SK hynix is reportedly continuing to invest capital, human resources, and strategic planning into its NAND Flash fab in China. According to recent reports, SK hynix plans to increase production at its plant in Dalian, China, by 50% in 2027 to meet the rising demand for NAND Flash. This fab is owned and operated by its NAND subsidiary, Solidigm, which is finally resuming expansion in mainland China after a four-year project freeze. In 2021, when SK hynix acquired Intel's 3D NAND and SSD business, it formed Solidigm to focus on this area of production. However, export controls and a downturn in the NAND Flash market halted any expansion plans that Solidigm and SK hynix had in China, effectively stopping strategic planning from the headquarters.

Currently, Solidigm produces about 100,000 wafers per month of NAND Flash, so a 50% increase in 2027 would raise production to approximately 150,000 wafers per month. Solidigm will also begin producing its floating gate 3D QLC NAND Flash, which will feature over 200 active storage layers per chip. This advancement will enable multi-terabyte SSDs for enterprise use, which is currently the primary consumer of these solutions. With global NAND Flash production expected to finally balance with demand in 2027, this introduction will be a welcome addition, though it represents just a small fraction of the global NAND Flash demand.

If it wasn't enough that memory, storage, and CPU shortages are driving up the prices of PCs, Microsoft is now imposing a price increase on OEM PC makers for their Windows licenses. According to Taiwanese UDN, Microsoft has notified OEM PC makers that Windows licensing fees are going up by 7% to 10%. One of the PC vendors noted that Microsoft typically increases its Windows license fees every year, but only by a low, single-digit percentage. However, this time the increase goes up to 10%, making it one of the most significant increases OEMs have experienced. We don't know the exact dollar amount, as Microsoft works with PC OEMs individually to set prices.

PC OEMs likely receive better prices based on volume, with large makers such as Dell, HP, ASUS, Acer, and others getting much better deals than smaller OEMs with fewer shipments. Thus, the 7-10% increase depends on the PC OEM's volume and individual deals with Microsoft. However, no one is spared from these increases as the Redmond-based giant seeks to extract a few more percentage points from the world's most widely used PC operating system. The company is investing heavily in engineering efforts to remodel its Windows 11 operating system for a better user experience after years of updates that compromised it. Increasing the license fees for OEMs is one way to offset these efforts.

NVIDIA's GeForce RTX 50 series of "Blackwell" graphics cards has seen significant price increases over the past few months, with a particularly notable jump between July and August. On Newegg, one of the largest computer hardware retailers in the United States, there has been a noticeable change in the median listed price for each GeForce RTX 50 model. For instance, the entry-level GeForce RTX 5050 had a median price of $300 in June, which rose by 5% to $315 in August. Moving up the range, the GeForce RTX 5060 was listed at an average of $370 in July, but this increased by 27% to $470 in August, marking a $100 difference. However, this is just the beginning.

The 8 GB edition of the RTX 5060 Ti was priced at $470 and has now risen to $530, reflecting a 13% increase. The most significant change occurred with the 16 GB version of the RTX 5060 Ti, which jumped from $570 in July to $805 in August, a 39% increase. The RTX 5070 experienced a similar percentage rise of 36%, going from $660 to $900. The RTX 5070 Ti remained steady at $1100, while the RTX 5080 saw a modest 3% increase from $1462 to $1500. The flagship GeForce RTX 5090 recorded a 9% increase, moving from $4300 to $4700.

Samsung Foundry has been consistently upgrading its chip-making tools. However, despite the availability of the new generation of EUV lithography, the company has not yet committed to using High-NA EUV lithography tools from ASML. According to ZDNet Korea, this is expected to change with the development of Samsung's 1 nm node, scheduled for 2030. This will be one of the last introductions of High-NA EUV machinery across leading-edge semiconductor node designs. Intel is gradually scaling up its 14A node to risk production with High-NA EUV tools, while TSMC is expected to introduce it in 2029. Samsung being one of the last does not necessarily mean the company is in a rush, as it is currently performing well with the existing Low-NA EUV tools.

For reference, TSMC claims it can maintain a competitive advantage using the existing Low-NA EUV technology for now, scaling nodes down and introducing improvements with each generation. Low-NA tools are also half the price of the premier High-NA EUV systems. The most advanced High-NA EUV machine from ASML costs around $400 million, which is an incredibly expensive piece of machinery, even for a modern semiconductor fab. Hence, TSMC has refrained from using it for now as it can achieve similar advantages with the current Low-NA technology, and Samsung appears to be following this decision.

Going disc-less appears to be beneficial for AAA game titles, as a new report estimates that game publishers will save more than $10 per game copy sold through digital licensing instead of physical discs. According to the GameFile report, game publishers are paying Sony and Microsoft between 15% and 20% of the game's MSRP to manufacture physical game discs, which are then sold through retailers. For modern AAA games, this represents a significant cost-cutting effort as the industry pushes for digital adoption. For example, with a game like the upcoming Grand Theft Auto VI, priced at $79.99 MSRP, this cost could range from $12 to $16. This presents a substantial cost advantage for game publishers, who would likely cease physical disc offerings if Microsoft and Sony stopped producing them.

The gaming console industry is currently divided, with some advocating for a move away from physical disc media, while others insist on the right to choose their preferred purchase medium. Microsoft's recent internal roadmap leak indicates that these disc-to-digital initiatives are imminent. Disc to Digital is Microsoft's initiative to promote a digital-only future for Xbox consoles. This feature allows players to insert their game discs into the console to convert them into digital licenses. When the disc is scanned, Microsoft uses the unique key to create a digital license, preventing the disc from being activated on third-party consoles and stopping it from being rented, borrowed, or sold.

TSMC reported its revenue for July, indicating that the company continues to increase its output capacity, customer commitments, and business execution. TSMC's July revenue stood at NT$467.58 billion, or about $14.49 billion at the time of writing. This marks a 5.6% increase from June's revenue of $13.75 billion and a significant 44.7% year-over-year growth. This suggests that TSMC nearly doubled its July revenue in an environment where it was already selling all its available wafer capacity, indicating rising demand from customers, increasing commitments, and a sustained AI expansion boom.

Revenue for the period from January through July 2026 totaled NT$2,872.06 billion, or about $89.03 billion, representing a remarkable 37.0% increase compared to the same period in 2025. In June, TSMC successfully began high-volume manufacturing of the N2 process. Currently, the N2 node accounts for only 3% of TSMC's revenue, while the 3 nm node contributes 30%, and the 5 nm node leads with 33%. TSMC's fabs currently produce about 20,000 wafers per month on the latest 2 nm N2 node. However, the company is rapidly scaling its high-volume manufacturing, aiming to produce about 100,000 N2 wafers per month by the end of 2026. This scaling up is driving revenue growth. Even with projected increases in N3 and N5 consumption, increasing revenue fivefold would mean the N2 node would represent more than 10% of the company's overall revenue share. Since N2 wafers are naturally more expensive, revenue growth for TSMC has only just begun.

Intel announced today that it will officially offer a $15 billion underwritten public offering of common stock to fund its capital-intensive foundry expansion. This move means the company is looking to dilute existing share ownership by issuing more shares to raise additional capital through its public offering. Supporting this deal is an elite group of investment banks, including JPMorgan, Goldman Sachs, Morgan Stanley, and Citigroup, who are acting as joint book-running managers. Intel operates in one of the most capital-intensive industries, where maintaining leading-edge semiconductor manufacturing requires tens of billions of US dollars. Therefore, the company needs additional funds without taking on loans or requiring external investors. Underwriters have a 30-day option to purchase an additional $2.25 billion worth of shares at the public share price, pushing the grand total to $17.25 billion

Regarding the main question of why Intel is doing this, the company noted that "Customers continue to signal a strong and sustainable demand environment, driven by unprecedented investment in AI compute. Progress in emerging areas, including physical AI, purpose-built silicon, advanced packaging, and external wafers, represents significant growth opportunities for Intel." This indicates that Intel's external customers are eager to use Intel Foundry services, including silicon manufacturing and advanced packaging, in quantities that exceed Intel's current capacity.

Noctua produces a range of cooling solutions, from low-profile air coolers and massive dual-tower air coolers to AIO liquid coolers. When fitting a large tower air cooler inside a case, PC case manufacturers list the maximum cooler height their chassis can accommodate. Recent testing by Noctua has revealed that more than half of the tested cases deviate from their original specifications. The room for error varies from being 3.5 mm less than the specified CPU cooler height to providing up to an extra 10 mm of cooler height. After testing over 100 cases, Noctua found that around 56 cases deviate from the manufacturer-provided specifications, mostly offering more space than advertised.

This is good news for case buyers, as there can be up to 10 mm more room for a CPU cooler. However, some cases are advertised with more space than they actually provide. For example, the ASUS Prime AP201 case claims to have 170 mm of CPU cooler headroom, but real-world measurements show it to be 168.3 mm. This means that when using the Noctua NH-D15 G2 dual-tower cooler, there's no way to increase RAM clearance by moving the front fan, necessitating the use of RAM kits that are only 32 mm tall. Anything taller will interfere with the CPU cooler. Another example is the Lian Li B4-mATX, which measured at 164.5 mm instead of the advertised 168 mm, making the NH-D15 G2 incompatible with this case.

According to Windows Latest, the default weather application in Windows 11 can consume more than 1 GB of RAM just by displaying the weather forecast. When the Weather application is open, it uses nearly 1.2 GB of RAM, even without any scrolling, clicking, or other interactions. By default, Windows 11 uses the "MSN Weather" application, which not only uses a significant amount of RAM but also displays ads. For an operating system that costs real money, having ads in a Microsoft-owned weather app, for which you paid a licensing fee, is a disappointing experience.

Last week, Microsoft removed documentation from its website and materials that recommended 32 GB of RAM for new PC builds or purchases. The company now emphasizes that 8 GB devices are optimized to run Windows 11 smoothly. However, with an 8 GB RAM configuration, the weather application alone consumes about 15% of the available RAM. This is before any other services start running or daily tasks begin. Below screenshots are taken by the Windows Latest team.

According to the latest information shared by long-time Intel leaker Jaykihn, Intel's upcoming Xeon 7 "Diamond Rapids" will feature a flagship SKU with up to 256 P-Cores. This contradicts previous reports suggesting that Intel's flagship Xeon 7 SKU would be limited to 192 P-Cores and 192 threads, as this generation does not include Hyper-Threading. From Intel's VLSI Symposium 2026 presentation, we learned that "Diamond Rapids" includes four Compute tiles, also known as "core building blocks" (CBBs), each built on the Intel 18A-P foundry node. Each Compute tile contains a CPU complex with 48 "Panther Cove" P-Cores and an L3 cache localized to the tile. However, a 256-core design implies CBBs with 64 cores each and possibly a larger L3 cache.

Using Intel's latest 18A-P node, the "Diamond Rapids" CBBs are expected to achieve a 9% increase in performance at the same power level compared to the standard 18A node, or an 18% power reduction at the same performance level. Since Intel aims for maximum performance, we can expect these improvements to enhance performance further or to optimize power consumption with the additional cores. Earlier claims suggested a 500 W TDP design for 192 P-Core version, featuring 16 DDR5 memory channels and running on a new LGA-9324 socket on the "Oak Stream" platform. However, a 256-core design could go much higher, and cooling manufacturers have indicated 800+ W TDP designs for this socket.

According to the reliable Intel leaker Jaykihn, Intel's upcoming "Nova Lake-S," featuring 12 Xe3P cores, currently has a Power Limit 2 (PL2) setting of about 40 W. This builds on the previous rumor that this integrated graphics solution will arrive with two VCCGT phases for up to 65 W of power. This SKU includes four P-Cores, eight E-Cores, and four LPE-Cores, along with 12 Xe3P GPU cores. This 4+8+4 CPU core and 12 GPU core configuration is one of the few processors where integrated graphics require two VCCGT phases. In regular SKUs, VCCGT is a single-phase power rail for the integrated graphics inside the processor.

However, we now know that the PL2 setting will only use 40 W, as the 65 W threshold is set for extreme scenarios like PL3, and the graphics processor's PL4 is capped at that 65 W figure. The Xe3P features more microarchitectural changes compared to the Xe3 found in "Panther Lake." If readers recall, a 12 Xe3-core iGPU found in "Panther Lake" has 16 MB of L2 cache, while the Xe3P iGPU dedicated to "Nova Lake" will have a 20 MB L2 cache configuration, a 25% increase. More L2 cache means that some workloads will naturally benefit, but we still don't know preliminary performance. The Xe3P microarchitecture is expected to carry more changes, especially since it has a larger L2 cache.

AMD's mysterious RDNA 4m integrated graphics have reappeared after a few months of silence, as the GPU has officially been merged into the Mesa 26.3 open-source driver stack. Running under the GFX1171 GPU target, this RDNA 4m is neither a true RDNA 4 design nor a simple RDNA 3 GPU IP. It sits just above RDNA 3.5, with supposed capabilities to support INT8 data types and native FSR 4 technology. However, since AMD announced that older RDNA generations are now getting FSR 4 upscaling, the true purpose of RDNA 4m remains unclear.

As readers may recall, AMD plans to keep its RDNA 3.5 iGPU in use until 2029, meaning most consumers buying laptops in the coming years will experience the same integrated graphics capabilities as those found in today's Ryzen AI 300 and 400 series processors. This suggests that the RDNA 4m, despite its limitations, will be a welcome addition to integrated graphics. "Redstone" features such as ML-based upscaling and rendering technologies rely on the RDNA 4 IP to function effectively. With RDNA 4m support now merged into the Mesa open-source graphics driver, the RADV and RadeonSI drivers are being updated for RDNA 4m.

Lenovo climbed 43 places to #153 on Fortune's annual Global 500 List, making it the highest ranking in the company's history and one of the most significant advances in recent years. The recognition underscores Lenovo's position as one of the world's 500 largest companies by revenue and concludes the strongest year in its history, with record revenue of US$83 billion and adjusted net income growing 42% year-on-year to US$2 billion.

The results demonstrate Lenovo's continued progress in delivering its Hybrid AI vision, positioning itself at the forefront of the AI inferencing and democratization era. AI-related revenue doubled year-on-year and accounted for 33% of group revenue for the full financial year, highlighting AI as the leading growth engine across Lenovo's businesses.

NVIDIA has officially brought its RTX Neural Texture Compression (NTC) technology to Windows-on-Arm, gearing up for a general release on its RTX Spark PC platform. First demonstrated at GTC 2026 in May, NVIDIA showcased how NTC can significantly reduce GPU VRAM usage by up to seven times. In a technology demo at one of the GTC 2026 sessions, NVIDIA revealed that NTC can lower VRAM usage from 6.5 GB to just 970 MB in certain scenes. NTC technology is an AI-driven texture output that enables games to feature high-quality, complex materials without sacrificing performance. Games can take advantage of the substantial VRAM reduction that NTC offers while maintaining the same texture quality.

For the upcoming RTX Spark, NVIDIA is enhancing the software ecosystem so gamers and future buyers receive a mature software that works seamlessly out of the box, eliminating the need for workarounds. Even Japan's legendary SEGA has released AAA titles optimized for RTX Spark and is dedicating more resources to support the ecosystem. NVIDIA has already started shipping its first preview build of the CUDA toolkit for the RTX Spark platform, specifically as a native Windows-on-Arm software toolkit. Now, emerging technologies like NTC will run on RTX Spark, which includes configurations such as one with 6,144 CUDA cores and another with 5,120 CUDA cores. Both configurations are based on the "Blackwell" GPU microarchitecture and are paired with a 20-core Arm CPU and up to 128 GB of LPDDR5X memory.

TSMC reportedly has a large stockpile of Apple chips as the company waits for DRAM delivery from Apple's suppliers to integrate into a complete package. Using Integrated Fan-Out Package on Package (InFO-PoP), TSMC is developing a 3D wafer-level fan-out package. This package is designed to hold memory directly above the SoC die, resulting in a smaller PCB design without the LPDDR5X module occupying over 100 mm² of PCB area. With less than six weeks until the launch of the iPhone 18 series and the iPhone Ultra, Apple is rushing to place DRAM orders to secure new supplies. The Cupertino-based giant primarily relies on DRAM from Micron, whose LPDDR5X memory is used in a range of devices and is the main source for the next-generation iPhone. SK hynix and Samsung are also suppliers, but Micron remains the primary choice.

Apple has reportedly been working to get more manufacturers on board, even negotiating with CXMT for a discount based on volume, but CXMT reportedly rejected this idea. As TSMC waits for Apple's DRAM suppliers, especially Micron, the company can't assemble Apple's processors just yet. InFO-PoP requires DRAM dies to be integrated into the package, and TSMC can't ship any processors before packaging is complete. This process reportedly takes up to two weeks, which means that TSMC and Apple will be working with a tight deadline around the launch of the latest iPhone update.

Apple is one of the largest supply chain operators in the world, maintaining a vast network of suppliers and manufacturing capacity. However, even a company of its stature is struggling amid the current DRAM shortage. According to Tim Culpan's Culpium newsletter, Apple is scrambling to secure DRAM just weeks before its annual iPhone update cycle. With less than six weeks until the launch of the iPhone 18 series and the iPhone Ultra, assemblers are rushing to place DRAM orders to secure new supplies. The Cupertino-based giant primarily relies on DRAM from Micron, whose LPDDR5X memory is used in a range of devices and is the main source for the next-generation iPhone. SK hynix and Samsung are also suppliers, but Micron remains the primary choice.

Interestingly, another report has surfaced about Apple's efforts to secure more DRAM. The company has reportedly entered negotiations with Chinese memory giant CXMT for LPDDR5X memory needed for its current product lineup. Apple attempted to negotiate better pricing with CXMT, likely due to the volume it plans to purchase, but CXMT rejected the proposed lower price. This is one of the first instances where negotiation power lies with the suppliers, as CXMT is producing one of the world's most in-demand commodities and is offering it at market rates without discounts.

Electronic Arts is officially back to private ownership after nearly 36 years in the public markets. If readers recall, in September 2025, EA announced that the company would be going private again under a consortium managed by PIF, Silver Lake, and Affinity Partners. This private undertaking has been valued at $55 billion, primarily led by Saudi Arabia's PIF. Interestingly, this means that the company's decision-making process will be handled by the consortium instead of public-market investors. This will have a significant impact on individual game development, as the new company management will likely dictate how the funds are spent, alongside the chairman and CEO, Andrew Wilson. He has even promised to continue investing in the next generation of titles boldly, stating the following:

Andrew Wilson, Chairman & CEO of Electronic ArtsWe're entering this next chapter from a position of strength with partners who share our vision and ambition. Together, we'll invest boldly, accelerate innovation, and build the next generation of games and experiences for the hundreds of millions of players and fans who inspire us every day.

Huawei has just unveiled its latest MateBook Pro laptop, featuring a homegrown processor, operating system, and a fully integrated experience. Priced at 9,999 RMB for the base model with 24 GB of RAM and a 512 GB SSD, this equates to about $1,480. The top-end configuration offers 32 GB of RAM and a 1 TB SSD for 13,999 RMB, approximately $2,072 at the current exchange rate.

Inside, the previous Kirin X90 has been replaced by the new Kirin X90 Plus, which is kept cool by a foam-copper vapor chamber, allowing the chip to maintain a 40 W performance envelope. Huawei seems to have retained the same core counts and GPU specs with a 10-core, 20-thread design, but the improved thermal management suggests they are extracting more performance from this generation. No official specifications have been released yet, so this remains speculative.

Microsoft is retroactively removing documentation from its website and materials that mention 32 GB of RAM as a recommendation for new PC builds or purchases. Instead, the company is now emphasizing 8 GB devices that are optimized to run Windows 11 smoothly, even with that RAM configuration. Thanks to Windows Latest, which spotted the changes, Microsoft's Windows Learning Center has been updated to remove mentions of the 32 GB RAM recommendation. Previously, Microsoft published a document suggesting that 16 GB is a baseline for a modern PC, describing it as a good "starting point." However, the company also used to recommend that gamers try to surpass this baseline, suggesting 32 GB as a "no worries" upgrade for gaming PCs.

At that time, this sparked significant controversy as gamers were vocal about Microsoft recommending more RAM when purchasing additional memory could cost as much as an entire PC, depending on the configuration. Amid a severe DRAM shortage, Microsoft faced pushback from gamers and quietly removed the mention of the 32 GB "no worries" recommendation. This also raised concerns about whether Microsoft would optimize the Windows 11 operating system to perform better on lower-spec PCs with less than 16 GB of RAM, such as the 8 GB PCs that Microsoft sells itself, like the Surface Pro 12 and Surface Laptop 13, both of which come with 8 GB base configurations.