Having jettisoned 4F² memory cells from its roadmap, Micron must use different methods to continue scaling its DRAM devices in the coming years to keep up with rivals planning to use smaller DRAM transistors towards the end of the decade. Instead, Micron intends to introduce wafer bonding with its upcoming 1ε DRAM process technology and adopt High-NA EUV lithography for subsequent generations. Meanwhile, the company is developing new types of memory, including vertical high-bandwidth memory (vHBM) and vertical edge inference memory (vEIM), which will be integrated directly on top of processors.

"DRAM is nearing atomic limits at this point," said Scott DeBoer, president and chief technology and products officer at Micron, during the company's event this week. "Although we have confidence in several more planar nodes, we are very focused on making sure that we introduce 3D DRAM at the optimal time from a cost and performance point of view."

Wafer bonding, High-NA EUV, and new materials

Micron's current 1γ DRAM process is already in high-volume manufacturing, whereas its successor, 1δ, is scheduled to enter volume production in the second half of 2027. The latter promises a mid-20% increase in bit density and is optimized for future AI memory products, including HBM5, vHBM, and vEIM. Beyond 1δ, Micron plans to introduce wafer bonding with its 1ε process, which is currently under development. However, the company has not disclosed how it intends to implement wafer bonding or what improvements it expects from the technology.

For subsequent generations — ζÅ and ηÅ — Micron plans to adopt High-NA EUV lithography (ζÅ, zeta angstrom, Micron's 1^st Generation sub-10nm-class DRAM node) and new materials (ηÅ, eta angstrom, Micron's 2^nd Generation sub-10nm-class DRAM node) to continue shrinking DRAM cells. For many of its upcoming fabrication technologies that will be used over the next five to 10 years, Micron is also collaborating with Albany NanoTech. Of course, as with all organizations like Albany NanoTech, we are mostly talking about fundamental, or pre-competitive, IP that could be used for DRAM production at some point.

 "Albany NanoTech, our announced focus on advanced lithography [is] really driving High-NA EUV roadmap for DRAM," said DeBoer. "In particular at Albany with our team there, focused on the next five to ten years."

Micron is also developing 3D DRAM technologies and lists two generations — 3D1 and 3D2 — in its roadmap already. However, the company has not disclosed its architectures or launch schedules, as it intends to introduce 3D DRAM only when the technology offers sufficient advantages in manufacturing costs, complexity, and performance over conventional planar DRAM.

"When we look at three 3D DRAM we have that on the roadmap, as I mentioned, and it is shown here in between one zeta and one eta," said Micron's CTO. "But in reality, this is a parallel path, and we will implement it [3D DRAM] in a fashion that is ideal for cost, complexity, and device performance."

Going vertical, becoming critical

Beyond developing new process technologies for its upcoming DRAM products, Micron is also working on all-new types of DRAM products that are tailored for very specific use cases, including vertical HBM (vHBM), vertical edge inference memory (vEIM), and various custom HBM (cHBM) implementations.

vHBM will place memory stacks directly on top of processors (such as AI accelerators and AI GPUs) rather than next to them. This approach promises substantially lower latencies and higher bandwidth thanks to shorter connections. However, placing memory above a processor creates significant thermal challenges and limits the number of memory stacks that can be integrated on top of an ASIC. Therefore, Micron expects vHBM to complement conventional HBM, which will continue to provide the bulk of memory capacity in many AI accelerators.

Another technology under development is vertical edge inference memory (vEIM), which shares architectural and packaging characteristics with vHBM but targets different applications. Whereas vHBM prioritizes maximum bandwidth for high-performance AI accelerators, vEIM is focused on energy efficiency for edge AI processors used in autonomous vehicles, robotics, and other power-constrained devices. Nonetheless, even the 1^st Generation vEIM targets formidable 2 TB/s bandwidth at

Both vertical memory architectures will require advanced packaging technologies, particularly hybrid bonding, which enables much denser connections than conventional microbumps. Thermal management will also be critical, especially for vHBM. To meet cooling requirements, Micron is developing specialized heat-spreading structures to transfer heat from hot regions of memory base dies through the DRAM stacks.

The company has not disclosed when vHBM, vEIM, or cHBM will enter commercial production. What is important about the development of such types of memory is that it indicates that future improvements in memory performance will increasingly depend on advanced packaging and closer integration with processors, which will require collaboration between DRAM makers and logic developers. In turn, this transforms memory from a commodity to a critically important ingredient of an ASIC design.

DRAM featuring 4F² memory cells has long been considered one of the holy grails of the DRAM industry, as they promise to reduce cell area by 33% compared with conventional 6F² designs and therefore enable building DRAMs with more memory cells in the same area. However, after working on 4F² memory cells for a couple of decades, Micron has concluded that the architecture offers no meaningful density advantages, increases manufacturing complexity and costs, and has fundamental scaling limitations. This has a dramatic impact on Micron's future plans.

"Micron invented 4F² DRAM," said Scott DeBoer, president and chief technology and products officer at Micron, during the company's event on Wednesday. "The path with 4F² DRAM is always been an option through time, and we have evaluated it and not pursued it due to increased complexity and cost through this past two-decade period of time."

Interestingly, both Samsung and SK hynix still have 4F² memory cells in their roadmaps (according to TechInsights), so Micron's rejection of the technology puts the company on a unique trajectory. Several years ago, the company delayed integration of EUV lithography into its manufacturing flow, contrary to what its rivals did, and still managed to retain competitive positions both in terms of costs and in terms of performance. Therefore, skipping 4F² memory cells is not completely surprising.

4F² is not the future of Micron's DRAM

The main attraction of 4F² DRAM is its theoretically smaller memory cell. The 4F² naming refers to a cell area equivalent to four times the square of the minimum lithographic feature size (F), compared with six times F² for conventional DRAM.

At identical feature dimensions, a 4F² cell occupies approximately 33% less area than a 6F² cell, which could theoretically increase the number of memory cells per wafer by 50%. Such an improvement would reduce manufacturing costs per bit and increase per-device memory capacity even without an equivalent reduction in feature dimensions. However, to make these advantages possible, DRAM makers must develop a suitable transistor architecture and a sufficiently compact capacitor arrangement. This is not easy at all.

The CTO of Micron says that the theoretical advantages of 4F² DRAM do not necessarily translate into higher memory density in practice. DeBoer explained that dimensions and pitch of DRAM capacitors, rather than the nominal area of DRAM transistors, ultimately determine how many memory cells can fit on a wafer. Both 4F² and 6F² transistor structures must fit beneath the capacitors, which means that reducing transistor area alone does not guarantee a corresponding reduction in overall cell size because transistors need a capacitor. Furthermore, Micron believes that 4F² introduces additional manufacturing steps and complexity that increase costs without delivering sufficient benefits.

"It is important to understand the capacitor pitch, the capacitor size, and dimensions, is what sets the DRAM bit density, not whether it is 6F², 4F², 8F²," DeBoer said. "From our investigation of this technology over the past couple of decades, it is a bit of a myth that 4F² actually provides anything versus 6F². It is capacitor pitch that defines the DRAM. You can build the cells, the transistors underlying them in different ways, but they still have to fit under the capacitor pitch, which sets the DRAM [cell size]. So, some of that is just smoke and mirrors in terms of talking about 6F² versus 4F²."

4F² memory cells may curb new process technologies

Instead of adopting the 4F² DRAM cell architecture, Micron plans to continue scaling conventional 6F² DRAM. The company believes it can increase bit density by reducing capacitor pitch and shrinking process technologies even without introducing a fundamentally different transistor architecture.

Advanced lithography, new materials, and wafer bonding should allow Micron to extend planar DRAM scaling for several more generations, according to the company. For example, its upcoming 1δ process technology, scheduled to enter volume production in the second half of 2027, promises a mid-20% improvement in bit density compared with its predecessor. Micron also plans to adopt High-NA EUV lithography for future DRAM generations.

Interestingly, DeBoer fully acknowledges that 4F² DRAM is technically feasible and that Micron itself has manufactured experimental devices based on the architecture. However, its manufacturing complexity and costs become increasingly problematic as feature dimensions shrink, so essentially Micron claims that it can either adopt 4F² or use more sophisticated process technologies, a claim that perhaps implies why China-based CXMT recently praised 4F². Although DeBoer did not identify any specific physical limitations involved, he stated that 4F² has no sustainable scaling path beyond a certain point. This suggests that Micron sees little value in introducing an intermediate architecture that would eventually require another major transition.

"When you look at a scaling path, you can go back a few nodes and absolutely, 4F² can be built back a few nodes, no question about it," DeBoer noted. "It is more complicated, it is more complex, it is more costly, but you can do that. When we look at scaling 4F² beyond where we are sitting today, there it is; it is physics. There are fundamental problems with 4F² scaling that ultimately will lead to 3D DRAM coming in instead of 4F²."

3D DRAM is the future

Indeed, Micron's preferred long-term solution is 3D DRAM, which would increase memory density by stacking memory cells vertically rather than relying exclusively on reductions in their lateral dimensions. Alongside development of multiple next-generation planar DRAM process technologies, the company is already developing multiple generations of 3D DRAM production nodes (which is not surprising, as 3D NAND nodes will rely on fundamental innovations developed for planar processes), although it has not disclosed when the new architecture will enter commercial production. DeBoer indicated that Micron will make the transition when 3D DRAM offers the right combination of density, manufacturing complexity, and cost.

Although both Samsung and SK hynix have 4F² DRAM products in their roadmaps, Micron believes that the assumption that reducing the nominal footprint of DRAM cells improves manufacturing economics is wrong and that extending 6F² scaling and eventually adopting 3D DRAM offers a better long-term strategy.

The two biggest NAND flash memory manufacturers on the planet, Samsung Electronics and SK Hynix, are reportedly working to upgrade their Chinese factories to produce more advanced chips for use in AI data centers, Reuters reports. The move comes as the two companies continue to fight off competition from the fast-growing Chinese memory makers like CXMT. Fierce competition in the lucrative AI RAM market has already seen manufacturers slash consumer-focused production to allow them to increase production capacity for the more advanced — and money-spinning — components.

Citing four people familiar with the companies' plans, the report claims that Samsung Electronics aims to switch its Xian plant to the production of 286-layer flash chips by the end of the year. SK Hynix has its sights set a little lower, targeting chips with 240 or more layers at its Dalian facility. That factory was taken over back in 2021, but a timescale for the new upgrades isn't clear.

However, while Samsung Electronics and SK Hynix are well placed to deal with the growing threat from Chinese chipmakers, both are faced with additional complications in the form of U.S. regulations. Export controls on chipmaking equipment were applied in 2022, with the aim of slowing China's technological progress. Both companies, based in South Korea, must now get licenses in order to install U.S.-built hardware in Chinese factories. The report points to this as a reason for the recent growth of Chinese firms like Yangtze Memory Technologies Co. (YMTC). The company told investors at a recent IPO preparation meeting that it aims to become the world's largest NAND flash producer by the end of 2027.

While both Samsung Electronics and SK Hynix say that they are working in consultation with the U.S. authorities, there's little doubt that this extra step has hampered the companies' ability to upgrade their factories to better meet the new AI memory demand.

It's impossible to overstate just how key Samsung and SK Hynix are to NAND supply. Samsung's Xian factory produces around 40% of the company's global output. It currently produces chips that have around 200 layers, but Samsung's plans aim to boost that to 286 layers for V9 NAND chips. For its part, SK Hynix's Dalian plant currently produces chips with 192 layers, with the company's aim for 240-layer chips reportedly confirmed by three people familiar with the plans.

Meanwhile, CXMT, China's premier chipmaker, is looking to move beyond DRAM production. A report from September 2026 noted that it was eyeing production of 3D NAND memory for use in solid-state storage (SSD) products. The country's largest DRAM producer, CXMT, posted $22.41 billion in revenue and $11.57 billion in net profit in the first half of 2026 after years of losing money. But despite this purple patch, the company is still much smaller than the likes of Samsung, SK Hynix, and Micron.

Micron faces a one-day labor stoppage at its Taoyuan operations in Taiwan on October 22, Reuters reports. Micron's employees are striking for concrete terms for this year's bonus and a clearer mechanism for sharing the company's future profits. The strike's teeth come from the resultant financial impact on Micron's operations, as semiconductor fabs are supposed to operate around the clock. The scheduled protest comes amid strong demand for memory chips and will have a minor impact on Micron's memory output.

The Taoyuan union wants Micron to present a specific proposal for this year's additional employee bonus and establish a transparent formula for distributing a portion of future profits. Previous negotiations between the company and the union ended unsuccessfully in September, which prompted employees to approve industrial action in early October. Nonetheless, the union stresses that the October 22 strike is intended to pressure Micron into reaching an agreement rather than terminate negotiations. The union is prepared to conclude a deal if the company presents terms acceptable to employees across its Taiwanese operations.

The situation could become more serious if employees at Micron's Taichung operations, its largest manufacturing site in Taiwan, decide to join the action. A separate labor organization that represents workers at that facility is scheduled to meet with Micron for mediation on October 22, the same day as the Taoyuan protest. However, the Taichung union has not yet authorized a strike.

Micron's fabs in Taiwan produce commodity DDR5 and LPDDR as well as expensive high-bandwidth memory (HBM) for AI accelerators. Amid record demand for memory of all types, even a one-day suspension of operations would have a material impact on Micron's output and earnings.

Modern semiconductor fabs are supposed to work 24/7/365, as many processes cannot be halted. However, if management has sufficient time and personnel to prepare, a one-day warning strike does not necessarily mean a complete shutdown of Micron's Taoyuan manufacturing operations, not to mention permanent damage to wafers or equipment. The actual consequences of the strike will depend on employee participation, which manufacturing operations are affected, and Micron's contingency plans. While it is likely that the output will drop, the decline will likely be negligible.

Micron employs approximately 15,000 people in Taiwan. The workers point to compensation arrangements at Samsung Electronics and SK hynix as examples of how semiconductor manufacturers can distribute profits among employees as AI-driven demand generates unprecedented cash flows for memory firms.

What if you could buy a cheap used DDR5 SODIMM (laptop memory) kit and transform it into a super-fast DDR5 desktop kit? That’s what electronics hobbyist Blighted Agent recently did when they succeeded in converting a SK hynix 2x 16GB DDR5-5600 SODIMM kit they got from Facebook Marketplace for $120 into a 2x 16GB DDR5-7800 desktop kit, claimed to be worth around $1,000. They also had to buy some blank DDR5 desktop memory PCBs from AliExpress for around $16 each, and a set of heatsinks for around $18.

In the above video, our hero says they bought the donor laptop 2x 16GB DDR5-5600 sticks for just $120, which is a bargain. With extra bits and pieces from AliExpress (desktop DDR5 PCBs, etc.), the total bill of materials came to $175. Blighted Agent comments that the SODIMMs used what he believed to be SK hynix A-die modules, which are the same as the ICs used in expensive desktop DDR5 modules like the G.Skill Trident Royal line.

The process began with taking the memory ICs from the donor modules using a hot air station. Then they were cleaned and had 0.45mm leaded solder balls added, using a DDR5 stencil, reballing jig, and hot air. Next, the blank desktop DDR5 PCBs were prepared. A little solder was added to the plethora of BGA connectors to be ready for the memory chips to connect to.

The old DDR5 ICs and the new desktop modules could now be mated, using flux, the preheater, hot air, and watching and prodding carefully with a dental pick. Blighted Agent’s experience lets him know when the BGA chips have successfully mated with the PCB. You can eyeball when they pop precisely into place due to surface tension, and offer up a modicum of resistance to pick prodding.

Finally, the enthusiast added the SPD chip and BIOS chip from the SODIMMs to the new desktop DIMMs. These new desktop DIMMs were quickly tested in a PC before they were finished off with carefully applied thermal pads and the fancy-looking heatsinks.

Blighted Agent felt lucky that there were no issues with the newly created desktop modules. There’s a lot of intricate things that can go wrong during such a process. The initial boot verification showed them recognized in his Asus ROG Maximus Z790 Apex Encore motherboard as two “SK hynix 16,384MB 5600MHz” modules, which was correct. Some tuning resulted in rock-steady performance at DDR5-7800 CL38 and 1.40V. This was verified by stress testing for 90 minutes, which the enthusiast reckons was ample for a gaming PC. It is very likely a better result than using those mechanical SODIMM-to-DIMM converters you can buy, which can be problematic.

Blighted Agent seems to specialize in BGA soldering work, and has also successfully upgraded several graphics cards with double the VRAM. Though videos like this are edited to be brief, and they make the process look easy, the TechTuber admits there is a lot that can go wrong during this kind of rework. Moreover, with RAM modules, they lack the testing equipment to locate an issue with any particular IC on the PCB. So, if this module didn’t work, they might have ended up in a world of pain, not knowing which memory chip had an issue due to either an internal fault or a BGA soldering glitch.

Looking around at used DDR5 SODIMMs, it looks like Blighted Agent was very lucky to get a used working 32GB kit for $120. I just checked eBay U.S., and the best deal there was $320. Meanwhile, the best price for a new 32GB DDR5-6000 kit seems to be $449 at the time of writing.

We reported on Russian enthusiast VIK-on succeeding in a similar project earlier this year, but all their costing was in Rubles, and they went the extra mile in flashing custom firmware so the remade modules would be auto-detected with 6400 MT/s XMP support on most motherboards. In late 2026, DDR5 RAM continues to be too expensive.

Are you looking to buy RAM for an upgrade or an existing PC? You've no doubt seen the news about the extortionate price rises on RAM and storage, thanks to the ongoing AI boom ruining the component market for PC enthusiasts; well, today, we have some slight light in the darkness, with a $108 price cut on some DDR5 memory from Corsair. You can pick up 32GB of Corsair Vengeance RGB RAM (DDR5-5600) for $409 direct from Corsair. I've checked against our own price tracking info and used various other online resources to confirm that this is indeed the cheapest available to purchase memory kit with these speeds.

● Check out this deal at Corsair

This 32GB kit comprises two 16GB matched memory sticks, both with RGB lighting running along the spine of the heat shroud. The Corsair Vengeance RGB RAM is DDR5 with speeds of 5600 MT/s and a CAS latency of 40. This kit comes with Intel XMP profiles that can be selected in the BIOS.

We can't compare prices of RAM now to those of even a year ago. That's just unrealistic in the current AI economic climate, and whether you love or hate AI, it's not going anywhere soon; the rollouts of large data centers and manufacturers focusing resources on profit-heavy AI component production will continue for the foreseeable future.

There's no denying that RAM prices are tough right now. All the signs point to prices going up even further as scarcity becomes more of an issue for anyone wanting to build a PC in the next few years. If you're waiting for a fall in price, or the AI bubble to burst, you could still be waiting for up to a decade for memory and RAM production to normalize to anywhere near previous figures.

If you desperately need RAM now, this is one of the better prices on a 32GB kit. There are faster memory kits available, but you have to, of course, pay more for them, and the performance upgrades in most situations won't even be perceivable for the average user.

Elon Musk has criticized memory makers for what he calls 'extortionate' pricing amid an unprecedented shortage of both conventional DRAM and high-bandwidth memory (HBM). Although Musk did not identify the companies he criticized, his remarks seem to be directed at Samsung Electronics and SK hynix, as he praised Micron for its more 'reasonable' approach at the same time. Interestingly, Micron also enjoys exceptionally high profitability: its latest quarterly gross margin reached 81%.

"Micron has been reasonable regarding future memory prices, while others have been extortionate, which is something today's customers will not forget," Musk wrote in a post on X. "Micron is making the smart move for long-term customer loyalty."

Musk did not provide specific details of said extortionate pricing or disclose terms offered by different memory suppliers (which tend to be confidential). Consequently, it is impossible to determine whether Samsung and SK hynix charge substantially more than does Micron for comparable products. Nevertheless, Musk's remarks suggest considerable frustration with the pricing policies of South Korean memory manufacturers.

Over the past several quarters, memory has transformed from an easy-to-obtain commodity to a strategic asset, and makers of 3D NAND and DRAM have certainly benefitted massively from this shift. Historically, memory makers have responded to strong demand and rising prices with aggressive capacity expansions, which eventually led to oversupply and falling prices. This time around, however, manufacturers have raised memory prices severalfold without comparable increases in production capacity. As a result, supply still lags behind demand and will continue to do so for years to come, which puts further upward pressure on memory prices.

What makes Musk's praise of Micron particularly interesting is the company's financial performance, which clearly suggests that the company is taking advantage of the situation in the market. In its fourth quarter of fiscal 2026, Micron reported revenue of $54.23 billion and a GAAP gross margin of 81%, up from 40% a year earlier. The company's cloud memory unit earned $16.283 billion, and its non-GAAP gross margin totaled 83%; its core data center storage business sold goods worth $18.002 billion with a 90% gross margin; while sales of the automotive and embedded units totaled $6.824 billion with gross margins reaching 84%.

Perhaps the most surprising part of Micron's last quarter report was its mobile and client revenue of $13.114 billion and non-GAAP gross margin of 90%, which means that Micron earned higher profit selling memory and storage for client devices than it did for selling memory solutions for data centers. Such results clearly indicate that even supposedly reasonable memory prices can generate extraordinary profits in the current market.

Memory manufacturers are increasingly shifting toward long-term supply agreements (LTAs) that guarantee future supply and provide customers with predictable prices of DRAM, HBM, and NAND. So far, Micron has signed 26 strategic customer agreements that cover more than 35% of its expected revenue through 2030. These contracts include $32 billion in customer financial commitments, most of which consist of cash deposits. The company acknowledges that its new agreements reflect higher prevailing memory prices, and it by no means is trying to sell memory at a discount.

Samsung has reportedly secured long-term contracts covering 75% – 80% of its planned memory output in 2027, according to South Korea's ⁠Korea Economic Daily (via @Jukan05). SK hynix is reportedly pursuing similar arrangements as AI infrastructure investments continue to absorb available memory capacity.

Market insiders and analysts expect memory supply shortages to persist through at least 2028, and that gives manufacturers considerable leverage in contract negotiations. Memory suppliers have little incentive to reduce prices and are eager to enjoy their position, but Musk's warning suggests that today’s pricing policies could influence customer relationships once supply and demand eventually return to balance.

The AI-induced PC component shortage, which has caused drastic increases in the cost of DIY building, continues to get worse. Since last year, RAM prices have increased by more than 500% and show no sign of slowing down. In August, we reported that DDR5 6000 RAM — the sweet spot benchmark for both performance and pricing — had reached $400; now, the price is $450.

• Save by going slower: CORSAIR Vengeance RGB 32GB (2 x 16GB) DDR5 5200 - $399 at Newegg

As mentioned, because DDR5 6000 RAM is a de facto standard configuration and capacity in terms of performance, tracking the price of this particular spec has proven to be a strong barometer of the industry overall. While we track RAM prices across multiple models and kits, we also track the price of 32GB DDR5 6000 RAM daily.

We've been creeping closer to $450 for some time. As of today, the cheapest 32GB DDR5 6000 kit on the market costs $449.97, or three cents shy of $450. Cheaper RAM is still available; this 32GB DDR5 4800 kit is $399 at Corsair, but good, speedy RAM is now harder to buy than ever before. The 4800 Corsair kit once cost just $80.

As RAM prices and other component hikes have a knock-on effect on PC pricing as a whole, it was recently reported that PC shipments in 3Q26 have tumbled by a massive 20%, with the top three vendors shipping 11.6 million fewer units than the same period last year.

Combo deals remain a rare bright spot. Various Newegg bundles could save you as much as $358 when you buy RAM alongside a motherboard, CPU, and even an SSD, and we've seen a recent resurgence in stock on these.

Pre-builts also continue to offer some solace to wannabe DIY builders, with price hikes trickling down to systems more slowly. With industry figures warning that memory shortages could persist well into next year, or even last up to a decade, in the space of just 12 months RAM has gone from being a PC gaming commodity to a luxury few

CXMT expects to launch its first 4F² product, a server DDR5 RDIMM, by the end of the year, company president Cao Kanyu said in remarks reported by Yicai. The memory will be built on a denser cell layout, not yet commercialized, using vertical transistors to reduce the cell’s footprint by about one-third on paper. No speeds or capacities have been announced yet for the new product.

Cao spoke about the new RDIMM at the Integrated Chips and Chiplets Conference on Saturday. A 4F² cell is 2F by 2F at about two-thirds the area of current 6F² cells at the same feature size, and designs so far all stand the transistor up. The company has produced research and patent filings on the technology with papers in 2021, 2023, and 2025, and an 8Gb full-array research example was demonstrated at IEDM 2023.

Cao said CXMT’s version, the Vertical Surrounded Gate (VSG) transistor, has a channel that stands vertically with the gate wrapped around it, allowing stronger gate control with less leakage. Most importantly, standing the transistor up shrinks the cell. CXMT now has more than 300 published patent families tied to 4F², he added. The report says the industry as a whole is moving toward 4F² with 3D solutions that decouple DRAM and CMOS processes, combining them with wafer-level hybrid bonding, which allows logic and cell density to be optimized independently. Cao also covered CXMT’s own hybrid-bonding work, which began in 2019, with published patents.

Aside from this announcement, China’s primary DRAM maker has been in the news lately for its manufacturing and market-share gains. Its fifth-generation G5 DRAM technology platform, announced Sept. 20, is in mass production. TrendForce puts CXMT’s share of global DRAM revenue at 9.5% in the second quarter, up from 7.6% in the first, good for fourth place. Last November, CXMT showed off DDR5-8000 and LPDDR5X-10667. It has revealed itself as a manufacturer to be taken seriously by the biggest existing players.

Samsung showed off a 16Gb, 10nm-class 4F² DRAM design in its ISSCC 2026 paper in February. That one also used vertical-channel transistors (VCT), with the array bonded on top of the periphery using wafer-to-wafer hybrid copper bonding. The company claims the bonding raises dies per wafer by more than 20% with the same design rules. In April, The Elec reported that Samsung had a working 4F² VCT die on its sub-10nm “10a” process, with mass production targeted for 2028. Later South Korean reports clouded that date: the VCT version has since been split off as “D0a-V” with no timeline, and there is no firm timeline for adopting 4F².

SK hynix laid out plans at VLSI 2025 for a 4F² vertical-gate platform and 3D DRAM at nodes of 10nm and below, with circuitry bonded under the cell array. The April report also puts SK hynix’s 4F² debut at the 10b node, one step after Samsung’s 10a. Micron’s plans remain less clear, with no widely reported 2026 statement.

The announcement is perhaps more interesting for what wasn’t mentioned: capacity, speed, die density, process node, named customers, or intended volume. All we have is a date and a purported cell design, admittedly well-supported by the literature, but no product that has been tested by anyone outside of CXMT. Beyond Cao’s talk, CXMT hasn’t posted a release about the product in its Chinese or English newsrooms.

South Korean reports have said that CXMT’s initial products may only partly use 4F² and that producing stable units at volume is a separate challenge. With about 11 weeks left in the year, it will be interesting to see what CXMT actually delivers.

If you're in the market for a deal on DDR5 RAM during the Amazon Prime sales, Corsair might have the answer with a sale on 32GB of DDR5 5600 memory. Right now you can save $103 and grab this kit for $414.99.

• Check out this deal at Corsair

This is obviously a lot more expensive than the RAM prices from a year ago, but it's as good as it gets in today's market, with prices continuing to rise and showing no sign of letting up anytime soon. The cheapest DDR5 6000 kit — the sweet spot for AM5 gaming — is $437. For slightly more modest speeds, this is a decent saving, and the cheapest kit of this capacity and clock on the market. Crucially (pun not intended), this kit supports both AMD EXPO and Intel XMP, so will work with either platform for best speeds. If your rig runs with an Intel CPU, you can save $5 more with this kit.

There's no denying that RAM prices are tough right now. Unfortunately, all the signs point to things getting worse, rather than better, with some industry figures suggesting the RAM shortage could last an entire decade.

If you need RAM right now, either for a new build or to upgrade your existing system, then this is about as good as it gets.

Corsair is actually hosting a RAM sale across lots of its products, with savings of up to 30%. Prices remain high, but you can check out the whole sale here.

Best Tech and PC deals | Best gaming PC deals | Best RAM combo deals | Best 3D printer deals | Best RAM deals | Best gaming laptop deals | Best gaming monitor deals | Best Wi-Fi Router deals | Best GPU deals | Best SSD deals | Best hard drive HDD deals | Best CPU deals | Best gaming chair deals | Best PC building tool deals | Best PC peripherals deals | Best filament and resin deals | Best motherboard deals | Best CPU cooler deals | Best PC case deals | Best Dell and Alienware deals | Best USB charger deals | Best gaming and productivity laptop deals under $1,000 | Best laptop PC deals

Also, you can join the Tom's Hardware deals Discord for up-to-the-minute hardware deals.

A YouTube channel best known for its skillful graphics card fixes has published its first RAM module repair video. Tony from NorthWestRepair’s latest video shows his attempts to mend a pair of high-capacity DDR5 sticks worth a purported $16,000. It isn’t the repair skills that are newsworthy here. Rather, it’s the shift in economics that makes the chuck-it-and-replace-it attitude to bad RAM a philosophy that simply doesn’t apply in 2026. The impact of the AI-pocalypse means that spending a couple of hundred dollars on repair labor to save a several-thousand-dollar DDR5 DIMM is now sensible. If you can repair RAM, you can now start a business.

Two DDR5 sticks were sent in to Tony, described as having physical damage. They are both ECC sticks, meant for servers, one with 128GB and the other with 256GB of DRAM ICs onboard. Tony sounded happy at the start of the video, as this is an excuse to use his Unified DDR Flasher Main Board, for which he had just acquired an RDIMM extension that works with these server sticks.

Initial close physical inspection didn’t reveal any conclusive evidence of the source of the problems. So, it was hoped electrical tests would provide more clues. A data line tester didn’t highlight any issues, though. Tony used flux and hot air to reflow the solder on the first DIMM, which is a far simpler process than a reball (where all the memory ICs are taken off, cleaned, and given new solder balls). Initial signs were that the module was now working OK. So, he moved to the next DDR5 DIMM.

The second DIMM “may have a faulty PMIC chip,” which is why it seemed more severely broken than the first module, Tony indicated. Powering on a system with this DIMM inserted showed no signs of life, and the DIMM didn’t even warm up (thermal camera checks). That contrasts with the first fix.

Because this DIMM is completely dead, Tony did some work on the non-DRAM chips like the PMIC and SPD EEPROM first. Then he decided to reflow some other chips, including memory ICs, like before, while the board was warm. Some ICs were also reballed after inspection, making Tony suspicious of them.

The repairer found a short circuit via voltage injection and thermal camera monitoring. This was fixed with a new capacitor and by fixing the short on the damaged edge of the PCB. Finally, a new PMIC chip ended the quest, and this second DDR5 DIMM powered on and trained successfully.

The repair is good to see, saving these valuable components from e-waste. However, the bigger picture is that repairing memory modules could now be a viable business. Previously, it didn’t make sense to spend a hundred or two on labor to fix a DIMM that would cost about the same or even less to replace with a new one. Now the repair fee might be just a fraction of a replacement price.

Micron released its fiscal Q4 2026 financial results on Wednesday, expectedly setting another record for revenue with $54.23 billion for the quarter, up nearly five times since the same period last year. Micron also hit a record 87% gross margin, the largest contributor to which was Micron's Mobile and Client (consumer) business unit. Further, Micron's client business was the only one that shipped less memory last quarter, despite bringing in the highest operating margin at 88%.

Micron's margins overall are up significantly year-over-year; Core Data Center surged from 25% to 85%, and Automotive and Embedded have climbed from 20% to 79%. The Mobile and Client unit also saw significant growth, with a 29% operating margin in fiscal Q4 2025, and now a margin of 88%. Cloud Memory margins grew, though not to the same extent as other units, moving from 48% to 76%.

The gross margins are interesting to look at, too. In a financial statement, the gross margin equals revenue minus cost of goods. Operating margin, on the other hand, equals revenue minus cost of goods and all other operational expenses. The operating margin simply excludes taxes and interest.

The gap between gross and operating margin is what's interesting in the breakdown above. The Mobile and Client unit has the smallest gap, with only a 2% difference. The Core Data Center unit matches the Mobile and Client unit with a 90% gross margin, but it has a lower operating margin at 85%. In other words, the cost of running Micron's consumer business is extremely low relative to the amount of revenue it currently generates.

It also saw the lowest amount of growth for the quarter, with revenue up 14% quarter-over-quarter. Micron says this was "driven by higher pricing, partially offset by lower bit shipments." Out of Micron's four business units, the Mobile and Client unit was the only one that shipped less memory in the past quarter.

For clarity's sake, a bit shipment is just a shipment. Micron measures shipments in bits rather than units. It's the capacity Micron has shipped in a quarter, not the actual number of DRAM or NAND chips.

Outside of Mobile and Client, Micron attributes higher revenue to higher pricing and bit shipments. The situation in the Mobile and Client unit echoes a sentiment we've heard elsewhere: growth in the consumer market is still happening, but it's driven by higher pricing, not growth in shipments.

Earlier this year, Intel's David Zinsner, chief financial officer, attributed a 13% YoY growth in Intel's consumer business to higher average selling price (ASP), not increased unit sales. Zinsner clarified that Intel "thought we had seen some inflation on our cost and needed to pass that on to the end customer."

Micron's outlook for fiscal Q1 2027 has revenue set at another record of $61.5 billion ± $1.5 billion, with gross margin rounding from 87% in fiscal Q4 2026 to 86.25%.

Another record quarter signals more squeeze on the DRAM and NAND markets. Echoing rival SK hynix, Micron says it expects memory and storage to be "much tighter" in 2027 and 2028, claiming that the company expects the industry to be supply-constrained in both years. That's despite a growth in supply that Micron anticipates.

The company says it expects industry NAND bit shipments to grow in the mid-20% range, and DRAM bit shipments to grow in the low-20% range. The industry broadly will produce more memory over the next two years, says Micron, but that still won't keep pace with demand.

Worrying still is this line in Micron's earnings slides: "we do not have line of sight to when supply and demand will return to balance." However, the company says it plans to increase its capital expenditure (capex) in fiscal 2027 compared to its prior estimates.

A "majority" of that increase is set aside for construction of additional clean room space, which Micron has set for 2028 through the end of the decade. Hopefully, that additional clean room space coming online will ease supply constraints, but we're still a long way from that point.

Micron said during its FY 2026 earnings call that RAM and storage chip shortages will worsen in 2027 and 2028, even as it announced a record gross margin of 86.25%. CEO and chairperson Sanjay Mehrotra said that the company has already sold most of its capacity for next year and that customers will have to pay more compared to 2026.

“In calendar 2027 as well as 2028, we see demand exceeding supply. In fact, we see greater tightness in the industry in 2027 and in 2028 versus 2026. Overall, supply-demand environment is only getting tighter,” Mehrotra said. “We do not have line of sight to when supply and demand will return to balance.”

The memory and storage industry is not holding back on increasing production, though, with Micron saying that NAND and DRAM shipments are expected to grow in the low-to-mid 20s percentage range for the next two years. However, it seems that this is not enough to match growing demand, as it expects “the industry to remain supply constrained in both years.” Still, it seems that manufacturers are prioritizing HBM demand for AI data centers, with the company expecting shipments of these chips to grow faster than conventional DRAM.

Micron and other chip manufacturers are bringing new factories online to increase output, but it’s going to take years before they can start producing chips. But even as we’re amid an unprecedented multi-year memory chip shortage, Micron also announced record financial guidance. It forecasted a record $61.5 billion FQ1 revenue and a gross margin of 86.25%, implying a gross profit of more than $53 billion. After accounting for an estimated $2.06 billion in operating expenses and a tax rate of around 15.5%, the company said its earnings per share sit at $38.15, plus or minus $1.00, for approximately 1.15 billion shares.

These are record numbers, but it’s not surprising given the amount of money being poured into AI data centers. It has gotten to the point where the memory workers from Samsung and SK hynix have demanded and received bonuses worth hundreds of thousands of dollars, with Micron’s workers in Taiwan threatening to strike if they do not receive a bigger slice of the AI pie. While this is good news for the average employee from these big companies, it’s unfortunately bad news for the average person who must pay for RAM and storage at higher prices.

Micron’s prediction goes directly against the Acer CEO’s prediction that PC prices will decline by late 2027, especially as Chinese RAM chips enter the market and consumers simply refuse to pay more for expensive memory chips. We can only wait and see which of these CEOs will be correct when it comes to DRAM and NAND pricing.

A PC RAM maker has honored its guarantee promises despite the memory kit in question’s price rising eightfold since it was bought. This is news in 2026, as we’ve heard of so many companies dealing with AI-inflated component pricing refusing to replace items, instead refunding the original purchasing price despite the products now being worth many multiples of the original price. These glad tidings come via Redditor Double-South8863, who says their $150 G.Skill Trident Z5 Neo RGB 64GB DDR5-6000 RAM kit, now priced at $1,200 on Amazon, was back at full capacity after the company replaced the stick. They received the replacement within two weeks.

“I will never buy any RAM other than G.Skill for as long as I live,” pledged Double-South8863 on Reddit. They informed their fellow social media posters that one of the 64GB RAM kit’s sticks died at the beginning of September. After some A/B testing, they were sure one of the 32GB sticks was dead.

On September 8, about a week after the issue first reared its head, Double-South8863 decided to start an RMA process and get the dud G.Skill Trident Z5 Neo RGB 32GB DDR5-6000 stick replaced. They remember buying this highly desirable RAM kit online in May 2025 for the princely sum of $150. Now it is worth a king’s ransom at $1,200 on Amazon, and price tracker sites show it's even been as high as $1,300.

Perhaps the Redditor was aware of the reports of companies and their dastardly deeds regarding RMAs, as they are active on the PCMR subreddit. However, they optimistically went ahead and completed the RMA paperwork, then received instructions, including a shipping address to initiate the warranty exchange. Double-South8863 says they mailed the misbehaving memory module back to G.Skill on September 15. The Redditor’s post celebrating G.Skill’s customer service, proclaiming that they were now a “customer for life,” was published on September 26.

If you check the G.Skill site, you will note that this RAM-first firm still offers a limited lifetime warranty on all its popular RAM kits. Please note that, according to the official terms, this lifetime warranty “automatically terminates upon discontinuation or end-of-life ("EOL") designation of that specific model by G.Skill.” That means you shouldn't expect to swap out a faulty DDR5 module for a replacement for the whole of your natural life.

PC enthusiasts and DIYers are having a hard time right now as several key components required for performance builds are at all-time-high pricing. Meanwhile, those with a working setup pray that nothing will go wrong with the precious equipment they already have. Sadly, things do break, some components more than others. However, warranty obligations seem to have come under increased strain in recent months as consumer product scarcity means component makers can’t or won’t follow RMA norms that had been established for decades. Some of the worst cases we’ve seen involve PCs and component makers trying to refund the original (low) prices paid, substituting failed parts with non-equivalent parts, or even worse.

Memory prices are at record highs due to demand from the AI sector and are not going to drop any time soon. In fact, the cost of manufacturing memory chips on a per-area basis can be higher than the price of silicon processed using TSMC's N2 and N3 fabrication technologies, which are used to make compute chips, according to an observation made by Kurnal Insights. While the comparison has merit, it has an important caveat.

Based on media leaks, Kurnal from Kurnal Insights considers that a 300-mm wafer processed using TSMC's N3 technology costs $20,000, which translates to approximately $0.283 per mm² when the wafer price is divided by its area. For TSMC's N2, the analyst assumes a considerably higher wafer price of $30,000, or around $0.424/mm². These calculations are based on unofficial information and do not account for unusable wafer edges, dicing, test structures, defects, or yield, so they should be viewed as nominal wafer-area prices rather than the actual price of usable silicon.

The calculation for DRAM works differently. Kurnal assumes a DRAM price of $1.50 per Gb and then multiplies it by the bit density of different generations. At 0.219 Gb/mm², 1y DRAM works out to $0.329/mm², whereas 1z DRAM with its 0.273 Gb/mm² density reaches $0.410/mm². Finally, 1b DRAM at 0.436 Gb/mm² produces a figure of $0.654/mm², considerably higher than the nominal $0.424/mm² calculated for an N2 wafer.

There is an important distinction here. The N2 and N3 numbers represent assumed prices charged by TSMC for processing a wafer, whereas the DRAM figures represent the potential selling value of the memory contained within a square millimeter of silicon, thus illustrating how much revenue a DRAM maker can theoretically get from a given die area at the assumed memory price.

Interestingly, Kurnal Insights' $1.50/Gb assumption is almost exactly in line with the current spot price of DDR5 eTT memory. According to DRAMeXchange, 16Gb DDR5 eTT chips carried a session-average price of $24.80 on September 21, equivalent to $1.55/Gb. At that price, 1b DRAM with a density of 0.436 Gb/mm² would be valued at approximately $0.676/mm², just 3% above Kurnal's $0.654/mm² estimate.

It goes without saying that producing DDR5 memory is considerably cheaper than making logic chips using TSMC's N3 process technology. However, with DRAM in short supply, memory makers are getting considerably more money for their silicon than they did historically.

There are a few important caveats that make the comparison somewhat less straightforward. The calculations do not account for packaging costs, with logic requiring more expensive techniques. Additionally, TSMC manufactures chips on a contract basis, and the actual price of a processed wafer depends on the volumes, customer, and other commercial terms. Thus, comparing estimated TSMC wafer prices with DRAM spot prices is not exactly an apples-to-apples comparison. Even TrendForce notes that DDR5 spot procurement remains limited and transactions are sporadic, which means spot prices do not necessarily reflect the prices at which Micron, Samsung, and SK hynix sell the bulk of their DRAM.

Therefore, contract DRAM prices would provide a considerably better basis for an economically meaningful comparison. However, we do not know contract prices, and while they are clearly higher than historical memory prices, they may not be as high as spot prices. To that end, we cannot state for sure that DRAM makers get more money for their wafers than TSMC does for its logic wafers.

China's DRAM champion CXMT has begun mass production using its 5th-generation DRAM process technology.

CXMT says its new process technology — G5, as the company calls it — reduces DRAM’s active-area half-pitch to 11.95nm using quadruple-patterning lithography. The 11.95nm half-pitch corresponds to an active-area pitch of 23.9nm, something that puts this particular dimension roughly in the range of other advanced 10nm-class DRAM technologies. However, it cannot be directly translated into a conventional DRAM node designation. Separately, CXMT says its DRAM-optimized high-K metal gate (HKMG) process reduced the height of the core cell array to 6,762nm.

The company also says it modified its process flow and introduced unspecified new materials to enable storage capacitors with a depth-to-width aspect ratio of about 45:1.The figure is certainly difficult to compare directly with competing DRAM process technologies, but SK hynix says that capacitor aspect ratios will need to exceed 100:1 as DRAM critical dimensions eventually fall below 10nm.

A 45:1 aspect ratio means CXMT’s storage capacitors are 45 times deeper than they are wide, which is what enables the company to reduce DRAM cell area and fit more memory arrays onto a wafer and retail predictable capacity. However, increasingly high aspect ratios make etching, deposition, mechanical stability, and ultimately high-yield manufacturing considerably more challenging.

The first disclosed mass-produced devices using G5 are said to be 24Gb LPDDR5X devices, which have already entered mass production. The chips offer 50% more capacity than CXMT's previous comparable products and are available in two package formats for different mobile-device designs. According to the Chinese report, the 24Gb LPDDR5X products are already entering mainstream Chinese flagship smartphones.

A nanometer caveat?

Meanwhile, there are some important caveats.

First up, CXMT's 11.95nm figure represents the active-area half-pitch rather than a conventional process-node designation, so calling G5 a 12nm process is a large degree of simplification at best.

Secondly, the two revealed physical dimensions do not mean technological parity with the latest DRAM processes from Micron, Samsung, and SK hynix, as density, capacitor scaling, transistor characteristics, power consumption, and, arguably most importantly when it comes to commodities, yields matter.

Thirdly, CXMT says G5 provides at least 50% more gross dies per wafer than G4, but without actual numbers, this is obsolete. CXMT also did not disclose yields, which makes it impossible to determine the increase in productivity, not to mention known-good dies or the actual manufacturing cost per bit. Yet again, even a claimed 11.95-class DRAM node is an achievement.

Acer chairperson and CEO Jason Chen is expecting that prices for components will start reversing in the latter half of next year. According to DigiTimes, Chen told reporters that only high-end DDR5 parts like LPDDR5X-9600 and niche CPUs like Nvidia’s N1 and N1X chips are in short supply. Since supplier pricing lags consumer pricing by about a few months, he also added that PC prices will still rise between 5% and 20% towards the end of this year, plateau by the first half of 2027, before finally declining after years of AI-drive hikes.

The AI boom has driven shortages for various PC components, beginning with GPUs in late 2022, before expanding to memory chips by late 2025. This made memory prices climb 500% in 12 months, although the increases have seemingly cooled as consumers refuse to absorb further hikes. Some memory companies like SK hynix are saying that the shortage will be worse next year and that it won’t be until 2030 before pricing will start to normalize. The Adata chief even said that the DRAM shortage would last another 10 years. This is plausibly true, especially as HBM demand from AI hyperscalers remains strong, and the various memory chip fabs under construction aren’t expected to come online until the 2030s.

However, Chen disagreed with this take. He said that the major memory suppliers naturally want to keep their margins as high as possible for longer, “so they keep putting out the message: let me tell you, prices won’t come down until the year 20-whatever.” The Acer chairperson argued that there’s already ample memory and SSD supplies, while the purported CPU shortage is now limited to specific models. He also added that Chinese memory makers churning out cheaper alternatives would disrupt the market — something that SK Group Chairman Chey Tae-won feared. In fact, Acer, alongside HP and Asus, has started using CXMT chips in some of its products, while some Lenovo models sold in Germany were found to have YMTC SSDs.

Nevertheless, he says that price increases are becoming their own trend, saying that SSDs, PCBs, and fiberglass cloth used in motherboards are seeing their own hikes. “A lot of people come and tell us they want to raise prices, and we find it a bit baffling — this needs to go up, too?” Chen said to the reporters.

The overall tenor from Chen is good news, though, if computer manufacturers could finally lower prices after years of shortages and expensive parts.

The memory chip that makes a router remember how to be a router is a world away from the sleek GPUs that are attracting eye-popping investments and alarming valuations, as well as sending stock markets shooting upwards. They’re small, historically have been cheap, and are based on technology that has been around for years.

But despite being a world away from GPUs, the price of these often overlooked chips is skyrocketing, thanks to the all-encompassing memory price crisis caused by the AI boom.

While public and press attention has focused on the expensive chips, there’s an equally large impact beginning to be felt on older, less attractive memory chips. HBM is vital for AI accelerators, while DRAM and high-capacity NAND are being swallowed up by rapidly expanding data centres. A June report from Morgan Stanley reckons memory prices have risen more than sixfold over the last year, breaking with decades in which memory became steadily cheaper as production increased.

It’s not just HBM and DRAM that’s being affected. The crunch is also spreading down into much older forms of memory, including NOR flash and single-level cell, or SLC, NAND. Morgan Stanley expects NOR flash to remain undersupplied through 2026, while JPMorgan has warned its forecasts don’t fully capture a potential supply crunch in SLC NAND. The effects are already showing up in prices. TrendForce says contract prices for both NOR flash and SLC NAND rose by more than 100% during the first half of 2026, while it expects SLC NAND prices to rise another 120% to 170% in the second half of the year compared with the first half.

Picking winners

An increase in prices will have an impact on the tech we use day in, day out. NOR flash is commonly used to store boot and program code, and is a core part of automotive, industrial, and networking equipment. SLC NAND is deployed across a number of uses because of its reliability and endurance when placed in embedded hardware with long lifespans.

Both are vital. And both are being overlooked in favour of higher-margin chips — pushing the supply crunch to tech that previously never faced any issues. “The SLC NAND market is probably under a billion dollars a year,” said Jim Handy, a semiconductor and SSD analyst at Objective Analysis, in an interview with Tom’s Hardware Premium. That tiny scale adds up to a big problem, because it disincentivises any new investment.

“What you've got going on is a purely economic phenomenon,” said Handy. Hyperscalers and cloud providers are “all trying to outspend each other”, pouring unprecedented sums into semiconductors to build AI infrastructure. That willingness to spend big means the most profitable customers naturally move to the front of the queue.

Companies including Nvidia, Broadcom and Marvell need huge amounts of semiconductor manufacturing capacity for chips destined for AI systems. “They’re sucking up all of the wafers,” says Handy. “And then the companies who make NOR flash and SLC are having a hard time getting wafers to build their product, and so they have to raise prices.”

The problem is even starker in the NAND market. Bryan Ao, research manager at TrendForce, told Tom’s Hardware Premium in an interview that major NAND manufacturers, including Micron, Kioxia and SK Hynix, have been cutting the wafer capacity devoted to SLC because they can make considerably more money using it for newer NAND technologies. Ao estimates that a 12-inch wafer devoted to mainstream NAND can ultimately generate close to $20,000 in revenue. Use the same space to produce SLC and the figure is closer to $6,000 to $8,000.

Even if they wanted to, smaller SLC suppliers in China and Taiwan can’t just spin up new production. Lead times for some semiconductor manufacturing equipment have stretched to between 12 and 15 months, said Ao. The result is what he calls “severe undersupply”.

Big prices, big returns

BNP Paribas forecasts the average NAND price will hit $279.50 per terabyte during 2026, up from $73.10 in 2025. JPMorgan expects the memory shortage to persist for at least another two years, with customers getting just 70% to 80% of their orders fulfilled. TrendForce says manufacturers are shifting capacity towards advanced, higher-value memory products, with mature processes increasingly squeezed as a result.

There are some alternatives available. Kioxia says its serial SLC NAND is an alternative to NOR flash. But moving an existing industrial or networking product onto a different chip can itself require engineering work and qualification. Nor is there much incentive for memory manufacturers to fix the problem by building new SLC capacity – which means manufacturers are unlikely to invest billions in capacity whose useful market may disappear. That creates an unusual trap: there may not be enough demand to justify new factories, but there is still more demand than the shrinking supply can satisfy.

Hardware manufacturers can eat the higher component bill and accept lower margins, or pass it on. “We'll just have to either have lower margins, or we'll have to raise the prices to the consumer,” Handy said.

An ongoing issue

The problem is one that seems to have no solution – at least in the short term. Ao expects memory prices to remain high over the next five years and does not expect them to return to 2023 or 2024 levels. That broadly fits with the structural nature of the shortage identified by TrendForce, which says there are no significant capacity expansion plans for NOR flash or SLC NAND.

Handy sees one possible way out, but it is hardly reassuring. “As long as the race between the hyperscalers keeps up to spend, then it will continue to be an issue,” he said.

Handy compares the AI buildout to the internet infrastructure boom of the late 1990s. Rather than enough capacity eventually arriving to restore balance, he thinks spending may simply overshoot what the market can economically support.

“I'm expecting the same kind of a thing to happen here that we've got too many people spending too much money on AI, and not really making any return on it yet,” he said.

Until then, the least exciting memory chips in a computer may become some of the hardest to replace. Or, as Ao put it: “Pretty much we have to get used to this high price, no matter which segment of memory.”

We've talked about the RAMpocalypse extensively, but there's seemingly no end to the continuing rise in memory pricing at the moment. Luckily, Corsair is keeping its store stocked with some gamer-friendly Vengeance RAM, with a 32GB DDR5-6200 set for $419.99, or a slower option for $409.99, offering the best price on these kits in the current market.

● Check out this deal at Corsair

No, we're not talking record-low pricing here, but memory prices aren't a joke right now. AI is pushing them higher and higher, and after $400 for 32GB DDR5 was breached recently, you won't find better-priced DDR5 RAM elsewhere. You'll need a kit like this if you're planning a new gaming PC build or upgrade.

The Corsair Vengeance range continues to be a popular option among gamers, and while we've not given this particular kit a test, the similar spec'd Corsair Vengeance RGB DDR5-6000 C36 performs well. You're getting even faster RAM here, at 6,200 MT/s. It isn't the fastest, but this is priced right in the middle of slower kits that range from 4,800 to 5,600 MT/s, so you're at least getting better bang for your buck.

This particular kit comes with two 16GB modules, with those rated speeds of 6,200 MT/s. It's significantly faster than older DDR4 RAM modules, which typically cap out at 3,600 MT/s. It has CAS latency and memory timings of 36-46-46-100, comes in a black colorway, and has a solid aluminum heat spreader to keep things cool. 6,200 MT/s is a good speed for a gaming RAM kit, and while you can get faster, you'll typically pay more.

That said, every dollar counts when gaming PC builds can go for thousands more than they did last year. You can save $10 and get this DDR5-5600 Corsair Vengeance 32GB kit for $409.99. Like the DDR5-6200 model above, this has two 16GB modules, but with a slower speed of 5,600 MT/s, with CAS latency and memory timings of 40-40-40-77. It also comes with adjustable RGB lighting, a feature missing on the spec above, if you want a more colorful kit. You're probably better off spending the extra $10, but if you're building a mid-tier or budget PC, the speed difference isn't likely to matter as much, but any extra savings might.

The $419.99 price for this 32GB Corsair Vengeance DDR5-6200 RAM is the best price you'll find for RAM with these speeds and capacity, although you can go a little cheaper with the $409.99 DDR5-5600 kit instead. Corsair even has cheaper, and slower RAM, but these two offer the best value for performance that you'd expect for a DDR5 gaming PC build. You'll want to grab this deal while you can before the RAM prices inevitably go up even more.

Micron this week introduced its first 512GB DDR5-9200 memory module that is designed for servers used for applications that demand a lot of memory. The new modules — which are currently being validated by AMD and Intel with their next-generation server platforms — will enable server makers to build machines with up to 12TB of fast memory. What remains to be seen is the price of such modules and servers.

To build its 512GB DDR5 RDIMM, Micron uses advanced packaging that stacks multiple DRAM dies vertically and connects them using through-silicon vias (TSVs). The company does not disclose which memory devices and how many of them it uses, but claims that a single 512GB RDIMM consumes over 60% less operating power than four 128GB modules — based on 16.0W for one 512GB module compared with the 44.2W total for four 128GB modules — which suggests that we are dealing with fairly advanced ICs.

Micron's 512GB module is not the industry's first 512GB DDR5 RDIMM — that achievement belongs to Samsung — but it is certainly the industry's first 512GB module certified to operate with a 9200 MT/s data transfer rate with standard 1.1V voltage (which implies on usage of Micron's 16Gb DDR5-9200 devices made on its 1γ (1-gamma) fabrication process that uses EUV lithography and consumes 20% less power than predecessors, though we are speculating).

Truth to be told, 512GB DDR5 memory modules are rather niche products, which is perhaps why Samsung's 512GB RDIMMs formally introduced in 2021 have not become widespread even after AMD and Intel introduced processors with over 100 cores. Micron positions its 512GB DDR5 RDIMM primarily for servers running analytics, in-memory databases, simulations, virtualization, agentic AI, and other workloads demanding high-core-count processors and plenty of memory. As processors with 200 or more cores emerge, 512GB modules may become more relevant as 12TB of memory in a server running two 256-core CPUs means 24GB per core, which no longer looks particularly excessive for applications like in-memory databases, analytics, or caching.

Micron claims that in memory-constrained Spark Support Vector Machine (SVM) analytics workloads, systems featuring 512GB memory modules can provide up to 1.4 times the performance of configurations equipped with 256GB DDR5 modules, though the company does not disclose how much memory in total these systems use. Micron also says the higher-capacity memory can increase throughput and concurrency for memory-intensive database and caching applications such as RocksDB and Redis.

AMD and Intel are working with Micron to qualify the new modules for their upcoming server platforms. Micron plans to begin volume production of its 512GB DDR5 RDIMMs sometime in the second half of 2027 and intends to align the schedule with customer requirements.

One of the more pressing questions about Micron's 512GB DDR5-9200 memory modules is their price. A 256GB DDR5-6400 RDIMM currently retails for around $19,000. Given the unique proposition that 512GB modules have for memory-constrained applications, such modules can cost significantly more than 256GB memory sticks, which will not help their broad adoption.