TechPowerUP NEWS

[NextPowerUP]

Biwin's new Mini SSD squeezes NVMe-class performance into a SIM-tray style card barely larger than a MicroSD, offering quoted sequential read speeds of up to 3,700 MB/s and write speeds of up to 3,400 MB/s over PCIe 4.0 x2 in 512 GB, 1 TB, and 2 TB options. The 15 × 17 × 1.4 mm module ejects with a pin, similar to a smartphone SIM, allowing for easy swapping without tools. Biwin markets this module as IP68 dust and water-resistant, with three-meter drop protection for rugged mobile use. The format promises a sweet spot between MicroSD convenience and M.2 speed, and early support from handhelds such as GPD's Win 5 and the OneXPlayer Super X means we could soon see whether controllers, firmware, and thermal limits let the tiny drives live up to their claims. Additionally, if the format lives up to the promise, we could see adoption from more handheld device makers.

More...

[NextPowerUP]

SK hynix has taken the lead from Samsung Electronics as the top DRAM maker worldwide signaling the first change in leadership in over three decades. The main reason for this market shift is high bandwidth memory (HBM) where SK hynix has an exclusive deal to supply NVIDIA while Samsung hasn't yet secured this partnership (it does however supply HBM3E to AMD and Broadcom). Samsung's slice of the global DRAM market dropped significantly from 41.5% at the end of 2024 to 32.7% in the first half of 2025. This 8.8% drop is Samsung's biggest market share loss since they started keeping track in 1999. The company's market position has gone back to levels not seen since 2014.

SK hynix grabbed a 36.3% market share in the first half of 2025 beating Samsung for the first time since 1992. SK hynix DRAM market share has been growing from 27.7% in 2022 to its current top spot. A key component for this growth was also SK hynix's U.S. subsidiary which reported sales of $17.79 billion in the first half of 2025, representing a 103 percent increase from the previous year. Industry analysts expect SK hynix to maintain its competitive advantage in the HBM market despite anticipated increased competition from other manufacturers.

More...

[NextPowerUP]

Honestly, software subscriptions are exhausting. Everything's trying to charge you monthly—from your music to your to-do list app. But here's a rare win: you can get Microsoft Office Professional 2019 and Windows 11 Pro for just $45.97 (reg. $428). No recurring fees, just a one-and-done payment. For real. Let's break it down:

• You're getting the full Microsoft Office Pro 2019 suite: Word, Excel, PowerPoint, Outlook, Publisher, Access, and OneNote. If you've been limping along with some web-based alternative or an expired trial, this is your glow-up.
• On top of that, you get a legit Windows 11 Pro license—unlocked features like Remote Desktop, BitLocker encryption, Snap Layouts, and even Copilot, Microsoft's GPT-4 powered assistant that can help you generate content, answer questions, and more.

This deal only lasts a little while longer, so grab your lifetime licenses for Microsoft Office Pro 2019 and Windows 11 Pro for just $45.97 while supplies are still available.
Read full story

More...

[NextPowerUP]

TSMC is reportedly preparing to transition its advanced packaging strategy from CoWoS to a panelized approach called CoPoS, which stands for Chip-on-Panel-on-Substrate. The company plans pilot lines as soon as 2026 and aims for mass production between late 2028 and the first half of 2029. The change involves swapping around 300 mm silicon interposers for large square and rectangular panels, with early formats measuring around 310 × 310 mm and later options extending to 515 × 510 mm and even 750 × 620 mm. That larger, square format is intended to reduce wasted edge area, allow bigger reticles and masks for AI accelerators, and make it easier to place more dies and HBM on a single package.

Technically, CoPoS combines CoWoS ideas with fan-out panel-level packaging, building RDL layers on glass or sapphire panels rather than on circular silicon. TSMC is lining up CoPoS development and production at its AP7 Chiayi campus, where phases 4 and 5 are expected to host panel work, and it plans related capacity in future Arizona facilities. Equipment and tooling orders are already being placed, and the supplier list includes global names like KLA and Applied Materials, as well as many Taiwanese manufacturers, such as Innolux. The panel route promises higher area utilization, lower cost per packaged accelerator, and fewer warpage issues as package sizes grow, while TSMC continues to support its existing CoWoS lineup during the transition.

More...

[NextPowerUP]

Intel Foundry has been releasing content on its products and services, and today, an interesting update came across. Intel showed a reference "Deer Creek Falls" SoC of what they call a "non-x86" design, manufactured on the 18A node. However, the design is actually Arm-based SoC. Consisting of seven cores, the reference SoC packs a single performance core, two efficiency cores, and four low-power efficiency cores, alongside the embedded PCIe and memory controller IPs and PHYs from ecosystem partners. While this may not be an impressive design, it demonstrates that a heterogeneous design of hybrid cores and IPs from ecosystem partners exists, and it's likely that we will see some designs emerging from this.

While Intel didn't explicitly mention that this SoC is Arm, later in the video, in the section on performance tuning, there are mentions of AArch64, which is Arm's designation for the 64-bit Arm ISA flavor. Alongside the design, Intel also demonstrated that the company offers a suite of performance optimization tools, even for Arm-based processors, allowing its customers to optimize and enhance software performance on their preferred platform, not just the x86 ISA. Hopefully, external foundry customers are slowly becoming aware of Intel Foundry's capabilities, and the scale can commence, especially now that Intel is under supervision from the U.S. government.

14:10 UTC: Update: Intel has delisted the video from its YouTube channel, but our screenshots remain.
Read full story

More...

[NextPowerUP]

Intel has introduced a new driver feature called Shared GPU Memory Override, which allows integrated Arc GPUs on select Core Ultra systems to reserve a significantly larger portion of system RAM for video memory. The control appears in Intel Graphics Software as a slider that increases the percentage of system memory available to the iGPU. Early release notes and driver listings indicate default values near 57%, while vendor examples suggest that the control can allow significantly higher allocations on high-memory laptops, at 87% on Intel's demo. Intel positions the capability as a way to close the gap between integrated and discrete solutions for memory-bound graphics workloads and to give developers and power users greater flexibility for non-gaming tasks such as running local AI models.

Intel's implementation mirrors AMD's Variable Graphics Memory feature for Ryzen AI platforms, which similarly reallocates system RAM to the GPU and has been demonstrated alongside driver-level enhancements, such as AFMF, to increase frame rates in specific titles. Both approaches exploit unified memory architectures to enlarge the effective VRAM pool; however, this is not a universal performance win. Some games will opportunistically load larger assets when more VRAM is available, which can blunt or even reverse the gains. Users must install Intel's latest driver and reboot to enable the override. Minimum RAM and platform requirements also apply. OEMs may expose similar options in BIOS, and manufacturers will publish guidance. However, for workloads that are always VRAM-limited, such as local LLMs, it is essential to have an option to extend the iGPU's memory capacity, allowing larger models to run without an internet connection.

More...

[NextPowerUP]

A Reddit user bought a Radeon AI PRO R9700 graphics card in the retail channel for $1,220. This card was believed to be unavailable in retail, and restricted to the OEM and SI channels, meant to be sold by workstation builders. The card in question is a GIGABYTE Radeon AI PRO R9700 AI TOP, a custom-design board by the company that features a rack-friendly lateral airflow cooling solution that's build in GIGABYTE's distinct silver color scheme the company uses with its AI TOP branded graphics cards, notebooks, mini PCs, and motherboards.

The Radeon AI PRO R9700 is intended by AMD to be an edge AI accelerator, combining the AI acceleration capabilities of the RDNA 4 architecture, with 32 GB of memory, for accelerating AI models with larger parameter counts. It's based on the same 4 nm "Navi 48" silicon as the Radeon RX 9070 XT gaming GPU, which it maxes out, enabling all 64 compute units present on the silicon. The GPU offers a peak INT4 throughput of 1,531 TOPS, and peak FP16 throughput of 95.7 TFLOP/s. GIGABYTE includes access to its in-house AI TOP utility with this card.

More...

[NextPowerUP]

The Radeon RX 9060 (non-XT) is only released in the OEM-SI channel, for now, which is why AMD's add-in board (AIB) partners are rather muted with their announcements of custom-design RX 9060 graphics cards. A Sapphire-branded card was recently tested where it was shown presenting a good price-performance proposition against the GeForce RTX 5050 and RTX 5060, and now we see ASRock put up the product page of its first custom RX 9060 graphics card on its website. The new ASRock Radeon RX 9060 Challenger OC is based on the same board design as the RX 9060 XT 8 GB Challenger OC. The product page lacks pictures or renders of the retail box of this card, indicating that much like the Sapphire Pulse, this card will be available only to the OEM/SI market until AMD decides to bring the RX 9060 to the DIY retail channel.

The card features an aluminium fin-stack heatsink that's ventilated by a pair of 90 mm fans. Much of the airflow from the second fan flows through the heatsink, and out the backplate through a large cutout. Although branded "Challenger OC," the card does not appear to come with any factory overclock—the RX 9060 ticks at 2400 MHz Game clock, with 2990 MHz boost, while the memory is left untouched at 18 Gbps (GDDR6-effective). The RX 9060 is carved out of the 4 nm "Navi 44" silicon by enabling 28 out of 32 compute units present, working out to 1,792 stream processors, 56 AI accelerators, 28 RT accelerators, 112 TMUs, and 64 ROPs. The card comes with 8 GB of 18 Gbps GDDR6 memory across a 128-bit wide memory bus. The card draws power from a single 8-pin PCIe power connector; display outputs include two DisplayPort 2.1a, and one HDMI 2.1b. Although not yet available in the retail channel, it's been reported that AMD is targeting a $275 price point with the RX 9060.

More...

[NextPowerUP]

Relic Entertainment is thrilled to launch Warhammer 40,000: Dawn of War - Definitive Edition. The game is now available on Steam and GOG for $29.99 USD, with a 30% Veteran's Discount available to Warhammer 40,000: Dawn of War - Anniversary Edition owners (original game). Additionally, for 2 weeks a 10% launch discount will also be available to everyone.

This is Dawn of War as players remember it, faithfully enhanced for today's hardware. The award-winning, genre-defining real-time strategy returns with visual and performance upgrades that preserve the gridmark soul of the original while enhancing the experience on modern systems. The Definitive Edition includes the original Dawn of War and all its expansions; Winter Assault, Dark Crusade, and Soulstorm.
Read full story

More...

[NextPowerUP]

ASRock, a global leader in motherboards, graphics cards, mini PCs, power supplies, and gaming monitors, announces the latest Thin Mini-ITX motherboard, H810TM-ITX, designed for commercial environments that require exceptional stability and performance. The H810TM-ITX supports Intel Core Ultra Series 2 processors with AI acceleration, dual DDR5 memory configurations, four storage options, and up to five display outputs (supporting three independent displays). It further offers highly customizable I/O configurations, including optional HDMI + DisplayPort, USB4 or Intel Thunderbolt 4, and TPM 2.0—making it an ideal choice for AIO solutions, mini PCs, and specialized deployments.

For more advanced requirements, the B860TM-ITX and Q870TM-ITX also feature five display outputs - driving up to four independent displays, come standard with Intel Thunderbolt 4 and TPM 2.0 security modules, and provide richer I/O options—meeting the diverse needs of smart retail, interactive kiosks, AI-powered education, and digital-first workplaces.
Read full story

More...

[NextPowerUP]

Industrial IoT solutions provider AAEON (Stock Code: 6579) today announced the released of the uCOM-M700 and uCOM-M510, two SMARC modules powered by the MediaTek Genio IoT platform. The uCOM-M700 and uCOM-M510, which are AAEON's first products equipped with MediaTek technology, offer SKUs powered by either the octa-core MediaTek Genio 700 or hexa-core MediaTek Genio 510 SoC, respectively. Both platforms leverage ARM big.LITTLE architecture for efficient processing, as well as an integrated NPU and VPU to go along with an on-chip Arm Mali -G57 MC2 GPU. With these SoCs, the uCOM-M700 and uCOM-M510 can draw from a foundation that prioritizes efficient, low-power operation alongside up to 4 TOPs of AI performance, making them well-suited to applications such as advanced human-machine interfaces and smart retail kiosks.

The modules' specifications reflect the needs of their targeted application sectors, particularly with respect to the choice of display interfaces that can be used in conjunction with the Arm Mali -G57 GPU. Each module is equipped with HDMI 2.0, DP 1.4, and eDP 1.2, with an additional MIPI DSI connector available upon request. This grants users a broad choice to suit application need, while the combinations of said connectors allow for up to two simultaneous displays, one 4K and one 2K.
Read full story

More...

[NextPowerUP]

The United States government, under the Trump administration, is reportedly considering taking a stake in Intel to help the company, and more importantly, its foundry division, stay afloat. At the beginning of the week, we reported on the idea that Intel is collaborating with the Trump administration, as Intel's CEO Lip-Bu Tan was summoned to a meeting in the White House. However, it seems that a deal is nearing materialization. The latest proposed idea is that the U.S. government will take a state-owned stake in Intel, utilizing various methods of equity investments, guaranteed government purchases, loans, private financing, and other means to maintain the feasibility of Intel's and its leading-edge manufacturing operations. According to Bloomberg's private sources, a potential deal will focus on establishing two new, leading-edge node manufacturing plants in Licking County, Ohio.

Another interesting tidbit is that it is not the first time the U.S. government, specifically under the direction of President Donald Trump, has gotten involved in making a deal with a third-party company. A lesser-known name, MP Materials, also got a deal with the Department of Defense to accelerate the build-out of the U.S. rare earth magnet supply chain and reduce foreign dependency. Under the Trump administration, domestic companies are securing new deals to ensure that strategically important companies remain operational. For the U.S. national interest and strategic importance, continuing Intel Foundry's research, development, and manufacturing of leading-edge nodes is crucial. When asked about the deal, an Intel spokesperson commented, "We look forward to continuing our work with the Trump administration to advance these shared priorities, but we are not going to comment on rumors or speculation." Intel stock is up 7% on the news, so the market is pricing in the deal with high confidence that it will entirely go through.

More...

[NextPowerUP]

Sandisk Corporation today reported fiscal fourth quarter financial results.

"Sandisk delivered strong results this quarter, with revenue and non-GAAP EPS exceeding our guidance. We continue to execute with discipline, balancing innovation and operational focus," said David Goeckeler, Sandisk CEO. "The ramp of BiCS8 brings new levels of performance, density and energy efficiency to our customers. With High Bandwidth Flash (HBF), we are creating a new paradigm for AI inference solutions. With demand improving and industry fundamentals strengthening, we are well-positioned to drive sustainable growth, expand margins, and generate strong cash flow."
Read full story

More...

[NextPowerUP]

Today at its inaugural Level Reforge event, HP Inc. unveiled the next evolution of its OMEN and HyperX gaming ecosystem - powered by AI, built for creators, and engineered for players who demand more from every session.

PC gamers continue to raise the bar for what they expect from their gear, from smarter performance to more expressive setups. Eighty-five percent of gamers say high FPS is extremely important to their experience, while 76% are more likely to purchase from brands that offer personalized options. Building on the momentum from CES and HP Amplify earlier this year, the company's latest launch delivers OMEN AI across more devices, including the powerful new OMEN MAX 45L desktop and HyperX's latest personalized audio lineup.
Read full story

More...

[NextPowerUP]

According to a new patent filed by Sony Interactive Entertainment Inc. at the World Intellectual Patent Organization, the console maker may be moving to a simpler model for game development by putting the necessary hardware for game development into a host PC or server. Patent W02025170845 details a design that effectively puts "compute resources substantially similar to compute resources of a retail game console" onto a PCIe card and allows communication and translation between the host system and the on-board APU, storage, and other console hardware via a specialized management engine.

The patent details that this idea was partially inspired by the difficulties developing games for proprietary console hardware, but it also references cloud gaming, suggesting these slot-in cards may also come in handy as cloud compute modules that can be used in servers to support game developers who don't have access to dev kit hardware. A slot-in solution like this could also theoretically allow developers more freedom when it comes to testing and diagnosis, since they would be able to use hardware and software testing tools connected to the host system, instead of having to rely on built-in diagnostics and hardware support. The patent also specifically mentions being able to capture video from the host system, which could also help diagnostics and game optimization.

More...