ComputerBase conducted a massive online blind test, allowing thousands of gamers to vote on their preferred game video output. The options included NVIDIA's latest Deep Learning Super Sampling 4.5, AMD's FSR Upscaling AI with FSR 4 on Radeon RX 9000 series GPUs, and native resolution rendering without AI upscaling. Interestingly, DLSS 4.5 received the highest percentage of votes, winning in every scenario. ComputerBase tested various games like Anno 117, ARC Raiders, Cyberpunk 2077, Horizon Forbidden West, Satisfactory, and The Last of Us Part II. Only in Cyberpunk 2077 did gamers almost equally prefer native resolution rendering, but DLSS 4.5 still came out on top with a slight edge.

Initially, the ComputerBase team conducted the comparison using videos labeled only with options like 1/2/3/4, without revealing which rendering method was used, to ensure a fully blind test. After two weeks of testing and thousands of votes, the results showed a significant preference for NVIDIA's latest DLSS 4.5 technology. This conclusion highlights that NVIDIA's newest technology is now so refined that its video output offers superior visual appeal. DLSS 4.5 uses a second-generation Transformer model trained on a much larger dataset of video games, with five times the computational investment of the first DLSS 4 Transformer upscaler. It utilizes the FP8 data format, allowing for more processing with minimal precision loss. The new algorithm is more context-aware, and NVIDIA confirmed that the input for upscaling is used more intelligently, resulting in much more detailed and precise video output.

AMD has released its latest Ryzen Chipset Software version 8.01.20.513 driver, which includes general bug fixes and performance improvements. These updates should provide Ryzen CPU owners with some quality-of-life enhancements. In the release highlights, AMD mentions that the installer file has been updated with a new installation wrapper called Install Shield. Interestingly, AMD notes that this latest package does not include the AMS Mailbox and S0i3 filter drivers, which now need to be installed independently on systems that require them or may produce errors without them. Finally, AMD acknowledges that some issues persist, such as the regression from branch version 7 to the earlier chipset driver branch version 6, but the company offers a workaround, which is listed below. We have included the entire changelog below, as well as a download option on TechPowerUp's download section page.

DOWNLOAD: AMD Chipset Software 8.01.20.513

According to reports from Chinese local media, Apple is considering partnering with Chinese semiconductor manufacturers CXMT and YMTC for its upcoming iPhone 18 series and potentially other products like MacBooks and Mac computers. With memory and storage suppliers such as Kioxia, Samsung, and SK hynix raising prices due to a significant industry shortage, Apple is experiencing a squeeze on its profit margins while maintaining the same MSRP for its products. To diversify its supply chain, Apple is reportedly exploring the addition of DRAM from CXMT and NAND Flash from YMTC to reduce its dependence on South Korean and Japanese suppliers.

Recently, the U.S. Government's Bureau of Industry and Security reportedly removed CXMT and YMTC from the list of restricted Chinese companies. This move could benefit Apple, which may now utilize products from both firms. CXMT offers LPDDR5X and DDR5 memory, which aligns with Apple's needs for its products. Since Apple's iPhones and M-series SoCs primarily use LPDDR5X memory, CXMT's 12 Gb and 16 Gb LPDDR5X capacities could serve both the smartphone and computer sectors effectively.

ASUS has released an official statement in response to the recent decision by a Munich court to prohibit the sale of its PCs and notebooks in Germany. Although the company is now unable to sell these products in Germany, it will continue to provide after-sales services, including support and warranty claims. This decision stems from a ruling that ASUS infringed on Nokia's HEVC/H.265 patents. As a result, ASUS is prevented from generating additional revenue in Europe's largest economy. The court found that the two OEMs, ASUS and Acer, did not demonstrate behavior consistent with willing licensees under FRAND (Fair, Reasonable, and Non-Discriminatory) terms, leading the court to issue an injunction rather than limit the remedy to financial damages.

ASUSASUS has temporarily suspended its official website and online store in Germany following a Munich I Regional Court ruling (case no. Munich I - 7 O 4102/25) related to a patent dispute with Nokia. The suspension, resulting from a temporary injunction, affects select ASUS PCs and products using High Efficiency Video Coding (HEVC). All after-sales services in Germany remain fully operational, and existing customers will continue to receive uninterrupted support in full compliance with the current court order. ASUS is evaluating and pursuing further legal action to reach a fair resolution as soon as possible. ASUS continues to stand by its position and remains committed to our customers and partners.

According to SemiAnalysis, which recently released its InferenceX benchmark comparing the inference performance of the latest AI accelerators, AMD's next-generation Instinct MI455X accelerators, designed to compete with NVIDIA's "Vera Rubin," are facing serious manufacturing difficulties, causing delays in AMD's roadmap execution. Only low-volume production will occur this year, while customers will be generating AI tokens in Q2 2027. This means that NVIDIA's "Vera Rubin" VR200 rack-scale system will be the most powerful system shipping this summer, without competition. As AMD's first rack-scale system, the "Helios" AI rack with the MI455X UALoE72 configuration packs 72 GPUs and 6th Gen AMD EPYC "Venice" CPUs in a heterogeneous system designed to run as a single homogeneous GPU, delivering massive AI performance with multi-ExaFLOP compute output.

SemiAnalysisEngineering samples and low volume production of AMD's first rack scale MI455X UALoE72 system will be in H2 2026 while due to manufacturing delays, the mass production ramp and first production tokens will only be generated on an MI455X UALoE72 by Q2 2027.

Valve updated its Steam Deck website with a disclaimer, noting that the availability of the Steam Deck OLED version may be affected by the storage and memory shortage across the industry. While Valve has been able to secure a significant inventory of NAND Flash and DRAM modules for its Steam Deck handheld console, demand might be outpacing supply. The component shortage at Valve is not an isolated case, as many are expressing concerns over not being able to secure sufficient capacity. The only ones able to secure massive DRAM and NAND Flash capacities are hyperscalers building out data center infrastructure, who are securing long-term agreements with storage and memory makers.

Last week, we saw customers facing difficulty acquiring new Steam Deck OLED stock. Unfortunately, potential buyers in North America, Japan (through Komodo, Valve's regional distribution partner), Taiwan, South Korea, and Hong Kong were unable to get their hands on new OLED units. Initial reports suggested a "plentiful" supply for customers in the UK and Australia, but several key markets appear to be suddenly experiencing shortages, as the inventory is depleting across the globe. While other components are plentiful, memory and storage are in a dire shortage situation, which might be prolonged. Regarding the 256 GB LCD version, Valve has reiterated that it will only be available while existing stock lasts, as this model was discontinued in December of last year.

TSMC is reportedly planning to announce the addition of four new fabs at its Phoenix, Arizona site, known as TSMC Arizona. According to the Financial Times, TSMC will add four additional fab buildings at its Arizona campus, worth about $100 billion, with an official confirmation expected as early as April, just two months from the time of writing. In January, TSMC reportedly purchased additional 900-acres of land directly across the highway from the existing site north of Phoenix, preparing to expand its campus across. That extra production capacity will be supported by an estimated $100 billion of additional funding that TSMC is willing to spend on U.S. soil, despite making additional simultaneous investments back home in Taiwan.

However, many of the deal's specifics are still uncertain. According to Commerce Secretary Howard Lutnick, about $100 billion of those pledges are firm, which means that the total value of TSMC's investment in Arizona will amount to $265 billion. About $165 billion has already been invested in Arizona, and another $100 billion will make this one of the highest ever external investments on U.S. soil. Interestingly, TSMC is not alone in this investment round, as its supply chain partners are expected to contribute about $30 billion of the $100 billion figure, meaning that many more players are gathering funds to form the new supply chain independence.

Sony is reportedly postponing the launch of its next-generation PlayStation 6 console to 2028 or 2029 due to the massive rise in memory costs. This is reportedly impacting decision-making at the top of the company unit responsible for PlayStation strategic planning. Rising memory costs and a tight semiconductor supply chain are the core reasons behind any changes. Initially, we expected to see the wave of next-generation consoles in 2027, as AMD recently confirmed Microsoft's next-generation Xbox arriving in 2027. However, Sony could be abandoning the 2027 release window and pushing its PlayStation 6 console to a late 2028 or even 2029 launch, indicating that the company is not capable of securing sufficient DRAM allocation from its partners. This could leave Microsoft's new Xbox console as superior for an entire year or even two before Sony could join the next-generation console wars.

BloombergSony Group Corp. is now considering pushing back the debut of its next PlayStation console to 2028 or even 2029, according to people familiar with the company's thinking. That would be a major upset to a carefully orchestrated strategy to sustain user engagement between hardware generations.

According to the latest ruling from the Munich Regional Court, ASUS and Acer are banned from directly selling their PCs in Germany—Europe's largest economy. The judgment, handed down on January 22, 2026, has already led both manufacturers to suspend or remove affected product listings from their German online stores while they assess legal options and the scope of the ruling. The court's order targets the manufacturers' own sales and distribution channels, rather than retailers selling existing stock, effectively leaving the German market to supply only through other OEMs and the remaining inventory from ASUS and Acer.

The core reasoning behind this decision is patent claims asserted by Nokia, including patents related to core HEVC/H.265 encoding and decoding techniques. As both ASUS and Acer use GPUs, processors, and many other digital decoding/encoding engines for video, the use has to be licensed, and the sales can only proceed with approval from Nokia. The judges concluded that the two OEMs had not demonstrated the conduct of willing licensees under FRAND rules, which allowed the court to grant injunctive relief instead of limiting the remedy to damages. We don't have information on the length of the ban, which will likely require a followup with the court and two OEMs.

Intel has released its latest 32.0.101.8509 WHQL driver package, bringing support for XeSS 3 across multiple graphics generations and finally ending the launch exclusivity of "Panther Lake" integrated graphics. Now, Intel supports XeSS 3 multi-frame generation across Arc B-series "Battlemage" and A-series "Alchemist" discrete GPUs. For other integrated GPUs, only the Intel Arc GPU family is receiving official support, with "Meteor Lake," "Lunar Lake," and laptop "Arrow Lake-H" chips now offering full XeSS 3 support for 4x the frame output with AI-based frame generation. As Intel recently launched "Panther Lake," some game issues occurred, but thankfully they have been resolved. This includes color corruption in Ghost of Tsushima using DirectX 12 and game crashes in Mount & Blade II: Bannerlord using DirectX 11. However, some "Panther Lake" issues persist, with certain titles crashing and experiencing in-game corruption, which is expected to receive an official fix in upcoming driver releases.

DOWNLOAD: Intel Arc GPU Graphics Drivers 101.8509 WHQL

Intel has once again confirmed its product roadmap for the next generation of graphics IP, despite rumors that there might not be anything new after the launch of the Xe3P. According to a new presentation from Intel's Anil Nanduri, Vice President of GTM & Product Management AI Accelerators, responsible for data center AI accelerators, Intel will have a dedicated graphics portfolio after the recently introduced "Crescent Island." Labeled as "Xe Next," the next generation of GPU IP is still under development and currently without an official name. It is expected to follow the rollout of the Xe3P-based "Crescent Island" inference-focused GPUs. With "Xe Next," the data center lineup is seemingly expanding into two production lines: inference-focused "Crescent Island" GPUs and the long-awaited "Jaguar Shores" GPU for AI training and HPC workloads.

For AI training and HPC workloads, Intel's "Jaguar Shores" design is reportedly being finalized and will be completed by the middle of this year, with production starting shortly after. Assuming a perfect design, we could see Intel's "Jaguar Shores" GPU late this year or early 2027. It will likely involve a mix of Intel's internal production in collaboration with TSMC or Intel's pure 18A design, assuming yields are manageable. We have already seen that Intel confirmed HBM4 memory will be used in these accelerators, and SK hynix appears to be the main partner supplying Intel with this memory. As a rack-scale design, Intel will compete with NVIDIA and AMD in the highest-margin business.

Laptop and Mini-PC maker Framework is sharing another update with the community about the company's cost base for acquiring DRAM, such as DDR5 memory. In its latest February update, the company notes that the cost for DDR5 memory in its systems is now priced at $12-16 per GB of capacity, depending on the kit size and total capacity. This means that for a 16 GB kit, customers are expected to pay anywhere between $192-256, and as much as about $400 for 32 GB of DDR5 memory in its Laptop 12/13/16 models. According to the company blog post, this represents an average price that Framework is charging depending on the kit, as there is different pricing for a single higher-capacity sticks or dual lower-capacity DIMMs going into the system.

In late December 2025, we reported that Framework's pricing was $10 per GB for 8 GB, 16 GB, and 32 GB modules, and a bit higher for dual rank non-binary 48 GB modules. However, just two months later, the situation is now much worse, as we are seeing Framework's suppliers increasing memory costs anywhere from 20-60%, depending on the configuration.

Console gamers might not necessarily become console owners, as Sony's PlayStation UK division is experimenting with leasing the PlayStation 5. The lease starts at £9.95 per month and varies depending on the console version. This plan involves a 36-month lease for a PlayStation 5 Digital Edition with 825 GB of storage. Opting for a shorter lease, such as 24 months or 12 months, increases the monthly cost to £10.49 and £14.59, respectively. Other versions of the PlayStation 5, like the PlayStation 5 Pro, as well as accessories such as the DualSense Edge Wireless Controller, PlayStation VR2, and PlayStation Portal Remote Player, can also be leased for an additional cost per month, based on your lease choice. There is also a "rolling" lease option where users pay £19.49 monthly and can cancel at any time, provided they return the console.

A quick calculation shows that the 36-month lease is the most expensive, totaling £358.20 by the end of the term. Afterward, users have several options: upgrade the console by returning the old one, continue with the monthly subscription, or exit the Flex plan by returning the console. For the 24-month and 12-month plans, users pay £251.76 and £175.08, respectively, by the end of the lease period. These plans also offer the same options: return the console, upgrade the hardware and return, or continue paying the monthly fee for a while longer. The only flexible plan is the monthly rolling option, which allows users to pay £19.49 per month without any long-term commitment, and they can cancel at any time. This plan doesn't require any upfront payment for the console.

It is official: AMD is now capturing 40% of the entire server CPU market revenue share, according to data from Mercury Research. In the final quarter of 2025, AMD EPYC server CPUs managed to capture 41.3% of the revenue in the server/data center market that hyperscalers are spending on. This is a 1.8% increase from Q3 2025 and an impressive 4.9% year-over-year growth in a multi-billion USD data center market. For unit share shipment, AMD now stands at 28.8%, meaning the company is actually selling more SKUs at a higher average selling price. Intel, on the other hand, now holds 71.2% of the unit share while capturing 58.7% of the revenue share, indicating that Intel Xeon processors are now selling for a lower ASP with more units needed to reach this revenue.

The situation in the desktop segment looks interesting as well, with AMD's revenue share in the desktop CPU market now at 42.6%, while the unit share is at 36.4%. Again, this means that AMD's Ryzen processors sell at a higher ASP, nearly capturing half of the desktop CPU revenue with a bit more than a third of the unit sales. This sector also grew 1.6% sequentially, while Ryzen CPUs won the hearts of 14.6% more gamers for a yearly revenue share increase.

NVIDIA released a hotfix version 11.0.6.386 for its app that addresses a serious issue where users were unable to access advanced Optimus MUX switch options. These options were reportedly grayed out during random MUX switch scenarios. As a reminder, NVIDIA Optimus technology is designed to switch between integrated and dedicated graphics within the system to conserve power. This feature is not used on desktop systems but on laptop PCs that have an iGPU from an Intel or AMD processor and a dedicated NVIDIA GeForce GPU. During light workloads, such as basic image display, the iGPU is used, which saves laptop battery due to its low power consumption. However, when demanding graphics tasks are running, NVIDIA uses a MUX switch to activate its dedicated GPU for heavier workloads. In the previous NVIDIA app version, users reported that some advanced Optimus MUX switch options were "grayed out" during random MUX switch scenarios. The issue has now been isolated, and users can either update the NVIDIA app from the software itself or download it from the link below, which is a direct link to NVIDIA's .exe file.

DOWNLOAD: NVIDIA App Hotfix 11.0.6.386

Intel's 10 years of integrated graphics have yielded massive performance improvements, according to recent testing by Phoronix. The latest testing shows that moving from Intel Gen 9 integrated graphics in the "Kaby Lake" CPUs introduced in 2016 to the modern Intel Arc B390 with Xe3 cores in "Panther Lake," results in a 12x performance boost and a 8x performance-per-watt efficiency increase. This is remarkable progress for Intel's iGPU team, delivering steady performance improvements year-over-year, with a significant boost in recent years. Phoronix tested iGPUs of top-end Core models, including: Core i7 8550U "Kaby Lake," Core i7 8565U "Whiskey Lake," Core i7 1065G7 "Ice Lake," Core i7 1185G7 "Tiger Lake," Core i7 1280P "Alder Lake," Core Ultra 7 155H "Meteor Lake," Core Ultra 7 258V "Lunar Lake," and finally the Core Ultra X7 358H "Panther Lake" processor.

The oldest among these is the "Kaby Lake" generation, which utilized Intel UHD Graphics 620 on Gen 9 architecture, while the newest is Intel's most powerful creation to date—Arc B390 based on Xe3 cores. Comparing the 14 nm FinFET Intel node to the TSMC N3E node reveals a massive gap not only in performance but also in efficiency. In the geometric mean of all test results, Intel has achieved an 11.97x performance improvement from the 14 nm Gen 9 iGPU era to the modern 3 nm Xe3 iGPU era. This performance increase is accompanied by a significant efficiency gain, resulting from new nodes and more work done per watt, which Phoronix calculated to be 8x. While the "Lunar Lake" platform is the smaller power consumer with an average power draw of 13.82 W and a maximum of 36.97 W, "Panther Lake" uses a slightly higher average of 26.86 W and a maximum draw of 55.59 W for nearly twice the result.

Microsoft's Windows 11 26H1 update, initially expected to offer only new silicon support without a host of new features, is now dedicated exclusively to new and upcoming Arm-based processors like the Snapdragon X2 Elite/Plus. This marks a significant shift in Windows 11's development path—a first divergence in Windows deployment in the recent timeline, and a first for Windows 11. As a result, regular PC users with x86-64 and older Arm-based platforms will continue to use the Windows 11 25H2 as a feature update. According to the latest Windows IT Pro Blog, "Windows 11, version 26H1 is not a feature update for version 25H2," indicating that a different version will serve as the next feature update for Windows 11, rather than the anticipated 26H1.

Microsoft is concentrating on providing full support for the new Windows-on-Arm platforms, some of which are already available. These platforms include Qualcomm's Snapdragon X2 Elite processors and NVIDIA's N1 and N1x SoCs. This new hardware requires optimization from Microsoft to ensure the best experience on Windows 11, so the company is dedicating most of its efforts for the 26H1 update to this task. Since new silicon requires fine-tuning, such as specific power profiles and hardware optimizations to extract maximum performance, Microsoft is separating Windows 11 into another branch to make sure that servicing and updates are easier, and also that shipping customized power/performance settings doesn't end up corrupting the build for non-Arm CPUs.

TSMC just reported its January 2026 revenue results, with a net revenue of NT$401.26 billion (about $12.763 billion), an increase of 19.8% from December 2025 and an increase of 36.8% from January 2025. While these results are impressive, TSMC will have to spend more to keep customers coming back, and its board has just approved a $44.962 billion package to expand and upgrade its semiconductor facilities in 2026. This is a record capital expenditure for TSMC, indicating that the AI boom and sustained demand from its mobile customers are enough to keep the capital expenditure increasing every year. Originally, the plans included spending about $17.141 billion in Q1 of 2025, $15.247 billion in Q2, $20.657 billion in Q3, and $14.981 billion in Q4. However, most of these funds will actually be spent in 2026. Less than the new $45 billion figure was spent in 2025, making the new 2026 CapEx target the largest one to date.

TSMC's plans for this massive figure include expanding production capacity with hundreds of thousands of wafers per month, distributed across mature, current, and next-generation advanced nodes. Interestingly, mature node capacity is as important as maintaining the current node production, as entire industries like the automotive industry rely on TSMC's production and advanced packaging to satisfy all market needs. The current plan is to allocate about 70-80% of the new $45 billion package towards advanced nodes, with about 10-20% going to advanced packaging and mask making. The remaining 10% will be used for specialty technology expansion, likely including silicon photonics and other technologies.

On your Windows PC, the Unified Extensible Firmware Interface (UEFI) firmware has a Secure Boot Certificate that mandates only verified software starts the boot-up sequence. Microsoft is preparing to refresh these certificates, and the company announced that millions of Windows PCs in circulation will receive new Secure Boot Certificates in an industry-wide gradual rollout to replace aging certificates that are expiring soon. According to the latest Windows Blog, the original Secure Boot Certificates introduced way back in 2011 are reaching the end of their planned lifecycle, with the expiration date set for late June 2026. This not only mandates updating but also requires a massive staged rollout from OEMs and Microsoft's partners to ensure that all Windows devices stay secure.

According to Microsoft, this is one of the largest industry collaborations that spans the Windows ecosystem, including servicing, firmware updates, and countless device configurations from OEMs and other hardware makers. Firmware makers are at the center with their UEFI BIOS patches, which will now have to replace their aging Secure Boot Certificates. The blog also states that OEMs have been provisioning updated certificates on their new devices, with some devices from 2024 and almost all PCs from 2025 updated to support the new certificate. Interestingly, older PCs and devices that were shipped prior to these years will also be taken care of, with major OEMs providing their own guidance on updating the certificate. If you don't see your OEM offering an update, be patient as the rollout is gradual.

South Korean memory makers, SK hynix and Samsung, are preparing to showcase their next-generation LPDDR6 memory solutions at the International Solid-State Circuits Conference (ISSCC) 2026 in San Francisco, which will take place from February 15-19. As the premier event for showcasing advancements in silicon design, South Korean makers will present their best new technologies. For SK hynix and Samsung, this includes an update to their low-power DDR memory, now in its 6th generation. The LPDDR6 modules from SK hynix will arrive in 16 Gb capacities and offer a transfer rate per pin of 14.4 Gbps, built on the 1c generation (1γ generation) semiconductor node, which is the company's 6th generation of 10 nm DRAM. SK hynix runs these new modules at JEDEC's highest LPDDR6 speeds, meaning that the company is close to maxing out the new technology, and overclocked LPDDR6X versions might be arriving soon.

Samsung, on the other hand, has improved its LPDDR6 since the original CES 2026 presentation. The company will now present its 16 Gb LPDDR6 modules running at 12.8 Gbps, which is a significant improvement over the 10.7 Gbps modules from a few weeks ago. Samsung reportedly manufactures this LPDDR6 memory on a 12 nm process, which is slightly larger than SK hynix's 10 nm, but these modules also deliver great benefits. The company claims a 21% improvement in energy efficiency over its predecessor LPDDR5X. Additionally, Samsung's LPDDR6 memory uses NRZ signaling for I/O with a 12DQ subchannel, while SK hynix modules likely follow suit.

Amkor is preparing to greatly expand its Arizona-based operations, and the company will boost its spending not by a few percent, but by several multiples. The company is planning to triple its capital expenditures next year, with a dramatic increase from roughly $900 million in 2025 to as much as $3 billion in 2026, betting on massive demand from Intel and TSMC packaging technologies. This includes working with Intel and TSMC to enable their most advanced technologies like EMIB and CoWoS, all of which come in various form factors. We previously reported that Intel partnered with its long-time OSAT partner Amkor to take additional EMIB capacity that Intel's customers are interested in, in Incheon, South Korea.

However, as Amkor expands its facilities in Arizona, Intel will also collaborate with Amkor to deliver advanced EMIB packaging types on United States soil. While TSMC has been a primary choice for many high-density assemblies, growing interest in Intel's EMIB and Foveros options has led partners like MediaTek, Google, Qualcomm, and Tesla to consider alternatives. Interestingly, Amkor will also offload some of the CoWoS packaging work from TSMC by creating CoWoS packages on U.S. soil, instead of sending these chips back to TSMC's Taiwan fabs to finish production. Both CoWoS and EMIB/Foveros offer a list of benefits, making them highly sought-after packaging technologies for companies seeking to extract maximum performance from their chips. Amkor plans to be at the center of that supply chain and help Intel and TSMC handle more customers.

Microsoft is reportedly considering merging some of its subscription services into a single offering. According to The Verge, and later confirmed by sources from Windows Central, Microsoft is exploring the possibility of combining the PC Game Pass and Xbox Game Pass Premium subscription tiers into one "super" tier. This potential consolidation would address the increasingly complicated subscription lineup, which often confuses gamers and affects their subscription choices. Offering support for more than one platform in a single subscription could potentially revitalize the struggling subscription services sector at Xbox and align well with the timing of a new console release. The company is also looking into ways to incorporate more third-party service bundles into its Game Pass offerings.

Currently, PC Game Pass costs $16.49 per month after a significant price increase of nearly 40% last October, while the Xbox Game Pass Premium tier costs $15 but doesn't include the full library available to PC subscribers. At the top end is Xbox Game Pass Ultimate for $30 monthly, which offers day-one access to all Microsoft first-party releases, along with bundled perks from EA Play, Ubisoft+ Classics, and Fortnite Crew. Combining the PC and Premium tiers could simplify this structure, though it raises questions about pricing and feature access for current PC subscribers, or whether Microsoft will maintain the $16.49 price.

The next major refresh of AMD's Ryzen AI MAX APUs is still far away, but now we are putting together the pieces of the "Medusa Halo" APU puzzle. According to a famous leaker, @Olrak29_ on X, AMD's next-generation "Medusa Halo" APU will be complemented by LPDDR6 memory. This is one of the first LPDDR6 memory SoCs we are learning about, making it unique. Based on previous rumors, the silicon could have a 384-bit bus powering LPDDR6 memory, which would translate into massive bandwidth powering the SoC's new CPU and GPU configuration. This includes up to 24 "Zen 6" CPU cores and 48 RDNA 5/UDNA compute units for the GPU configuration. Paired with the added bandwidth from LPDDR6 memory—which these APUs greatly benefit from—"Medusa Halo" will be one of the best-performing SoCs when it launches.

Interestingly, memory manufacturers like Samsung and Innosilicon are already supplying LPDDR6 modules to customers for validation. Innosilicon's LPDDR6 modules boast an impressive speed of 14.4 Gbps, significantly faster than Samsung's initial modules, which achieve 10.7 Gbps. Innosilicon's modules offer a 1.5x increase in IO speed capability compared to the 9.6 Gbps of LPDDR5X previously available, along with improved efficiency. The latest LPDDR6 also increases the number of bits per byte of IO from 8 to 12. This results in LPDDR6's bandwidth at a single-channel 24-bit I/O speed being double that of LPDDR5X at a 16-bit single-channel. The company is reportedly collaborating with TSMC and Samsung to ensure sufficient production capacity for LPDDR6 IP, while Samsung relies on its own fabs for manufacturing memory.

Microsoft's next-generation Xbox console is reportedly taking an unconventional route by running on a customized version Windows 11 OS instead of the specialized console OS that typically powers these devices. According to Windows Central, the system will function essentially as a gaming PC that boots into an Xbox interface by default. This UI is likely similar to what the current Xbox Full Screen Experience looks and feels like, and will likely provide the same performance boost. We have already seen that Xbox FSE mode brings about a 9.3% reduction in RAM usage and about an 8.6% higher FPS due to the smaller system overhead. Users could exit that interface to access the full Windows 11 operating system, meaning the hardware would support Steam, EPIC, and other competing game stores, as well as standard PC applications alongside Xbox games.

This is Microsoft's first radical departure from the walled-garden approach that has defined console gaming for decades. What it could translate to is the first hybrid system that serves multiple purposes, from traditional gaming to running productivity suites of Microsoft 365 apps like Word, Excel, and others, all from the same system. Teams from the Windows and Xbox divisions are reportedly collaborating closely to adapt the operating system for living room use. Microsoft is also working with hardware partners like ASUS to create multiple devices at different price points rather than releasing a single standard console. Plans for a first-party handheld device are still under consideration, though the traditional console appears to be the main focus.

Intel has quietly deprecated its Software Defined Silicon initiative (SDSi), known as "Intel On Demand," according to a report from Phoronix. The company has archived the official GitHub repository for SDSi for Xeon, an effort intended to enable optional features on Intel's server processors that could be unlocked for an extra fee. Intel had hoped enterprises would pay to enable these features, but the initiative never gained mass traction and was only sporadically maintained. Because hyperscalers operate at massive scale, paying an additional fee to enable a feature on silicon they had already purchased made little sense, contributing to Intel's decision to abandon the project. Subscription services are similar in concept, but they generally apply to software on a monthly basis rather than one-time hardware activations.

Originally, Intel planned to make Quick Assist, Dynamic Load Balancer, and Data Streaming Accelerator available as On Demand features, alongside Software Guard Extensions and the In-Memory Analytics Accelerator. These were described on the Intel On Demand website as a "one-time activation of select CPU accelerators and security features." The Intel On Demand site has since been reworked to remove most information, leaving only a few documents and paragraphs. Thankfully, the idea of putting hardware features behind a paywall has not gained traction for now, leaving the paywall model to traditional software. At one point enthusiasts wondered whether Intel On Demand would trickle down to consumer CPUs, but with the project apparently dead, that possibility seems unlikely in the near term. Intel Upgrade Service existed in a similar format back in early 2010s, but was also short-lived.

AMD's FidelityFX Super Resolution 4 technology, now known simply as FSR 4, is currently supported in many games, but not across all AMD RDNA GPU generations. In response to an inquiry from Hardware Unboxed, AMD mentioned that it is still uncertain whether official FSR 4 support will be extended to the Radeon RX 7000 series and older GPUs, as the company reportedly has "no updates to share at this time." AMD official product separation stems from its RDNA 4 architecture and the support for 8-bit floating point instructions. While the latest RDNA 4 hardware supports Wave Matrix Multiply Accumulate in FP8 format, older RDNA generations like RDNA 3 and RDNA 2 lack this hardware instruction support and can't process 8-bit floating point data in this format.

However, older Radeon GPUs can instead rely on the 8-bit integer (INT8) data formats, which Radeon RX 7000 series fully supports. AMD accidentally leaked FSR 4 INT8 on its AMD GPUOpen platform, showing that FSR 4 on older GPUs is a possibility, which is just kept hidden for now. Later on, ComputerBase tested this leaked library, finding that FSR 4 offers a balance between native image quality and FSR 3.1 performance on both RDNA 3 and RDNA 2 hardware. In tests with Cyberpunk 2077 in 4K on Ultra settings using the AMD Radeon RX 7900 XTX, FSR 4 delivered 11% faster performance than native, but was 16% slower than FSR 3.1. Interestingly, performance may be the reason why AMD is holding these INT8 FPR 4 libraries back, but another point could be product separation.