Memory shortage driven by AI demand could keep prices high until 2027 or beyond

Report: Samsung execs worried company could lose money on smartphones for the first time

Selling smartphones used to be easy -- everyone wanted one, and every new phone was a lot better than the one that came before. Things are different now that smartphones are mature products. Plenty of manufacturers have thrown in the towel, leaving big players like Samsung to sell a new phone every couple of years. But even Samsung may find it tough to turn a profit in 2026 thanks to the ongoing race to build more AI capacity. According to Money Today (Korean), Samsung MX (mobile experience) head TM Roh has warned company leadership that it could be headed for the first net loss on smartphones in the company's history. Even during times of economic strife or amid pandemic-related supply chain chaos, Samsung still made money on smartphones. The skyrocketing price of DRAM and NAND may be what finally breaks the streak despite strong Galaxy S26 sales. Shortages of these components have affected all types of computing hardware, from consumer laptops to servers. The LPDDR5x memory found in most mobile devices is increasingly important for AI. Nvidia's Vera AI CPU, which will replace Grace later in 2026, will have up to 1.5 TB of LPDDR5x memory. The company's upcoming rack-scale AI platforms will have 36 Vera CPUs alongside 72 Rubin GPUs. The CPUs in one server alone will consume enough RAM for 4,600 Galaxy S26 Ultra devices (12GB each). Until recently, the application processor (which includes the cellular radio) was the most expensive component of most smartphones, and the display usually came in right below that. The AI era has upended the formula, roughly doubling the cost of memory and storage. According to Counterpoint Research, RAM will account for more than a third of the cost of building a budget phone in mid-2026. Even with more expensive devices, memory is still north of 20 percent of the cost. The good news for Samsung is that while the MX division struggles, its semiconductor division is raking it in. Samsung Semiconductor has smashed records in the first quarter of 2026, earning an estimated $38 billion (KRW 57.2 trillion) in profit. That's more than seven times its net from Q1 2025. Samsung, Micron, and SK Hynix are all accelerating plans to expand memory and storage production lines -- Samsung specifically has started spinning down LPDDR4 production to boost the supply of LPDDR5 -- but Nekkei Asia projects that won't be enough. Even with best-case improvements in output, DRAM production in 2027 could fall 40 percent short of expected demand. The only thing that could challenge that prediction is a substantial change in demand for AI applications. With most of the world's tech giants firmly committed to expanding AI compute through next year, it's unlikely that supply constraints will ease soon. Higher demand, higher prices There are already signs that RAM and storage costs are making phones more expensive. Motorola recently raised the price of its Moto G budget phones by up to 50 percent. Low-cost devices like the Moto G will feel the rising cost of components the most, making the very idea of a budget phone in the coming years suspect. With the prospect of sinking profitability in 2026, Samsung is also making changes. The recently released Galaxy A37 and A57 mid-range devices come with a $50 price hike over the last generation. The company has also increased prices on some more expensive devices, adding $80 to the Galaxy Z Flip 7 (512 GB) and Z Fold 7 (512 GB and 1 TB). Some of its tablets are also more spendy, including a $100 increase for the Galaxy Tab S11. With profitability in doubt, Samsung is on the verge of releasing new, ultra-expensive phones. This summer, the company will debut a new generation of Galaxy Z foldables, which are always priced even higher than the Galaxy S series. These devices come with ample storage and RAM to help justify the exorbitant price tags. That could make them prime candidates for price hikes that make them even more unrealistically expensive.

The RAM demand won't resolve in 2026, and 2027 doesn't look great either, claims report

* RAM and storage prices are skyrocketing; don't expect relief anytime soon. * Nikkei reports RAM shortage may not clear until 2027 or later. * Samsung's Pyeongtaek lines won't ease supply until 2027-2028; AI demand keeps pressure. Things look pretty dire in the world of PC hardware. RAM and storage prices have been skyrocketing, and any real chance of either of them coming down requires either more manufacturing lines to get set up or for the AI hype to die down, both of which will likely take years to manifest, if they ever do. Well, one report has given us a peek at what we could be looking at in the future. As it turns out, 2026 doesn't seem like the year we'll see prices go down...but 2027 isn't looking too hot, either. Samsung is reportedly doubling its DDR5 RAM prices as it runs out of stock Shocking, but not surprising Posts 2 By Simon Batt RAM prices may not come down for another year, at least I'm not shocked by the news, to be honest In a report by Nikkei Asia, the publication claims that the RAM shortage may not fully clear up until 2027 at the earliest. Given how the AI industry is showing zero signs of slowing down any time soon, the best bet we have to see relief in RAM prices is with the production lines. If RAM companies can flood the market with more sticks, there will be far more for consumers to purchase, reducing demand. Unfortunately, even the big leagues cannot magically whip up a new factory overnight. As Nikkei Asia claims: Samsung Electronics, the world's leading memory chip maker, plans to bring the fourth fabrication plant at its Pyeongtaek campus in South Korea online in 2026. But full-scale mass production will not begin until 2027 or later. The facility also will make logic chips for computing, limiting its capacity to increase production of memory. Nikkei Asia also reports that Samsung is creating a fifth production line in Pyeongtaek. This one will specifically produce RAM designed to handle AI tasks, which should alleviate the strain on the PC market for hardware. The bad news is, this one will take even longer to come online, with Nikkei Asia claiming it won't arrive "until 2028 or later." If there's any good news to take from this, it's that memory companies are taking steps to help bring RAM supply in line with demand. It's just that it's going to take a while for that to happen. Memory suppliers are playing favorites with RAM, and only 4 PC makers are getting priority This strategy is squeezing smaller PC makers in a big way, says a recent report. Posts 3 By Patrick O'Rourke

Chipmakers on track to meet only 60% of AI memory demand by 2027

Serving tech enthusiasts for over 25 years. TechSpot means tech analysis and advice you can trust. AI Economy: Big Tech players and hyperscalers are buying up virtually every memory chip they can secure, and even that isn't enough to satisfy demand driven by AI. The conversation has shifted from short-term shortages to a more structural question: how long it will take manufacturers to catch up. The answer, increasingly, looks like years. A recent report from Nikkei Asia yet again highlights how tight the market has become. It centers on the ongoing memory shortage, which is expected to persist until chipmakers have both the capacity and the facilities to meet demand for AI-focused components from hyperscale customers. The industry's largest players, including South Korean giants Samsung and SK Hynix, alongside US firm Micron, are ramping up investment in new fabs. Even so, analysts estimate that by the end of 2027, supply will meet only about 60% of demand for high-bandwidth memory products. SK Hynix recently opened a new fabrication plant in Cheongju, but broader capacity expansions across foundries and integrated device manufacturers are not expected to come online until 2027 or 2028. According to Nikkei, manufacturers would need to increase annual output by roughly 12% over the next two years to close the gap. Research from Counterpoint suggests they will fall well short, projecting growth closer to 7.5%. The longer view is even more sobering. SK Group chairman Chey Tae-won recently said he expects the shortage to last through 2030. In the meantime, the entire sector has pivoted toward high-bandwidth memory, which is critical for AI accelerators and advanced GPU workloads. SK Hynix alone holds about 32% of the global DRAM market and more than half of the HBM segment. That focus leaves less room for traditional consumer memory. Chips used in PCs and mobile devices are no longer the priority, and in some cases, they are being sidelined altogether. Micron famously shut down its Crucial brand for good, while other vendors are simply trying to squeeze more money out of their customers as long as the AI craze lasts. There are early signs of downstream effects. Some system integrators and OEMs are beginning to ship consumer hardware with reduced performance targets, reflecting tighter component availability. The message from hardware makers, implicitly or not, is that constrained tradeoffs are part of the current moment and we should just be happy for the effort in these trying times.

AI's chip hunger could keep memory prices painfully high for years

Memory shortages may haunt your next phone, laptop, and GPU for years While recent reports claimed that memory prices may not fall till 2027, it seems like the memory chip crunch isn't a short-term headache. And that's bad news for anyone hoping phone, laptop, and GPU prices will get cheaper again soon. Reuters reports that SK Group chairman Chey Tae-won said the global chip wafer shortage is likely to last until 2030, with artificial intelligence demand continuing to outpace the supply. Chey said the current shortage could remain above 20%, largely because AI systems require huge amounts of high-bandwidth memory and therefore burn through a lot of wafers. Why memory price hikes could stick around for a few years Memory chips didn't suddenly get expensive over some production shortage or artificial scalping/price gouging. Chey specifically pointed to AI's massive appetite for HBMs, or high-bandwidth memory, as a key reason the crunch is still sticking around. He believes that the industry needs at least four to five years to build up enough additional wafer capacity, which is why he believes that the shortage could stretch to the end of the decade. Recommended Videos Memory is not some niche component buried in the supply chain. It touches almost everything consumers buy, from budget phones and mid-range laptops to gaming handhelds, consoles, SSDs, and graphics cards. Why your wallet should care Chey expects SK Hynix's leadership to unveil a plan aimed at stabilizing DRAM prices. Companies generally do not talk about price stabilization unless they are worried about volatility, and that usually implies less predictability and higher prices for buyers. With SK Hynix holding a 57% share of the HBM market and 32% of the global DRAM market, it's clear that the company is not a minor player. Being the second-largest DRAM supplier in the world, these warnings are hard to shrug off.

Global memory shortage expected to get worse before it gets better

TL;DR: DRAM manufacturers are projected to fulfill only 60% of the global demand by the end of 2027, indicating a continuing shortage in the DRAM market. The global memory crisis doesn't appear to be easing anytime soon, with new reports suggesting DRAM shortages could stretch well into the end of the decade, which coincides with what I heard from memory manufacturer insiders at CES 2026. According to Nikkei Asia, DRAM manufacturers are only expected to meet around 60% of global demand by the end of 2027, despite the numerous aggressive expansion plans for DRAM production. The big three, including Samsung, SK Hynix, and Micron, are investing heavily in new fabrication facilities, but most of this additional capacity won't come online until 2027 or later, leaving a multi-year supply gap. At the heart of the issue is the explosive growth of AI infrastructure. High-bandwidth memory (HBM), critical to AI accelerators and data centers, has become a top priority for manufacturers. As a result, traditional DRAM used in consumer devices, such as PCs, smartphones, and gaming handhelds, is being deprioritized because it isn't yielding as high a profit. Production of legacy standards like DDR3, DDR4, and LPDDR4 is already being scaled back, further tightening memory supply. The numbers paint a grim picture. Industry analysts estimate production needs to grow by around 12% annually through 2027 to meet demand, but current projections are closer to 7.5%, indicating a looming shortfall and tightening of supply. Meanwhile, AI companies are reportedly pre-booking large portions of future supply, limiting availability across consumer markets and driving prices even higher. Even with new fabs under construction, including efforts by Chinese manufacturers, relief seems far off, with some industry leaders warning shortages could persist until 2030. For us, everyday consumers, that means continued inflated pricing on memory, which impacts a wide range of common devices, with no clear return to normal anytime soon.

AI-Driven Global Memory Shortage Might Not End Until 2030

Samsung, SK Hynix are reportedly also building new HBM fabrication plants The ongoing global memory shortage has shaken up the consumer tech market, with the rising component costs driving up the selling prices of devices such as smartphones, laptops, gaming consoles, and more. A new report now claims that it could take until 2030 before the supply matches the demand. One of the biggest DRAM suppliers globally, SK Hynix, is reportedly planning to bring up a new fabrication plant, but it is said that the focus would be on high-bandwidth memory (HBM), which is used by artificial intelligence (AI) data centres, and not consumer devices. Memory Shortage Could Last Years SK Hynix's Group Chairman, Chey Tae-won, told Reuters that the ongoing crisis might not end before the end of the decade. The company, which holds a 32 percent share of the global DRAM market (second only to Samsung) and leads the HBM market with a 57 percent share, is reportedly struggling to meet the demand from AI players for their data centres. Chey reportedly said that to meet the current demand of HBM, the company will need four to five years of time. This is primarily to stock up enough raw materials to produce the required silicon wafers. However, this production drive will only benefit the rising global AI inference demands, and not the consumer tech devices. In fact, the executive believes the shift in focus could lead to more than a 20 percent of shortage of components for the retail market. A new Nikkei Asia report has a more optimistic outlook on when the shortage could end. As per the report, the current state of rising prices and stock depletion could continue until 2027 with efforts from the three global leaders in the space -- Samsung, SK Hynix, and Micron. Samsung is reportedly planning to add a fourth fabrication plant at its Pyeongtaek campus in South Korea this year, which will not become fully operational until the end of 2027. However, the facility is said to also focus on building logic chips for computing, which can impact the total memory production. The tech giant is also said to be constructing a fifth fabrication plant, but it is said to be focused on HBM. Similarly, both SK Hynix and Micron are reportedly expanding fabrication of HBM, with SK's Cheongju facility already in operation and Micron planning mass production by 2027 in its Idhao and Singapore plants. But these are also not expected to impact the DRAM shortage significantly. A Counterpoint report claims that ending the ongoing DRAM shortage would require the suppliers to increase production by 12 percent per year; however, current plans only point to an expansion of 7.5 percent, which will only marginally solve the problem. Meanwhile, the consumer tech market is dealing with the consequences of this shortage. A report last month claimed that the cost of a 16GB LPDDR5X RAM and a 1TB UFS 4.1 storage has surpassed the cost of Qualcomm's Snapdragon 8 Elite Gen 5 chip. The cascading effect has impacted many consumer devices as well. Earlier this month, Sony raised the prices of the PlayStation 5 and the PlayStation 5 Pro consoles. The ROG Xbox Ally X also witnessed a price hike in Japan recently. Reports have also claimed that Apple might be paying more for the memory components, which can result in the prices of this year's devices. Additionally, it is said that Vivo and iQOO phones in China could soon witness a price hike. If the shortage persists, other consumer tech brands are also likely to follow suit.

Samsung's Memory Division Cares More About AI Than Its Smartphone Business, As Executive Warns Of A Deficit Due To Rising Costs

The AI industry has grown exponentially, but at the cost of gobbling up nearly all of the DRAM supply, leaving next to nothing for a multitude of manufacturers hailing from other businesses. Memory makers are also struggling to keep up with demand and may only meet 60 percent of the entire supply through 2027. Samsung, being one of the firms to mass produce DRAM chips, appears to favor its short-term growth and profits over lending a hand to its DX (Device Experience) and MX (Mobile Experience) units, which are responsible for smartphone launches. According to the latest report, the situation is dire to the extent that the executive of both divisions, TM Roh, has warned that a possible annual deficit could materialize for the MX sector. If the Samsung executive's prediction comes true, it will be the first time the company's MX division has witnessed an annual deficit since its inception Soaring DRAM and NAND flash prices have already forced Samsung to introduce a price hike to its flagship smartphone series, the Galaxy S26, leading to a reduction in consumer spending. The Korean manufacturer has also reportedly halted production of LPDDR4 and LPDDR4X memory chips in favor of faster, more efficient, and more expensive LPDDR5 and LPDDR5X RAM for AI customers. Despite LPDDR being a primary component in mobile devices, it has emerged as a core piece of the puzzle for the AI market, resulting in a massive shortage. For instance, NVIDIA's next-generation AI CPU codenamed 'Vera' is said to be equipped with a total of 1.5TB of LPDDR5X memory. Considering that a single Galaxy S26 Ultra flagship unit ships with 12GB of LPDDR5X RAM, the cutting-edge AI processor's memory requirement is 125 times larger, ultimately causing a shortage. Samsung also had the liberty to prioritize its MX business unit by manufacturing cheaper DRAM chips for the Galaxy S26 series, but it appears that two business segments belonging to the same company don't necessarily mean that they will work in favor of each other. Memory supply contracts are typically inked on a quarterly basis, and price increases that have continued since the second half of last year are expected to be fully reflected in manufacturing costs from the second quarter of 2026. Counterpoint Research has also predicted that premium devices priced at $800 or more will feature DRAM whose cost will account for 20 percent of the total. In short, Samsung's MX division will likely incur a loss before entering the recovery phase. News Source: Money Today Follow Wccftech on Google to get more of our news coverage in your feeds.

Memory Makers Will Only Meet 60% of DRAM Demand Through 2027, Locking In Years of Shortages and Price Pain

Memory shortages are expected to last several years, with no sign of stabilization anytime soon, as DRAM makers fail to meet demand. Memory supply constraints continue to grip the industry as the rise of Agentic AI has led to severe shortages. The demand is so massive that it will take years before we see any hint of relief. The latest report from Nikkei Asia seems to agree with this matter. It is stated that DRAM manufacturers are expected to meet only 60% percent of global demand by the end of 2027. As datacenters continue to pop and gobble precious DRAM, we will continue to see higher prices, and little to no memory available in the retail channel. A shortage of memory chips appears likely to continue until around 2027, with the top U.S. and South Korean suppliers raising DRAM production at a pace that will meet only about 60% of demand. via Nikkei Asia To address this matter, major DRAM manufacturers such as Samsung, SK Hynix, Micron, and China-based YMTC are bringing up extra production lines and starting rapid construction of new facilities to meet the demand, but these factories won't be ready anytime soon. Just yesterday, we saw that YMTC and CXMT were working on three new fabs, one going live already, and two planned to go live later this year, which will double their total output. But despite all of this, AI factories are pre-booking entire year's worth of supply, leading to severe shortages in the smartphone and PC segments. And the majority of the new production lines will be focused on AI-specific DRAM such as HBM. Samsung and Co. have already dropped production of legacy DRAM such as DDR3, DDR4, and LPDDR4, further straining the market. Several companies have discontinued their OEM / PC-specific brands, Micron's Crucial is one example, and moved to more profitable pastures such as HBM and SOCAMM2. Chinese manufacturers have stepped into play to fill this gap. As per Nikkei's report, DRAM manufacturers will have to increase their annual production by 12% between 2026 and 2027 to address this growing demand, but Counterpoint states that at the current rate, the production is only increasing by a measly 7.5%. Counterpoint also shared its stats on how bad it is for PC vendors going forward, with memory prices continuing to climb. Remember that we had already pointed out in a previous report that memory prices won't be back to normal till at least the end of 2028, and that is looking to be the case every passing day.