AMD's Madhu Rangarajan, corporate VP of Compute and Enterprise AI, says the company never set out to build its Zen 6 EPYC 9006 "Venice" chips around agentic AI, a not-so-subtle dig at NVIDIA and Arm, which both pitch their latest server CPUs as made for that exact job. Rangarajan still called EPYC 9006 the best CPU for agentic workloads during a media session covered by Tom's Hardware. Chief Zen architect Mike Clark said much the same in a separate talk, admitting AMD didn't see agentic computing coming, though he added that the company always expected the CPU to keep a role in AI software. Here's what Rangarajan said:

We think that the EPYC 9006 CPU is the best CPU for the agentic era, and spoiler alert, we didn't magically know four years ago that agentic was going to be a thing and purpose-built something for agentic. The reality is an agentic workflow is a diverse set of tasks… and in order to execute those tasks well, you need compute that is able to handle a diverse set of workloads with a full portfolio.

NVIDIA may be slowing down its chip orders at TSMC, according to a new report from Taiwan's Commercial Times. The outlet cites unnamed supply chain sources but doesn't name a product, node, size, or timeline. Neither NVIDIA nor TSMC have commented so far, and the claims are single-sourced for now. The report frames this as NVIDIA being squeezed from inside and out. On the inside, there's Jensen Huang's November 8 email to staff, in which he apologized for an extramarital affair, and the board's independent investigation into reports that he was targeted by a foreign intelligence effort. The board has said it has no reason to believe national security or NVIDIA secrets were compromised.

The outside pressure mostly comes down to export controls. The report points to tighter export reviews and H200 chips reportedly stockpiled in Hong Kong, both of which it says raise the risk of NVIDIA scaling back orders. It also cites foreign media reports suggesting that some firms approved to handle H200 could be shell companies, which feeds concerns that chips might reach China through fronts like these or middlemen in other countries. The report presents that as a suspicion rather than a finding, and if cuts do happen, upstream suppliers such as TSMC and ASE are the most exposed.

While we've always co-designed our TPUs in partnership with Google DeepMind, we're now increasingly getting feedback from external organizations on their specific needs. With computing moving from human to agent scale, businesses are asking for accelerators that can reason across AI models' massive context windows. In addition, some highly regulated customers are asking for specialized encryption capabilities for unique workloads.

The new TPU systems we are announcing today address this new reality. Coupled with our recent PyTorch announcements, our aim is to make TPUs more easily accessible to every developer and Cloud customer. By combining modular custom silicon with an open and easy-to-use software stack, we can deliver new hardware and system configurations at the pace of the AI industry.

Infineon Technologies AG (FSE: IFX / OTCQX: IFNNY) has successfully processed 300-millimeter silicon carbide wafers into functional power devices at its fab in Villach, Austria. Infineon is the first company to pioneer this innovation milestone in a full-scale semiconductor fab environment. It was realized by utilizing existing silicon (Si) manufacturing equipment. This early achievement of functional power devices reflects Infineon's strong focus on providing customers with a comprehensive portfolio of high-performance SiC power semiconductors and ensuring long-term supply. Chip production on 300 mm wafers is technologically more advanced and significantly more efficient than 200 mm wafers, since the larger wafer diameter offers 2.3 times more chips per wafer.

According to Nikkei, a procurement team from Terafab, the chip venture between Tesla and SpaceX, visited Japan and walked local equipment makers through what it plans to buy. The targets it reportedly shared are gigantic: 1 million wafers per month from 2032 onward, with around 700,000 of those going to memory chips and 300,000 to logic. Nikkei's headline puts that at 40 times the output of Japan's Rapidus. The same split first surfaced in a Jiji Press report last month, which said a single Texas plant would handle both memory and logic, with a pilot fab completed as early as July 2027. Intel is reportedly helping with logic using its 2 nm and 1.4 nm-class technology, while a Taiwanese company has been floated as a technology partner for memory. That could be Nanya, though nothing is confirmed.

For context, Samsung's entire DRAM output was estimated at 650,000 to 700,000 wafers a month late last year, while SK hynix was at about 550,000 earlier this year and hopes to reach roughly 1 million by 2030 or 2031. It isn't clear whether Terafab's memory figure covers DRAM alone or also includes NAND, so the Samsung comparison is a loose one. Musk has also been careful with his own dates. On an October SpaceX call, he put "big error bars" on reaching volume production by Q3 2028, with a first SRAM test chip planned for mid-2027 at Terafab's Austin research lab.

ASRock Rack is promising up to 4 TB of DDR5 on a single socket with its DMRD16-2L+, one of the first motherboards built for Intel's upcoming Xeon 7 "Diamond Rapids" processors. The board was shown at the OCP Global Summit 2026 and has since been detailed by ServeTheHome. It has 16 DIMM slots in a one-DIMM-per-channel layout, matching the 16-channel memory setup Intel detailed at HOT CHIPS. Regular RDIMMs top out at 128 GB per module and 8000 MT/s, while MRDIMMs run at up to 12,800 MT/s, with taller 2U MRDIMMs reaching 256 GB each. Populate all 16 slots with those, and you hit 4 TB, along with a theoretical bandwidth of roughly 1.6 TB/s.

Expansion is unusual, with no standard PCIe slots and no OCP NIC 3.0 connector. All I/O goes through up to 16 MCIO x8 connectors, each rated for PCIe 6.0 or CXL 3.0, for 128 lanes in total. The rest is fairly basic: one PCIe 5.0 x4 M.2 slot, two Intel I350 Gigabit Ethernet ports, an ASPEED AST2700 BMC with a dedicated management port, and VGA output. ASRock Rack fits it all on a 12 x 13-inch EEB PCB, about the size of a large E-ATX board, so it could squeeze into some bigger workstation cases. Pricing and availability have not been shared yet, but that is hardly surprising, since Intel has the Xeon 7 family slated for a 2027 debut on the new LGA9324 socket. The board is aimed at servers, though anyone planning a high-memory workstation for local AI will probably keep an eye on it.

Maxsun has shared full specifications for the SK-PTLNAS and SK-WCLAIO, two NAS-focused MoDT (Mobile on Desktop) motherboards built around Intel's Panther Lake and Wildcat Lake chips. Both were first shown at Computex 2026 and are expected to launch this month, though pricing is still unknown. The SK-PTLNAS now tops out at the Core Ultra 9 386H, a 16-core chip with four Xe3 GPU cores, a 25 W base power and up to 80 W of turbo power. The Computex listing mentioned options up to the Core Ultra X9 388H, which no longer seems to be on the table. It supports up to 128 GB of DDR5-7200 memory across two SO-DIMM slots, with five M.2 slots and three Mini SAS ports for storage. The board also gets a PCIe 5.0 x8 slot, 10GbE, 2.5GbE and two 1GbE ports, plus two Thunderbolt 4 ports. Maxsun lists Windows 11 and Ubuntu support.

The SK-WCLAIO measures just 90 x 165 mm, a big change from the mATX design shown in June. It runs the six-core Core 7 350 (2P + 4LPE) with two Xe3 cores, a single DDR5-6400 SO-DIMM slot for up to 48 GB, 1 TB of onboard UFS 3.1 storage and four M.2 slots. Wildcat Lake has very few PCIe lanes, so those drives probably won't each get full x4 bandwidth. Maxsun rates the system at 45 W when active and 16 W with drives on standby, which should suit an always-on home NAS. The power input is a bit unclear for now, since the listing mentions both a 12 V DC jack and USB-C PD.

HWiNFO's newest beta, v8.55-6115, adds recognition for an AMD Radeon RX 7500, a graphics card AMD has not yet announced. The changelog entry is a single line with only the name, so the specs, the GPU underneath, and any launch plans are all unknown for now. The most likely candidate is probably the 6 nm "Navi 33" RDNA 3 chip, which already powers the 28 CU RX 7400 and the fully enabled 32 CU RX 7600. Of course, a card squeezed between those two would leave AMD very little room to work with, so the name is a bit puzzling. AMD hasn't said anything about the card so far, but we have heard the name before.

When the RX 7400 and RX 7300 first appeared in rumors back in 2024, there were rumblings of an RX 7500 XT desktop graphics card, though nothing concrete ever followed. Moving onto other additions to HWiNFO, the same build also adds enhanced support for Intel's Nova Lake-E and Razor Lake-H, HX, and S processors. Razor Lake is the platform expected to follow Nova Lake, and Intel has yet to officially announce it. CPUID added Razor Lake support to CPU-Z 3.02 earlier this week, while HWiNFO picked up preliminary Razor Lake-AX support back in August. With the first Nova Lake-S desktop chips not expected until Q1 2027, retail Razor Lake chips are probably still a long way out.

Police in Oakland, California, are investigating the theft of six pallets of NVIDIA hardware worth $11.4 million that were bound for Fort Worth, Texas, the Bay Area News Group reports. A truck carrying a food transportation company's logo picked up the cargo in Oakland on August 29 as planned, but the truck and its drivers then went off the grid heading south. According to authorities, the truck made several multi-hour stops before police lost track of it in San Bernardino County. Apparently, nobody realized the cargo was gone until early September, and Oakland police say the theft was reported to them on September 13. No arrests have been announced.

What exactly was on those pallets hasn't been disclosed, so we don't know if it was AI accelerators, complete systems, or GeForce graphics cards. Speaking to ABC7, UC Berkeley professor Olaf Groth said NVIDIA chips can sell for roughly double their US price abroad, with a $550,000 compute system going for $1 million or more. He added that the heists are getting bigger and more complex, which points to increasingly professional crews.

AMD CEO Lisa Su now calls AMD a "data center first" company, two days after AMD reported a 40% drop in gaming revenue. Speaking to Yahoo Finance's Brian Sozzi on November 5, Su said AMD's data center business should grow 50% sequentially from Q3 to Q4, and suggested it could more than double next year. Here's how she put it:

Dr. Lisa Su, AMD CEOWe are a data center first company, that is our profile, and you know, people need more computing to power all these applications.

Data Center brought in a record $8.5 billion in Q3, up 97%, and now makes up 64% of AMD's $13.4 billion total, up from 47% a year ago. Gaming fell to $773 million, or under 6% of revenue. AMD blames it mostly on lower semi-custom sales, meaning the console chips it makes for Sony and Microsoft, along with weaker graphics demand. Gaming was flat against Q2, though, and last year's Q3 was boosted by a 181% console-driven jump to $1.3 billion. Su used the same "data center first" line back in 2024, when data center was just over half of revenue.

The NVIDIA RTX PRO 5500 "Blackwell" workstation card has shown up at a US retailer, and it surprisingly costs more than the flagship RTX PRO 6000. As spotted by Alex Ziskind on X and corroborated by Videocardz, SHI Government Solutions listed the PNY-built model (VCNRTXPRO5500B-PB) with an MSRP of $17,999 and a discounted price of $17,408.63. NVIDIA's own US marketplace, on the other hand, lists the RTX PRO 6000 for $16,000, even though that card has more CUDA cores (24,064 vs. 21,760) and more memory (96 GB vs. 84 GB). So even at SHI's discounted price, the cut-down card costs around $1,400 more. NVIDIA hasn't announced official pricing since it quietly added the RTX PRO 5500 to its lineup in September, so it isn't clear whether the $17,999 figure comes from NVIDIA or PNY.

Videocardz points out that Chinese demand is probably a big factor behind the RTX PRO 5500's price. According to Reuters, ByteDance, Alibaba, and Tencent have ordered over 3 million RTX PRO 5500 cards at roughly $11,400 each, with total orders from China reportedly topping 6 million. One source said NVIDIA can make no more than 3 million of them in 2027. With the B300 barred from China and the H200 reviewed case by case, the RTX PRO 5500 is one of the strongest NVIDIA GPUs Chinese firms can still buy. The card also packs 28 GDDR7 modules at a time, even as memory prices keep climbing. With orders far ahead of supply, it's reasonable to expect RTX PRO 5500 pricing to stay high for a while.

CPUID has released CPU-Z 3.02, and the first line of its changelog simply reads "Intel Razor Lake support." Razor Lake is the processor family expected to follow Nova Lake, and Intel has yet to officially announce it. The changelog doesn't say which Razor Lake variants are covered or how deep that support goes, so this could be anything from basic detection to early groundwork. The November 4 update also adds the AMD Radeon AI PRO R9700, along with fixes for an SPD reading issue on Intel Broadwell-E and a validation problem on Windows XP.

Of course, system tools tend to pick up new chips well before they reach store shelves. HWiNFO added early Nova Lake support back in August 2025, and the first Nova Lake-S desktop chips still aren't expected until Q1 2027. HWiNFO also added preliminary support for Razor Lake-AX in August this year, the rumored large APU that could pack up to 32 Xe3 GPU cores alongside on-package memory.

Google has confirmed that Googlebooks with removable keyboards are on the way. During its Android Dev Summit 2026 keynote last week, the company said its Chromebook successor "spans across large screen form factors from clamshells, convertibles, and soon detachables." That's all Google said, so there's no date, price, or specific device attached to it yet. The current lineup is mostly made up of clamshells, and the Acer Googlebook 14, with its 360-degree hinge, is the only convertible in it.

Reports claim that the first detachable will most likely come from Lenovo, which already sells the Googlebook 15. Its Googlebook Duet m, codenamed "Sapphire," showed up in the Bluetooth SIG database back in August, and a matching keyboard and Lenovo Pen Pro stylus have now cleared the FCC under the IDs O57DUET13BTKB and O57AP501U1. The keyboard schematics carry Googlebook branding on the left palm rest, and it seems to work both through pogo pins when attached and over Bluetooth LE on its own battery. Leaks point to a 13-inch OLED panel and a MediaTek chip, though reports disagree on which one, and Lenovo hasn't confirmed any of it.

Samsung has reportedly booked the entire planned output of its second fab in Taylor, Texas, and most of it is going to AMD. According to an exclusive from Korea's Maeil Business Newspaper, citing industry sources, Taylor Fab 2 will make HBM base dies for AMD and NVIDIA on a 2 nm process, with AMD's orders making up the bulk of it. Apparently, AMD's volume alone is enough to stretch the fab's capacity. Samsung declined to comment, saying it can't confirm customer-related details.

These aren't AMD's CPUs or GPUs, though. The base die is the logic chip at the bottom of an HBM stack, and Samsung plans to pair the Texas-made dies with DRAM produced in South Korea and ship them as finished HBM. Its current HBM4 and the sampled HBM4E use 4 nm base dies, so 2 nm or smaller seems aimed at later HBM generations. The deal also follows Lisa Su's visit to Samsung in October, where Samsung reportedly wanted foundry orders from AMD in return for the HBM that AMD urgently needs. A March MOU had already made Samsung the primary HBM4 supplier for the Instinct MI455X.

Intel's upcoming Core Ultra 4000 "Nova Lake-S" desktop processors will reportedly cap their maximum core frequency at 5.1 GHz whenever they run above 90°C, according to Intel leaker Jaykihn on X. He says the limit applies to the entire Nova Lake-S desktop lineup, works separately from regular TjMax throttling at 100°C, and can't be raised. He also notes that the information is still preliminary. In practice, it looks like a safety net that sits between normal boost behavior and full thermal throttling. Below 90°C, the chips should presumably boost as usual, and models that don't exceed 5.1 GHz in the first place wouldn't be affected at all.

Final Nova Lake clocks aren't official yet, though the current Core Ultra 9 285K already boosts up to 5.7 GHz, so a high-end Nova Lake chip that runs hot could lose a sizable chunk of its peak frequency. A 100°C TjMax would also be lower than Arrow Lake's 105°C, which can be manually raised to 115°C on the 285K. That puts more weight on cooling, since the 28-core Core Ultra 9 4970K BFC was reportedly set to pull up to 296 W at PL2 despite its 125 W rating, and staying under 90°C at that kind of power draw will probably take a really capable cooler. Nova Lake-S is currently expected to arrive in Q1 2027, and Intel hasn't confirmed any of this yet.

X user GOK has shared photos of a huge red graphics card that carries four Polaris GPUs on a single PCB, jokingly asking whether it might be the biggest PCIe graphics card in the world. The four GPUs sit in a 2x2 layout, and each one appears to be surrounded by eight GDDR5 memory chips, which would mean 8 GB per GPU and a gigantic 32 GB across the board. According to GOK, a sticker on the card identifies it as a quad RX 470, which would work out to 8,192 stream processors in total. Feeding all of that are four 8-pin power connectors, while the only display output is a single HDMI port. A second photo shows the card lying next to a motherboard, and it's noticeably bigger than the board itself.

The PCB does carry an AMD logo next to a "DYII-XQ570 V1.0" marking, but it looks like AMD may not have built this one. GOK pointed out that the card has no board OPN (part number) and that the marking doesn't follow AMD's official format, adding that small factories in China sometimes put the AMD logo on their own designs. He also identified the PCIe switch as a PI7C9X2G, a PCIe 2.0 part, and thinks the card was likely made by one of those factories for crypto miners. He hasn't tested it yet, since he doesn't have a heatsink for it. Multi-GPU Polaris mining boards aren't new either, as Sapphire had a dual RX 570 mining card with 16 GB and a single HDMI port back in 2021.

Leaker Haze2K1 says on X that Intel might put the Arc B580 on the end-of-life (EOL) list. Apparently, Intel wants to prioritize workstation and AI inference cards like the Arc Pro B50, which are reportedly selling really well and carry higher margins. Intel hasn't confirmed anything, and Haze2K1 only says the card "might" go EOL, so this is only a rumor for now. However, the logic checks out. The Arc B580 launched in December 2024 at $249 and beat the RTX 4060 in TechPowerUp's review. The Arc Pro B50 and B60 use the same BMG-G21 silicon, and the B60 pairs the same 20 Xe cores with 24 GB of memory instead of 12 GB. Intel estimated the B60 at around $500 at launch, roughly double the B580's price. Clearly, Intel stands to gain more monetary benefit from the workstation cards.

As for the consumer GPUs, their supply already looks quite thin. The B580 is reportedly out of stock at most retailers, and prices have also climbed well above MSRP. The leaker doesn't mention the Arc B570, so we don't know if Intel has similar plans for that card as well. Intel's larger BMG-G31 chip also showed up in March in the workstation-only Arc Pro B70 and B65, so the shift toward pro cards seems to be underway already. If the B580 does go away, gamers will lose one of the more affordable 12 GB cards, even though the Arc B580 was selling above its MSRP. That said, I wouldn't expect cheaper Arc cards anytime soon

The Semiconductor Industry Association (SIA) today announced global semiconductor sales were $517.7 billion during the third quarter of 2026, an increase of 24.1% compared to the second quarter of 2026. Global sales were $172.6 billion during the month of September 2026, an increase of 146.8% compared to September 2025 and 7.9% more than sales in August 2026. Monthly sales are compiled by the World Semiconductor Trade Statistics (WSTS) organization and represent a three-month moving average.

Michael Robbins, SIA president and CEO"The global chip market grew substantially throughout the third quarter of this year, with worldwide sales increasing by more than 20% quarter-to-quarter for the third consecutive quarter. Year-to-year semiconductor sales remained strong across all major regional markets, highlighting the industry's global reach and its indispensable role in powering innovation, productivity, and economic competitiveness."

As spotted by Club386, two motherboards based on Intel's Z970 chipset have shown up in the NBD shipping database, which logs Vietnamese customs records. Going by the names, both appear to be from ASRock: the Z970 LiveMixer WiFi and the Z970 Pro-A WiFi. The entries refer to PCBs shipping at the end of September 2026. The Z970 chipset is rumored to support CPU multiplier overclocking, but not the base clock tweaking that Z990 is expected to offer. Leaker Jaykihn has also previously reported that it will replace most of the market the B860 covers today. If that ends up being the case, it would be a big change, since Intel has kept CPU overclocking on its most expensive chipset for years, while AMD allows it on most of its B-series boards.

Intel's Robert Hallock has previously said the goal is more unlocked SKUs over time, and that overclocking shouldn't be reserved for the people paying the most. It looks like that approach could extend to motherboards too. Of course, Intel hasn't announced Z970 or Z990 yet, so this is all unofficial, and we don't know what the prices of either would look like. Moreover, shipments this early don't necessarily mean retail boards are close. Intel has reportedly promised more details at CES in January 2027, which fits the rumored CES 2027 launch window. If Z970 does arrive at B860 prices, it could be the most interesting platform for most builders.

Leaker Kepler_L2 says AMD will bring RDNA 5+ to its Zen 7 "Grimlock" APUs in 2029. He posted on the NeoGAF forum, replying to a user who guessed RDNA 5 would trickle down to regular mobile chips in 2028 or later. His answer was "RDNA5+ in 2029 with Grimlock." He gave no details on how RDNA 5+ differs from RDNA 5, and AMD hasn't confirmed the name or timing. Grimlock is the codename for the Zen 7 Ryzen generation. This update follows Kepler_L2's January claim that AMD is splitting its APU roadmap in two. Products that don't need strong integrated graphics, like office laptops or laptops with a high-end discrete GPU, stay on RDNA 3.5 until 2029, while "premium" parts get RDNA 5. Medusa Halo is expected to be one of those.

The picture is messy, though. We previously reported that Medusa Point should get RDNA 4m, a tweaked RDNA 3.5 with FSR 4 support, and AMD's HIPRT 3.2 release notes list Medusa support for gfx1170, gfx1171 and gfx115E. The gfx115x family is RDNA 3.5, so it looks like at least one Medusa chip could stick with it. If this ends up being true, anyone waiting on a mainstream Ryzen laptop or handheld with RDNA 5 graphics may be waiting a long time. That's a long run for one iGPU architecture, but of course, this is all leak territory until AMD says something concrete.

PC Games Hardware has put a modded GeForce RTX 2080 Ti 22 GB through their test gauntlet, and the extra memory seemingly pays off once games run out of headroom. At 4K with DLSS Performance and FSR 3.1 Frame Generation, the card averaged 52.5 FPS in Flight Simulator 2024 against 30.2 FPS on a regular 11 GB model, roughly 74% faster. Spider-Man 2 jumped from 39.4 to 64.0 FPS, and Indiana Jones and the Great Circle crashed to the desktop on the 11 GB card. At 1440p with DLSS Quality, the gap was smaller, though those first two games still ran about 20% slower on 11 GB. Of course, PCGH chose nine memory-hungry games on purpose, and most titles run fine on 11 GB. Keep in mind that these are just one outlet's numbers, and other results may vary.

These cards have shown up everywhere from Alibaba to eBay listings at $499, and this one came from a reader. All eleven memory chips are swapped for 2 GB Samsung parts. No firmware changes are needed, since relocating a single strap resistor tells the GPU about the extra capacity, and NVIDIA's regular drivers pick it up. Its weak blower-style cooler and 250 W limit held it back, so PCGH tuned both cards to identical 1.9 GHz, 15 GT/s, and 280 W settings. At around €500, the 22 GB card costs quite a bit less than a 16 GB RTX 5060 Ti, which goes for €770 in Germany. PCGH still calls it a gamble, citing eight-year-old hardware and soldering work that can fail weeks later, and advises checking the seller's reputation and whether any warranty goes beyond a 30-day return window.

Two days after arguing that the DRAM in home PCs would be put to better use in data centers, OpenAI CEO Sam Altman sat down with around 50 Samsung presidents on Saturday, October 31. The group went well past the chip business, with leaders from Samsung's finance and biotech arms joining the electronics side at the company's training center in Yongin. Samsung presented it as part of a leadership seminar series for its company-wide AI push, and the talks apparently covered where AI is headed, the markets it could open up, and how it changes the way large companies make decisions. According to Samsung, Altman told the room that adopting AI alone doesn't give anyone an edge, and that the payoff comes from using it to move faster and build better products.

Nothing came out of it in terms of a new supply contract, at least publicly, and Samsung's own write-up stuck to AI strategy. Still, it's hard to look at the visit separately from the memory shortage. Samsung signed a letter of intent with OpenAI back in October 2025 to act as a memory partner for Stargate, with OpenAI's demand projected at up to 900,000 DRAM wafers per month. Separately, on October 31, SK Group chairman Chey Tae-won also reportedly met Altman in Seoul, with SK hynix and SK Telecom executives joining, according to industry sources. Last month, AMD's Lisa Su spent hours with Samsung's chip chief chasing HBM supply, and amid all these meetings, Micron's CEO expects the memory crunch to get worse in 2027 and 2028. It seems like every major buyer is knocking on the same Korean doors.

Haswell owners finally have a way to get Resizable BAR running on Intel 8-series motherboards. As spotted by Haze on X, a developer going by roschank has released reBART, an open-source UEFI driver for Z87, H87, Q87, B85, Q85, and H81 boards. Back in 2020, it was reported that Intel CPUs since Haswell support Resizable BAR, pending firmware updates from motherboard makers that never came for 8-series. The developer went digging after failing to get Above 4G Decoding working on their own B85 board, and found that Intel's 8-series platform code skips mapping memory above 4 GB altogether, something both 7-series and 9-series firmware handle. Apparently, every board maker they checked inherited the same gap, so the 4G toggle some of these BIOSes expose likely does nothing.

reBART lets the board boot normally and only kicks in right before the OS loads. At that point, it shifts the GPU's memory window up past the 4 GB line and points the boot display at the new location too. Unlike ReBarUEFI, it doesn't need module patching, though a DSDT patch is still required. Newer builds added experimental support for Intel Arc cards and multi-GPU setups, which is important since Intel once told Arc buyers without ReBAR to stick with other vendors. One user reports an Arc A750 and an Arc B580 both working with an unlimited BAR size on a Gigabyte GA-Z87M-D3H under Ubuntu. Of course, going this route means flashing a modified BIOS, and at least one ASUS Z87-Plus owner couldn't get the board to boot after the DSDT patch. The developer hasn't shared any performance numbers yet either.

Intel is moving some manufacturing capacity away from consumer chips to keep up with demand for its Xeon server processors. During its Q3 2026 earnings call, the company said it will convert some Intel 18A resources at its Oregon facilities to Intel 3, the node behind Xeon 6 "Granite Rapids." According to CFO Dave Zinsner, those 18A wafers "are generally going to build Panther Lake products." Zinsner also indicated that weaker PC demand gives Intel room to make the switch, as the company expects the PC slowdown to last at least through the first half of next year. CEO Lip-Bu Tan said the added Intel 3 capacity should be in production in the second half of 2027.

Intel CEO Lip-Bu TanThe industry continues to navigate one of the most significant supply shortages in the history, driven by the unprecedented demand for Super Intelligence. Against this backdrop, I am pleased by the progress we made growing our server supply in Q3. I'm mindful about the ongoing challenges we face in the future quarters to keep pace with the server demand.

Intel shares fell roughly 29% on Friday to $65.76 after the company's fourth-quarter outlook came in well below Wall Street expectations. According to CNBC, it was the stock's worst day since the 1970s. Intel had closed at $92.07 on Thursday, right before it posted its third-quarter 2026 results. Oddly enough, the quarter itself was a good one. Revenue came in at $16.5 billion, up 21% year over year, and adjusted EPS of $0.52 easily beat the roughly $0.40 analysts were expecting. Data Center and AI revenue jumped 78% to $7.3 billion, while the Client Computing Group slipped 5% to $8.1 billion.

Investors, however, were looking at Q4. Intel expects revenue between $15 billion and $16 billion, against a consensus of around $17 billion, with adjusted EPS of just $0.25. Non-GAAP gross margin is also expected to fall to about 38% from 44% in Q3. Part of the shortfall comes from a plant shutdown in Israel. CFO David Zinsner said Intel sees "the PC slowdown continuing at least through the first half of next year." On the earnings call, the company said it expects the PC market to shrink by about 12% in units this year, and it reportedly sees another 10% decline in 2027.