RAM

DDR4 was supposed to be the safe fallback once DDR5 pricing went off the rails. That is no longer true. According to a new DigiTimes report, Taiwanese memory makers are quoting DDR4 8Gb contract prices for Q3 2026 up to 50% higher than Q2, blowing past what the market had already priced in as a worst-case scenario.

This time, the driver is not PC demand but, surprisingly, enterprise SSDs. High-capacity eSSDs need standalone DRAM chips to handle random reads and writes, and as data centers push toward 16TB to 30TB drives for AI workloads, that DRAM requirement has scaled up with them.

Samsung, SK hynix, and Micron have mostly wound down standard DDR4 production to chase DDR5 and HBM margins, though Micron reportedly restarted 1-alpha DDR4 and LPDDR4 output at its Manassas, Virginia fab, aimed mainly at automotive, defense, and medical customers rather than PC builders. With the other suppliers unable to cover the shortfall, the supply gap could reportedly stretch out for as long as two years.

Lexar's lineup of THOR DDR5 memory for desktops, compatible with AMD and Intel systems, is built for high-performance and PC gaming. With vibrant RGB lighting, aluminum heatsinks, and support for Intel XMP and AMD EXPO overclocking, these are the sort of modules and kits that you'd install in a high-end gaming rig.

According to a new report over at IT Home (via TechPowerUp), Lexar's upcoming THOR II DDR memory kits will be built using CXMT (ChangXin Memory Technologies) chips. Now, the use of memory chips from Chinese companies is making its way to global markets, which is not uncommon in 2026 thanks to the current memory crisis, but this move from Lexar is notable because it highlights the progress CXMT has made with its memory.

The report notes that these new Lexar THOR II DDR5 memory kits will arrive in 32GB dual-channel form, with two 16GB modules, supporting speeds up to DDR5-7600 (7,600 MT/s).

AMD introduced EXPO Ultra Low Latency at Computex 2026 as an updated memory profile standard, promising tighter timings and around 4% better gaming performance over standard EXPO. AMD also suggested pricing would stay close to existing EXPO kits. But the first retail listings from G.Skill show that unlocking those gains will cost you.

G.Skill's Trident Z5 NeoX RGB series has appeared on Newegg in 32GB DDR5-6000 configurations with EXPO ULL support. The premium over the standard Trident Z5 Neo kits ranges from 10% to as high as 79%, depending on the timing tier.

The C26 NeoX kit is listed at $1,099.99, compared to $699.99 for the standard C26, translating to about a 57% premium. The C28 NeoX is $999.99, compared to $559.99 for the standard version, which amounts to a 79% premium for a kit with two fewer CAS cycles. The gap narrows considerably at looser timings, with the C30 NeoX at $619.99 running 14% above the standard C30, and the C36 NeoX at $549.99 sitting just 10% higher. In other words, the price jumps are steepest where the latency gains are most aggressive.

Samsung Electronics, SK Hynix, and Micron Technology are being sued in federal court for alleged collusion and price-fixing in the memory market. The lawsuit was filed in the Northern District of California, alleging that these three companies have deliberately engineered DRAM scarcity in the consumer and commercial markets to inflate prices.

If you've been following recent news about Apple, Xbox, and other companies increasing the cost of their consumer-facing technology products due to the rapid, exponential rise in memory (and flash storage) prices, you might be nodding in agreement. The current situation, born of the memory crisis, is that consumer technology pricing will continue to rise with no real end in sight. Lenovo is calling it the "new normal" and that this trend is set to continue through 2030.

As the three big players in the memory space, Samsung Electronics, SK Hynix, and Micron Technology have all transitioned their manufacturing and production capacities to producing High-Bandwidth Memory (HBM), specifically for the booming AI and data center markets. And with that, the lawsuit claims that production of legacy technologies like DDR3 and DDR4 has been wound down to pursue higher profits. And with demand for commodity DRAM still there, prices have increased by around 700% over four years. The lawsuit even references the name many have given the situation: the "RAMpocalypse."

As home to two of the world's largest memory makers, Samsung and SK Hynix, the South Korean economy is definitely benefiting from the current memory crisis, which has seen prices and profits hit unforeseen levels. Long story short, in the AI era of LLMs, agents, inference, and cloud-based computing, the demand for high-bandwidth memory (HBM) chips has reached a point where supply would take years to catch up.

And with that, South Korean President Lee Jae Myung has announced a $520 billion USD public and private investment to expand the chipmaking capabilities of Samsung and SK Hynix so the country can remain a leader in the AI race and the memory market. And the announcement was made in a televised address with Samsung Electronics Chairman Lee Jae-yong and SK Group Chairman Chey Tae-won.

According to the announcement, the majority of the $520 billion will be put toward the construction of four new fabs, two for Samsung and two for SK Hynix. The two plants will reportedly be built outside Seoul, in a mostly rural area. They will focus on producing lucrative HBM for the AI and data center markets rather than memory for the consumer technology market.

Last week, Apple raised prices across its entire hardware line. It was a decision born of the steep increase in component costs, specifically memory and storage, made because the situation had become "untenable" and it was no longer possible to shield its customers from rising manufacturing and procurement costs. The MacBook Pro saw one of the steepest price increases, by $300, specifically because of its advanced processor, memory, and storage configuration.

The situation has reached the point where Apple is reportedly lobbying the US government and the Trump administration to allow it to buy memory chips from the Chinese company CXMT without ramifications. As for why it's lobbying the government, well, CXMT is currently on a list of companies that US-based tech giants like Apple shouldn't engage with due to alleged ties to the Chinese military.

Apparently, this Pentagon list doesn't have any legal ramifications: companies aren't explicitly banned or in legal trouble for procuring memory from a company like CXMT. However, CXMT is apparently also on another US Commerce Department list for a trade blacklist, which would certainly prohibit any such agreement with Apple for memory.

Although this is an issue that has been developing for over a year, the current memory crisis feels like it's accelerating (in the wrong direction) on account of existing stockpiles drying up and the fact that every consumer tech company, big or small, is now paying exponentially more than they were a year ago. This is why we've been seeing a string of recent price increase announcements from companies like Apple and Microsoft, and will probably see several more as the months roll on.

Throw in comments from big players like Lenovo saying that there's basically no end in sight, and we've now got even more bad news to throw into the mix. A new report from Jefferies Equity Research reinforces the bleak outlook, stating that memory prices will increase by another 40-50% in Q3 2026. And that's 40-50% on top of the current inflated prices.

Of course, it doesn't stop there. The research and analyst firm is also predicting that in Q4 2026, pricing will increase by another 30-40%. And with that, it's not hard to see why Apple's MacBook Pro is getting a $300 price increase, and the almost six-year-old Xbox Series X console is getting its price increased by $150. And to make the situation feel that extra bit existential, 2027 will also see price increases of around 40-45%.

At the ISC 2026 conference in Hamburg, Lenovo executive director Martin Hiegl delivered what was intended as a joke but landed like a gut punch. When asked about the memory price outlook, he told the room that prices will "never" come back down to where they were before the crisis. The laughter that followed from the audience was probably nervous.

To be fair, Lenovo and ComputerBase (who first reported the comments) were clear that "never" is relative. Hiegl was really talking about the next five or more years, with prices expected to remain structurally higher through 2030 and beyond. Still, the underlying analysis is dead serious, and that leaves us worried.

According to Lenovo's presentation, DRAM and NAND began their rapid climb in late Q3 2025, and have since hit levels the industry itself did not see coming. Even with continued capacity expansion by major manufacturers, prices are highly unlikely to return to early-2025 levels.

The RAMpocalypse reached a new milestone in Q2 2026. According to market research firm SigmaIntel, consumer memory prices rose sharply in the second quarter as supply-demand imbalances persisted across the industry. These are quarter-over-quarter figures, meaning prices didn't just climb versus last year; they also jumped relative to the already elevated Q1 2026 levels.

The LPDDR segment took the worst hit. LPDDR4X 4GB ICs rose 75%, going from $26.2 to $45.9, while 96Gb (12GB) LPDDR5X modules surged 89%, climbing from $77.1 to a whopping $145.9. This is the single biggest price jump in DRAM across any segment.

On the standard DDR side, a 16GB DDR4 stick now costs $207.1 versus last quarter's $137, a 51% increase, while a basic 16Gb (2GB) DDR module jumped from $19.2 to $28.5, up 49%. It is important to note that DDR5 pricing wasn't even included in SigmaIntel's report, and we've already documented how brutally expensive DDR5 has become.

The AI industry boom is undoubtedly unprecedented, and consumers are feeling the pinch of supply constraints on memory and storage prices. This problem is so widespread that it isn't just consumers who are impacted by higher prices; as new Xbox CEO Asha Sharma recently outlined in a letter, the price Xbox pays for storage and memory has increased by 4x since she took office.

With DRAM manufacturers now turning toward AI-specific memory, such as HBM production, we are beginning to hear rumblings of a race between leading DRAM manufacturers, SK Hynix and Samsung. Reports indicate that both companies are rushing to release the first HBM4E memory solutions, which will undoubtedly be gobbled up by the highest-bidding AI companies and deployed directly in data centers.

HBM4E memory will be a big part of the next generation of data centers, which powers the next generation of AI. The new generation of memory chips will be headed to leading products such as NVIDIA's Rubin Ultra, and AMD's Instinct MI500. Just these two chips, by themselves, are expected to drive millions in sales. As for specifications, it's reported the new HBM4E offering from SK Hynix will arrive in capacities of up to 48GB, and speeds of 16GBps. As for Samsung, the South Korean manufacturer showcased its latest generation, HBM5, on the showroom floor at Computex 2026.

MSI joins GIGABYTE, ASUS, and ASRock as the fourth manufacturer to add HUDIMM support to its DDR5 Intel motherboard lineup. MSI announced in a blog post that it has rolled out motherboard updates for its entire Intel 600, 700, and 800 series motherboard lineups, adding support for the budget-friendly memory form factor. If you own a compatible board, you can go to the product page for your MSI motherboard to download the appropriate BIOS update.

HUDIMM is a new memory form factor that was made in collaboration with ASRock, Intel and TeamGroup to counter the skyrocketing prices of DRAM thanks to the AI boom. These budget-focused modules are cut-down versions of traditional DDR5 DIMMs, featuring just a single sub-channel, cutting memory bandwidth in half compared to regular DDR5 DIMMs. This bandwidth reduction is so significant that DDR4 memory can outperform HUDIMM DDR5 - depending on memory speed. Despite this, synthetic testing has shown that gaming performance could be largely unaffected by HUDIMM's bandwidth-limiting single-channel configuration, especially when two memory sticks are present and running in a dual-channel configuration.

Motherboards supporting HUDIMM memory can mix and match the modules with traditional DDR5 DIMMs. This is beneficial in specific cases where buying mixed DDR5 memory setups consisting of HUDIMMs and traditional DIMMs is a more cost-effective and/or performant solution compared to buying just DDR5 DIMMs. For example, pairing a single 16GB DIMM with an 8GB HUDIMM will yield better performance than buying a single 24GB DIMM. Mixing and matching is also highly beneficial for existing setups where users want to upgrade their DDR5 setups with HUDIMM memory.

The current memory crisis means that one of the only real ways to ease the burden on the consumer market and meet the insatiable demand from the AI and data center sectors is for key players to ramp up production of existing and next-gen memory chips. As one of the biggest players in the space, SK hynix is looking to do just that and, in the process, triple its wafer production.

Although the construction of four new facilities to, well, facilitate this won't be completed and online until 2034, this is actually a shorter timeline than the previous target set by SK hynix of 2045. In a recent interview with Nikkei Asia (via German outlet ComputerBase), SK Group Chairman Chey Tae-won said, "Since we're proceeding with the plan to expand as much as possible, our calculations show that our wafer capacity will double within five years. But honestly, once all these facilities are built, it won't just double, it will triple by around 2034."

According to the chairman, the current 2034 target is the best-case scenario and that there's no way to "move faster than this." Due to the ongoing demand for DRAM and NAND Flash, prices have quadrupled over the past year as supply struggles to meet demand. And with that in mind, this news is welcome for the long-term supply of memory, but it doesn't address the current crisis.

It looks like DDR5 overclocking records are falling fast at Computex 2026, and GIGABYTE just claimed the biggest one yet. GIGABYTE's in-house overclocking team hit DDR5-13556 MT/s at Computex 2026, setting a new world record for memory overclocking. The run used Corsair Vengeance DDR5 memory paired with the Z890 AORUS TACHYON DUO X ICE motherboard.

The team was led by well-known overclocker Hicookie, alongside teammates Sergmann, Saltycroissant, Madness777, and Exaberries. That same team was also invited to G.SKILL's 12th Annual OC World Record Stage at Computex, where they took 10 global first places in CPU frequency and other benchmark categories, this time using the newly launched X870 AORUS INFINITY motherboard with AMD Ryzen 9 9950X3D2 processors.

The Z890 AORUS TACHYON DUO X ICE is a two-DIMM board purpose-built for extreme memory overclocking. It uses GIGABYTE's CQDIMM technology, which is designed to improve signal integrity and push memory performance well beyond what standard platform configurations can achieve. The two-DIMM layout is key here, as fewer slots mean less electrical load on the memory controller, which is how these extreme numbers become possible.

KLEVV, one of the leading memory companies, has showcased a selection of new memory and storage products at Computex 2026, demonstrating high speeds and high capacities.

At their booth, KLEVV walked me through each of the new memory kits, starting with the CRAS V RGB Prime. The company's very first kit adds an AMD EXPO-ULL (Ultra Low-Latency) profile. According to KLEVV, the new high-end memory kit offers DDR5-6000 speeds with a CAS latency of 26. KLEVV showed a 32GB kit that was running on a Ryzen 9 9950X3D. Moving on, we have the CRAS Va RGB kit, which KLEVV explained is its high-performance kit as it features DDR5-10000 speeds.

Next is the LITE V RGB, a new memory kit that comes in capacities up to 64GB and speeds up to DDR5-6400. Additionally, the heat spreader features a nice RGB crown for some aesthetic flair. However, what was easily the most impressive kit KLEVV showcased was a 256GB (128GB x 2) kit featuring stunning 4R CUDIMMs.

Patriot Memory came to Computex 2026 with some new products including DRAM, SSDs for handheld devices, and some external storage solutions. But rather than try to dazzle with attention-grabbing high-end kits or impressive overclocks, Patriot highlighted a more mainstream kit.

The Viper Steel 5 Infinite RGB memory kits range from 6000 MT/s to 8000 MT/s speeds, with capacities of 2x16GB, 2x24GB and 2x32GB. These are regular DIMMs, and not the more expensive CUDIMM variety.

The kit looks really lovely, with the individual LEDs barely visible. It creates a true infinity effect.

At Computex 2026, AMD announced its new AMD EXPO Ultra Low Latency (ULL) technology for automated DDR5 memory overclocking that it says is specifically designed to boost gaming performance. Although AMD says that gamers leveraging EXPO Ultra Low Latency can expect to see a 4% increase in FPS compared to non-AMD EXPO ULL memory, in the era of "every frame matters," it's still enough to make an impact, as there's also an improvement to the all-important 1% lows.

Yes, as noted above about non-AMD EXPO ULL-compatible memory, to take advantage of the new technology, you will need to use compatible memory kits. The good news on that front is that AMD EXPO ULL memory kits will be available starting this month, in June 2026, from its certified memory partners, including G.SKILL, Kingston, KLEVV, Lexar, TeamGroup, V-Color, and XPG.

Digging a little deeper into the expected performance gains, AMD notes that when compared to standard JEDEC non-overclocked memory, you can expect a 13% improvement to the average FPS, and a 4% improvement to the average FPS when compared to standard AMD EXPO memory. And it's a similar situation with the 1% numbers, with AMD confirming it obtained its figures by benchmarking 30+ games on a Ryzen 7 9700X CPU.

GIGABYTE has introduced D5 Single Boost, a new BIOS-level feature for its Z890 and B860 motherboards that automatically overclocks a single DDR5 module to up to 8,400 MT/s. Running a single DDR5 module has always come with a bandwidth penalty. With only one module installed, the memory controller has fewer sub-channels to work with, which means lower overall throughput regardless of clock speed.

GIGABYTE is trying to close that gap as much as possible for builders who are starting with a single stick and plan to add another later, or for those who would rather spend their budget on a better CPU or GPU. That is especially useful amid the ongoing "RAMapocalypse" and rising memory costs, which have made getting into PC building more difficult than ever.

D5 Single Boost automatically detects Hynix M-Die memory chips as soon as a compatible module is installed. From there, an Adaptive BIOS Tuning Engine profiles the module in real time and applies precision timing and voltage adjustments calibrated to that specific die, with no manual configuration required. The settings persist across every boot without any user input.

G.Skill is back with another showcase of high-end DDR5 performance, and this time the company is turning heads ahead of Computex 2026. The memory maker has unveiled a Trident Z5 CK CUDIMM DDR5 kit running at a blistering 9200 MT/s, and the headline spec is not just the speed itself but how it gets there.

The kit is a 32GB (2x16GB) CU-DIMM configuration running at DDR5-9200 with CL74-74-74-148 timings, and it does all of that at just 1.1V DRAM voltage. That last part is what makes this announcement notable. Running at the standard JEDEC voltage figure of 1.1V at this speed is a serious engineering achievement. High-frequency DDR5 kits typically require higher voltages to remain stable, resulting in higher power draw and more heat. G.Skill is managing to sidestep both of those trade-offs at 9200 MT/s.

So how does CUDIMM factor into this? CU-DIMM is a newer DDR5 variant that integrates a built-in Clock Driver (CKD) chip directly into the memory module. The CKD chip strengthens the clock signal traveling from the memory controller to the DRAM chips, improving signal integrity and overall stability during high-frequency operation. Without the CKD on board, sustaining these kinds of speeds cleanly would be a much harder problem to solve.

The memory crisis may finally be on the verge of ending by the second half of 2027, driven by a surge in Chinese manufacturing capacity.

That's at least according to former Samsung Electronics semiconductor division president Kyung Kye-hyun, who spoke at the National Academy of Engineering in Seoul. Kyung pointed to a coming wave of new memory production from China that could alleviate the ongoing shortage and stabilize pricing.

Kyung emphasized that the increase in Chinese manufacturing is expected to flood the market with additional memory chips, potentially lowering prices for consumers. This comes as Samsung and SK Hynix continue to ramp up production to meet surging demand, particularly from AI applications.

ASUS has officially thrown its hat into the desktop memory ring. After months of speculation, the company finally confirmed its entry into the enthusiast memory market with the launch of its very first Republic of Gamers-branded kit, the ROG DDR5 RGB Edition 20. The announcement landed during the ROG 20th Anniversary event in China, which makes sense given the module's gold, red, black, and silver anniversary color scheme.

The new ROG DDR5 RGB Edition 20 was co-developed with BIWIN, one of ASUS' ROG Certified memory partners. It comes in a single configuration for now: two 24GB modules for a total of 48GB, built on SK Hynix M-Die ICs. ASUS rates the kit at DDR5-6000 with CL26-36-36-76 timings at 1.45V, and it supports both Intel XMP and AMD EXPO profiles, so it should work out of the box across current Intel and AMD platforms.

The more interesting addition, however, is what ASUS calls "ROG Mode." This is a dual-profile system exclusive to ROG motherboards, including the Crosshair, Maximus, and Strix lineups. It lets users switch between a low-latency 6000 MT/s mode tuned for gaming and a high-bandwidth 8000 MT/s mode running at CL36-48-48-110 at 1.40V. The lower-latency option is the better pick for most gamers, while the 8000 MT/s mode trades tighter timings for raw memory bandwidth.