlier, it takes several months for cheaper parts to reach finished PCs. Chen also reportedly said suppliers keep pushing through price increases, some of which he considers hard to justify. He also pointed to expanding Chinese memory production, and by the sound of it, he doesn't think the shortage can run until 2030, as many in the industry have claimed. That capacity is still ramping up, though. Chinese giant CXMT is reportedly preparing to enter the NAND flash market with an R&D line, and YMTC is expanding with new fabs.

As next-generation imaging devices continue to evolve, Team Group Inc., a global leader in memory and storage solutions, today introduces the Team Group ACE 8K MicroSDXC Memory Card. The ACE 8K is designed for a wide range of handheld imaging devices, including Insta360 and DJI cameras, as well as drones. The memory card supports smooth 4K 120 FPS video recording, making it an excellent choice for 360° panoramas, sports footage, and handheld content creation. With a focus on high-speed data transfer and consistent recording performance, the ACE 8K allows creators to capture every brilliant moment while providing a reliable, high-performance storage solution for on-the-go content creation.

The ACE 8K supports A2 Application Performance Class and U3/V30 specifications, delivering read speeds of up to 160 MB/s and write speeds of up to 110 MB/s when used with a professional card reader. This product is available in storage capacities of up to 1 TB, providing ample storage for high-resolution video, extended continuous shooting, and large volumes of content across a variety of application scenarios. Whether capturing everyday moments, dynamic outdoor footage, or professional 360° imagery, the ACE 8K delivers the precision and capacity creators need to capture all the thrilling moments.

Micron Technology, Inc. (Nasdaq: MU), today announced a major milestone in memory innovation with the successful demonstration of the world's first 512 GB DDR5 module on multiple server platforms. The achievement reinforces Micron's leadership in high-performance, high-capacity server memory and provides enterprise data center and cloud customers with a path to support large and demanding AI, analytics, in-memory database and simulation-heavy workloads with greater efficiency. Leading semiconductor enablers, AMD and Intel, are both actively validating the module, which will deliver speeds up to 9,200 MT/s.

The module's high capacity is possible due to Micron's advanced vertical interconnect packaging innovations. This product vertically stacks DRAM dies interconnected by through-silicon vias (TSVs) to maximize density within individual DRAM packages while preserving the performance modern data center architectures demand. The module enables up to 12 TB of DDR5 DRAM in a single 24-slot dual-socket server.

Patriot Memory today announced the release of the Patriot Viper Steel 5 Infinite RGB DDR5 ROG Edition. Born from an official certified collaboration with ASUS Republic of Gamers (ROG), this high-performance kit combines extreme DDR5 speeds with a patent-pending 3D Infinite Mirror optical structure. Built with tested XMP 3.0 base profiles reaching up to 7,000 MT/s, the series features an exclusive ROG Certified BIOS integration that unlocks hidden performance profiles up to 7,600 MT/s on compatible ASUS ROG motherboards.

Designed to complement the ASUS ROG hardware ecosystem, the Viper Steel 5 Infinite ROG Edition is housed in a custom matte carbon-black aluminium heat spreader. The chassis is styled with an official ROG Certified red label, slashed dot-matrix typography, and a diamond-cut Viper logo. Atop the module, a solid-media optical light-bar projects a multidimensional 3D light tunnel within standard memory slot dimensions, syncing seamlessly across ASUS Aura Sync, SignalRGB, and major motherboard RGB software.

Global semiconductor revenue surpassed $425bn in 2Q26 reaching another all-time high as the market grew a record 31.4% quarter-over-quarter (QoQ), according to new research from Omdia. AI demand and strong memory prices continued to be the leading market dynamics, taking total semiconductor revenue for the first half of 2026 to $752 billion. The 31.4% QoQ increase surpassed the previous sequential growth record of 29.2% in 1Q26. Since Omdia began tracking the semiconductor market in 1Q02, only 10 out of 97 quarters have recorded sequential revenue growth above 10%. The last four quarters beginning in 3Q25, have all recorded double-digit sequential growth, with 3Q26 expected to follow the same trend.

Memory continues to be the primary driver of growth, while other semiconductor components are growing above seasonal patterns and contributing to the robust market.

NVIDIA today quietly added another card to its pro-viz product lineup with the RTX PRO 5500 "Blackwell" workstation GPU edition. Featuring 84 GB of GDDR7 memory, it is equipped with a GB202 die that includes 21,760 CUDA cores. This is the same CUDA core count as the consumer GeForce RTX 5090 SKU, but it is now certified for professional workloads and has a much higher memory capacity. The 84 GB of GDDR7 is connected to the GB202 with a 416-bit memory bus that delivers about 1,398 GB/s of memory bandwidth. This would put each GDDR7 memory chip at about 26 Gbps per pin, which is a standard GDDR7 memory speed, but not the fastest available on the market. Additionally, to achieve the massive 84 GB capacity, there need to be 28 memory modules. NVIDIA placed 14 GDDR7 modules on the front side with the GPU die and 14 GDDR7 modules on the back of the PCB, effectively sandwiching the PCB.

Compared to the rest of the RTX PRO lineup, the new RTX PRO 5500 "Blackwell" sits just a tick below the flagship RTX PRO 6000 SKU, which has a slightly higher CUDA core count of 24,064 and a 96 GB memory capacity. However, the RTX PRO 5500 is also much more powerful than the RTX PRO 5000 in the 5000 tier. The RTX PRO 5000 "Blackwell" has a much lower CUDA core count of 14,080 cores, with up to 72 GB of memory. Technically, the new SKU is much closer to the RTX PRO 6000 than the RTX PRO 5000 SKU, just with a slightly lower core count and memory capacity. The official specification sheet follows, courtesy of PNY.

G.SKILL International Enterprise Co., Ltd., the world's leading brand of performance overclock memory, is excited to announce the launch of the new Trident Z5 Royal NeoX series DDR5 memory, featuring "AMD EXPO Technology: Featuring Ultra Low Latency", with specifications ranging from DDR5-6000 CL36 to DDR5-6000 CL26. Building on G.SKILL's long-standing dedication to memory overclocking and high-performance memory, the Trident Z5 Royal NeoX series DDR5 is a flagship masterpiece that merges luxury craftsmanship with overclock performance hardware. Supporting the latest AMD EXPO Technology: Featuring Ultra Low Latency, also known as AMD EXPO-ULL, the Trident Z5 Royal NeoX series overclock profile leverages fine-tuned memory sub-timings for advanced performance with supported AMD CPU, motherboard, and BIOS version.

"AMD EXPO Technology: Featuring Ultra Low Latency" is the newly announced memory overclocking technology released by AMD. Compared to existing G.SKILL DDR5 memory kits that include AMD EXPO support, the new Trident Z5 Royal NeoX series come with AMD EXPO-ULL that includes multiple extra-tuned memory sub-timing settings in the overclocking profile, which help deliver better latency performance at the same memory speed. PC enthusiasts can experience the enhanced performance by enabling the AMD EXPO-ULL memory overclock profile on a system with compatible AMD CPU, motherboard, and supported BIOS version.

CXMT has achieved the remarkable feat of officially beginning high-volume manufacturing of LPDDR6 memory. This positions CXMT's LPDDR6 memory modules as some of the first in the industry, with South Korean memory makers on a similar timeline for volume production of this next-generation technology. Reportedly, Xiaomi is the first customer for CXMT's domestic LPDDR6 memory, set to feature in the upcoming Xiaomi 18 Fold smartphone. Since the Xiaomi 18 Fold is scheduled to launch in September, CXMT's LPDDR6 memory will need to be delivered before the end of the month, indicating that customer shipments are already underway. This news follows recent reports that CXMT is conducting risk production of LPDDR6, suggesting a very high yield in production.

CXMT's LPDDR6 memory will operate at speeds of up to 12.8 Gbps, with the base speed of LPDDR6 being 10.7 Gbps. LPDDR6 builds on where LPDDR5X left off, which could reach speeds of up to 10.7 Gbps when highly overclocked. However, the new standard aims for significantly higher speeds, with the first generation of LPDDR6 from CXMT's fabs reaching 12.8 Gbps. Additionally, CXMT is producing these chips with a 16 Gb capacity, equivalent to 2 GB, and the memory IC is housed in a package-on-package (POP) casing, featuring about 1,295 pins in a ball grid array.Update, 14:55 UTC, September 7th : CXMT's official website confirmed mass production, with Alibaba Cloud, ByteDance, Tencent, Lenovo, Transsion, Honor, Vivo, and Oppo as customers.

Microsoft is working on expanding its memory integrity protection feature across its global Windows 11 installations, with the rollout beginning in October. For those unaware, memory integrity protection enhances operating system security at the kernel level, safeguarding it from sophisticated attacks by malicious actors. This feature is straightforward, requiring little to no additional configuration from the user, meaning any complexity is hidden while the PC remains significantly more secure. Using Virtualization-based Security (VBS), the OS employs hardware virtualization to create Windows hypervisors that dedicate isolated virtual environments. The OS operates under the assumption that the kernel can be compromised.

Microsoft developed this robust protection mechanism using Virtualization-based Security (VBS), allowing for more security hotpatches to be installed on the fly without needing a hard restart. For installations running in critical environments, this means the OS will continue running its dedicated tasks while receiving constant security updates. For large organizations with managed Windows 11 devices, setup will be easy as all devices will automatically benefit from the memory integrity addition. All existing admin and user decisions and policies will remain fully in effect, while devices with memory integrity currently disabled will not have it automatically enabled by the October rollout.

Samsung presented its updated memory roadmap at SEMICON Taiwan 2026, with HBM5 and the upcoming zHBM architecture being the main focus points while setting ambitious goals and expectations. According to Samsung, HBM5 is targeting twice the performance and 20% higher performance per watt compared to HBM4E, alongside a 20% reduction in thermal resistance. As changes, the base die moves from the 4 nm process used in HBM4 and HBM4E to Samsung's in-house 2 nm node, and the stack options expand to 12, 16, and 20 layers. Mass production is expected around 2028. Recently, Samsung confirmed at the Hot Chips 2026 conference's Advance Program that the company is preparing to ship HBM4E memory with speeds of 16 Gbps per pin. Samsung is currently shipping HBM4 that works at 11.7 Gbps per pin.

zHBM uses a more radical approach, rather than placing memory beside the accelerator as conventional HBM does, zHBM stacks memory directly on top of the processor, shortening the data path and delivering higher bandwidth and efficiency. With zHBM, Samsung is targeting eight times the raw performance and three times the performance per watt of HBM4E, with thermal resistance reduced by 75-90%. At FMS 2026, Samsung unveiled the industry's first concept models of zHBM and is expected to launch this completely different architecture after 2029. Rethinking the HBM architecture appears to be the direction the industry is heading. NVIDIA recently announced its own NVHBM custom memory that moves the memory controller into the HBM base die rather than keeping it on the compute die.

NVIDIA has reportedly revived its "Rubin CPX" AI accelerator project, which had previously been put on hold. Today, well-known supply chain analyst Ming Chi Kuo reported that NVIDIA is reintroducing the "Rubin CPX," now featuring HBM4 memory instead of the previously planned GDDR7. Late last year, NVIDIA announced its plans to develop a specialized accelerator derived from the "Rubin" GPU family, designed for large-scale agentic AI workloads. After the announcement, the project was reportedly paused, but NVIDIA has now redesigned the memory architecture. Previously, the "Rubin CPX" was slated to have 128 GB of GDDR7 memory, but it will now include 168 GB of HBM4 memory. This GPU SKU is housed in a separate rack with configurations ranging from 64 to 256 standalone CPX GPUs per rack.

This GPU is dedicated to prefill workloads, powering racks that manage input context for LLMs and KV cache processing. Essentially, NVIDIA is separating prefill from decode tasks, with CPX GPUs handling prefill and regular "Rubin" GPUs managing decode. A rack tray with eight CPX GPUs can handle 1.34 TB of long-context prefill and the associated KV cache, all utilizing HBM4 memory for faster processing and increased bandwidth. Since the previous design did not require HBM4 integration, NVIDIA has redesigned the package, likely using TSMC's CoWoS-S or CoWoS-L for packaging.

ASUS has officially launched its ProArt P14 creator laptop, which features NVIDIA's new RTX Spark Arm SoC. The 14-inch laptop is currently listed on ASUS's site as the H7407BA, and it is available with two touch-enabled G-Sync display options: a 14-inch, 2880 × 1800 OLED display with a 120 Hz refresh rate and up to 700 nits of brightness in SDR mode and 1,600 nits in HDR mode; or a 120 Hz, 2880 × 1800 OLED display with identical color reproduction but with 500 nits peak SDR brightness and 1,100 nits peak HDR brightness. Both displays also feature 100% coverage of the DCI-P3 color gamut and support for a pressure-sensitive stylus, all of which should make for a pretty convincing laptop for photographers and video editors alike.

ASUS says the ProArt P14 measures in at just 13.9 mm thick and 1.48 kg, while packing 24, 32, 48, 64, or 128 GB of LPDDR5X unified system memory and 512 GB or 1 TB of NVMe PCIe 4.0 storage—it does not seem as though there is a 16 GB RAM variant available. The P14's I/O relies largely on USB4 ports, featuring 3x USB4 Gen 3 Type-C ports with USB-PD and DisplayPort Alt Mode support alongside a single USB 3.2 Gen 2 Type-A port, a 3.5 mm audio combo jack, an HDMI 2.1 FRL port, and an SD 4.0 card reader. Pricing has not yet been revealed.

China's largest DRAM manufacturer, ChangXin Memory Technologies (CXMT), has sued the Pentagon over its designation as a company linked to the Chinese military. The lawsuit, filed Friday in the U.S. District Court for the District of Columbia, names Defense Secretary Pete Hegseth among the defendants and argues the classification was arbitrary, lacked evidentiary support, and violated CXMT's due-process rights. CXMT was first added to the Pentagon's Section 1260H list in January 2025 under Biden, and the Trump administration kept the designation in a June update.

CXMT's core argument centers on JEDEC compliance: its DRAM chips are built to standard commercial specs rather than military-grade requirements. These parameters include temperature tolerance, packaging, and testing. In February, the Pentagon reportedly issued a notice that CXMT would be removed from the list, then withdrew it the same day without explanation. CXMT says it spent over a year supplying the Department of Defense with information to support its removal, all without success.

Chinese memory giant CXMT has reportedly begun risk production of HBM3E memory, according to The Information. This indicates that Chinese memory manufacturers are currently two generations behind South Korean companies, which are gradually transitioning to HBM4E memory production. However, this is a significant milestone for CXMT, which is on track to achieve a manufacturing capacity of around 350,000 DRAM wafers per month by the end of the year, with plans to increase that capacity in the coming years. It is reported that HBM3E manufacturing has started in very low quantities, signifying the commencement of risk production.

While we don't know the specifications of CXMT's HBM3E memory, we know that the standard JEDEC specification states that this must be a 1,024-bit width, and speeds of anywhere from 9.2 Gbps per pin to 12.4 Gbps, with most of the standard configurations targeting 9.6 Gbps per pin. The total bandwidth is around 1.2 TB/s, and capacities come in anywhere from 24 GB using 8-hi stacks or 36 GB using 12-high stacks of 24 Gb layers. However, given that CXMT progresses rapidly from risk production to mass production, we could see high volume manufacturing take place in just a few weeks.

G.SKILL International Enterprise Co., Ltd., the world's leading brand of performance overclock memory, announces the launch of the new Flare X5X series DDR5 memory, with initial launch specification of DDR5-6000 CL36-36-36-76 32 GB (2x 16 GB). Continuing G.SKILL's long-standing commitment to memory overclocking technology and high-speed memory, the Flare X5X series DDR5 is designed for users seeking high performance and a clean minimalist hardware aesthetic. Supporting the latest AMD EXPO Technology: Featuring Ultra Low Latency, also known as AMD EXPO ULL, the Flare X5X series overclocking profile leverages fine-tuned memory sub-timings for performance with supported AMD CPU, motherboard, and BIOS.

"AMD EXPO Technology: Featuring Ultra Low Latency" is the newly announced memory overclocking technology released by AMD. Compared to existing G.SKILL DDR5 memory kits that includes AMD EXPO support, the new Flare X5X series come with AMD EXPO ULL that includes multiple extra-tuned memory sub-timing settings in the overclocking profile, which help deliver better latency performance at the same memory speed. PC enthusiasts can experience the enhanced performance by enabling the AMD EXPO ULL memory overclock profile on a system with compatible AMD CPU, motherboard, and supported BIOS version.

SK hynix expects the current memory shortage to persist through the end of 2030, with no clear signs of a slowdown yet, according to CEO Kwak Noh-Jung. Speaking after the company's groundbreaking ceremony for its new Indiana packaging facility, Kwak said any eventual slowdown would likely be the result of a gradual easing rather than the sharp price crashes memory markets have previously seen, per Reuters. SK hynix had previously flagged tight supply lasting through 2028 for commodity DRAM, so this latest comment pushes that prediction out further.

Kwak's reasoning is that memory products have moved away from being purely commodity parts. AI-driven demand has pushed customers toward customized DRAM and HBM configurations, which makes future demand easier for manufacturers to forecast and less prone to the abrupt oversupply swings that caused previous crashes.

SK hynix Inc. (or "the company", www.skhynix.com) is opening a new future for AI cooperation between the United States and South Korea by establishing a first next-generation High Bandwidth Memory (HBM) production hub in the US. SK hynix announced today that it held a groundbreaking ceremony for an advanced packaging production facility for AI memory on the 27th (local time) at the Holloway Gymnasium, Purdue University, located in West Lafayette, Indiana.

The groundbreaking ceremony was attended by prominent US government and political figures, including Indiana Governor Mike Braun, US Senator Todd Young, Purdue University President Mitch Daniels, and Chairman of the Korea-US Alliance Foundation Robert Abrams. Kang Kyung-wha, Ambassador of South Korea to the United States, was also present at the event. From the SK Group, Vice Chairman Yu Jeong-Joon, Vice Chairman Lee Hyung-hee, CEO of SK hynix Kwak Noh-Jung attended.

Kioxia Corporation, a world leader in memory solutions, announced today that it has commenced site preparation and other related work for the construction of Fab3, a new state-of-the-art manufacturing facility at the Kitakami Plant in Iwate Prefecture, Japan, to expand production capacity for advanced three-dimensional flash memory BiCS FLASH. Fab3 will be constructed on land south of Fab2, targeting operations to commence in fiscal 2029.

The flash memory market is expected to continue to expand significantly in the medium to long term, driven by agentic AI, physical AI, and on-device AI. Through the construction of Fab3, Kioxia aims to meet rising market demand by expanding production and ensuring stable supply of cutting-edge flash memory products.

NVIDIA has announced NVHBM, a custom high-bandwidth memory technology built into the HBM base die rather than on the compute die itself. This approach, which NVIDIA plans to use in its future GPUs, is now extended to third-party XPU customers through NVLink Fusion. The main idea behind it is rather simple, since standard HBM puts the memory controller on the compute die, that is consuming silicon that could otherwise be used for compute. NVHBM moves that controller into the 3D HBM stack, which has a notable effect on performance and efficiency. Compared to standard HBM4E, NVIDIA claims up to 30% more memory bandwidth, 15% lower HBM power consumption, and up to 25% more usable compute die area on the processing unit.

The area savings come primarily from a redesigned physical memory interface. Standard HBM uses wider interface connections that increase package footprint, on the other hand, the NVHBM custom PHY reduces that I/O area by up to 67% compared to JEDEC HBM4E. Also, the narrower interface simplifies interposer routing, freeing up to 80% more usable silicon across the layout. The reduced PHY footprint also means fewer routing constraints on the interposer, which matters when you're trying to pack multiple HBM stacks alongside a large compute die, explains NVIDIA in a dedicated blog post.

According to TrendForce's latest memory industry research, major global cloud service providers (CSPs) are accelerating AI infrastructure investment, with total CapEx projected to surge 98% YoY in 2026 and rise another 50% in 2027. The rapid increase will be driven in part by soaring memory contract prices and robust procurement demand. TrendForce estimates that DRAM and NAND Flash combined will account for 47% of CSPs' total CapEx in 2026, with their share rising further to 68% in 2027.

TrendForce observes that the sharp rise in memory contract prices since the second half of 2025 has significantly increased memory's share of CSP spending. Server DRAM—a key procurement category for CSPs—saw contract prices rise by a cumulative 64% in 2H25, with a further jump of approximately 270% expected in 2026. Meanwhile, a similar trend is unfolding in NAND Flash, with enterprise SSD prices rising by around 35% in 2H25 and projected to surge by a cumulative 235% in 2026.

XCENA, a semiconductor company developing memory-centric computing solutions for AI infrastructure, today presented the architecture and measured performance of MX1, its computational CXL memory architecture, during the Hot Chips 2026 Memory session. Presented by Harry Kim, Chief Product Officer of XCENA, the session examined how MX1 combines large-scale memory expansion, SSD-backed capacity and programmable near-memory computing in a single CXL Type 3 device to address the growing memory constraints of AI infrastructure. The MX1 presentation was selected through the Hot Chips program committee's technical review process.

As AI models continue to grow, system performance is increasingly constrained by memory capacity, data movement and the economics of high-bandwidth memory scaling rather than compute alone. MX1 is designed to complement CPUs, GPUs and AI accelerators by expanding memory capacity and offloading selected memory-bound operations—including vector search, KV-cache retrieval and data preprocessing—so host processors can focus on compute-intensive inference.

According to the latest report from Chosun, the United States government is reportedly in talks with the government of South Korea to secure major investment in domestic memory production on American soil, as pressure on the supply chain continues to grow. Since the majority of the world's DRAM production occurs in South Korea by companies such as SK hynix and Samsung, the U.S. government, during the Trump administration, is negotiating with South Korean officials to encourage more South Korean memory makers to invest in American DRAM fabs, which would significantly improve the supply chain situation.

There has reportedly been dissatisfaction with the level of Korean companies' investment in the U.S. and the delays these companies are causing despite promised expansions. However, the new strategic investment would be separate from the already-promised $350 billion expansion in the U.S. that was agreed upon by both parties last year. This would represent a new buildout, also worth billions of U.S. dollars. A new investment package is expected to be announced sometime in September, which is just days away, meaning we will soon see how South Korean memory makers plan to structure their U.S. investments.

Intel Foundry is achieving another significant success as SK hynix, a leading memory maker, is reportedly becoming a customer for EMIB advanced packaging. During Hot Chips 2026, SK hynix mentioned that it is collaborating with TSMC to integrate HBM memory into CoWoS-L, CoWoS-S, CoWoS-R, and most importantly, Intel EMIB. For instance, SK hynix's discussion focused on memory progressing from approximately 1,024 pins to 2,048 pins in the latest HBM4. The number of TSVs is increasing, and integration requires 2.5D packaging, hybrid bonding, as well as 3D DRAM integration to achieve next-generation memory technology with advanced AI accelerators.

As customers increasingly seek to diversify from TSMC's CoWoS, Intel's EMIB family offers many attractive options, particularly with EMIB-T. This variant includes TSVs to enable vertical power delivery across multiple stacked chips, supporting SK hynix's concept of vertical power delivery for 3D memory modules at the package level with EMIB-T.