RAM

In the middle of the RAM crisis, your cheapest computer was a bargain too good to last

by

If there is a product in Apple’s portfolio that was a real candy, it was the Mac Mini. This has been a reality for years, but in these times that are even more so: the Mac Mini M4 It came to the market with the power of the M4 chip, 16 GB of base RAM, a 256 GB SSD (the most stingy, Apple style) and a RRP of 719 euros, which in practice was much less. I bought it myself for less than 600 euros. Well, that bargain has come to an end: in the context of current shortages, the 256GB Mac Mini is no longer an option. We had already seen it with its models with more RAMbut this decision is dramatic for the general public. Goodbye to the 256 GB Mac Mini. Apple has made a decision that directly affects the pockets of those who want to buy the Mac mini in its most basic version. Since yesterday, May 1, 2026, Cupertino has removed that entry model from its catalog, as Joe Rossignol advances for MacRumors. It is not that it appears out of stock, it is that it has directly disappeared, as can be seen on the Spanish website. Of course, there is still stock and offers of the old base model in stores like amazon, at Media Markt either in El Corte Inglés. The entry price of the Apple desktop computer starts at 969 euros and corresponds to the version with M4 chip, 16 GB of RAM and 512GB SSD. In the United States the jump has been from 599 dollars to 799 dollars, in Spain it has gone from 719 euros to 969 euros. The versions with the M4 Pro processor remain as they are. This decision is framed within a structural RAM supply crisis and whose main culprit is the voracity of the AI ​​infrastructure. Prices and delivery for the Mac Mini. Apple Why is it important. Raising the entry price of one of its star products by almost 35% (in the United States it is 33) more is an aggressive move that has implications for both the individual consumer and the technology market in general. It is true that technically speaking Apple has not raised its prices, it has simply eliminated the lower step, leaving orphans those people who considered that base version sufficient, which are quite a few: it is my main computer for mixed tasks, basic editing, office automation and the Internet and the performance is more than good. In short: for many users, students or professionals, with tight budgets, this increase of more than 200 euros is a real chore. The problem is not just the price: the impact is worsened by delivery times. I have tried different Apple Stores and shipping is delayed until the end of May or beginning of June. Context. Tim Cook gave an explanation during the conference results for April 30, 2026 recognizing that the supply of Mac mini and Mac Studio is severely restricted and that normalization could take months. The reason given by the still CEO of Apple is that both devices have become popular platforms for artificial intelligence and agentic tools, which has triggered demand above forecasts. And he anticipated something: Apple will face significantly higher memory costs in the current quarter, according to MacRumors. This places the Mac Mini in a paradox: that the configuration of this compact desktop computer makes it ideal for working with AI locally and that precisely this reality is what has exhausted the stock, forcing Apple to cut its catalog. The AMR crisis continues to claim victims. In March of this year Apple already removed the 512 GB RAM option from the Mac Studio and in April several models of the Mac mini and Mac Studio they directly stopped being able to order in the Apple Store in the United States, with delays of up to five months for versions with more RAM. The memory chip supply crisis is not something exclusive to Apple, but a trend that crosses the entire sector and caused by the demand of the hyperscalers. Apple needs to ensure that every machine sold is capable of fluidly running its new digital agents and AI tools, making lower memory and storage configurations no longer viable or cost-effective under the company’s current standards. The particular thing about Apple’s decision is the timing: just when it launches its best chips for local AI processing, the global RAM market is strained to unsuspected limits precisely because of that fever. The result is paid by the final consumer. In Xataka | Not even Apple is free from the new reality of the technology industry: RAM goes first for hyperscalers In Xataka | The RAM crisis was supposed to make computers and smartphones very expensive. Apple has another opinion Cover | Apple and Alberto García

BYD promised them very happy by putting very advanced ADAS in very cheap cars. Until the RAM crisis came

by

In recent years, BYD had turned its brand new advanced driving system into one of the biggest arguments to confront Tesla. And having this type of technology in affordable cars can be attractive to the consumer, but it has a cost that other companies can hardly absorb. BYD thought so, but the RAM crisis It has stopped him, and the context is now much more complicated. Prices go up. BYD just announced in China a 21% increase in the price of the ‘DiPilot 300’ option (basically its “God’s Eye” in its version with LiDAR), which goes from 9,900 to 12,000 yuan (about 1,560 euros). The company justifies the measure by the “significant increase in global storage hardware costs.” In other words, DRAM memory and storage have become so expensive that they can no longer absorb the cost without passing it on to the customer. Until now, no major manufacturer had so explicitly linked a price increase to the memory market, according to collect South China Morning Post. In detail. The ADAS Modern ones (and especially those that integrate LiDAR like those from BYD) are very demanding on memory. They need high-performance chips to process LiDAR point clouds in real time, run driving models, and store route data. The problem is that this same type of memory is being absorbed en masse by artificial intelligence data centers, which account for most of the global production of DRAM and NAND. The prices of these chips have entered what analysts call a “supercycle,” with increases that according to TrendForce are around 55-60% in conventional DRAM this year, but that in premium automotive segments (which also use DDR5) have reached up to 300% in free market price. A problem of scale. BYD’s colossal deployment makes the problem especially bulging in its case. The company has installed your “God’s Eye” system in more than 2.85 million vehicles as of March 2026, generating approximately 180 million kilometers of driving data per day, according to own data of the signature. At that scale, every extra cent in memory multiplies into millions. On the other hand, BYD closed the first quarter of 2026 with its worst net profit in three years: 4.08 billion yuan, a drop of 55% compared to the same period of the previous year, according to figures published by the company. In this context, maintaining prices without making a move has become unsustainable for the company. They are not alone. Chery, Xiaomi and the Huawei Aito brand prices have also increased on models with similar advanced driving systems in recent months. William Li, founder and CEO of Nio, counted in January that the biggest cost pressure of the year would not come from raw materials, but from memory. What changes for the buyer. The founding promise of “God’s Eye” was that autonomous driving would no longer be an expensive privilege. As we counted almost a year agothe experience of the system on the highway (even in the most economical model, the Dolphin Surf/Seagull, which sells for around 9,000 euros in China at the exchange rate) was genuinely impressive. Lane keeping was impeccable, autonomous lane changes were well executed and traffic management rivaled other premium range systems. BYD even planned to distribute it as standard in all its models, regardless of the price. Although that narrative is not dead, it is beginning to have nuances. At the moment, the version with LiDAR (the most capable) is already a payment option that has just become 21% more expensive. And now what. From Counterpoint Research they point that the blow will be uneven: low-end models simply will not carry this technology, and high-end ones have less price-sensitive buyers. The greatest impact falls on the mid-segment, where BYD’s value proposition was most disruptive. As the markets are, we will have to wait to find out what direction the company finally takes. Cover image | BYD In Xataka | Cuba is experiencing a brutal energy crisis, so a Cuban has used ingenuity to fuel his car: charcoal

The RAM crisis is very good news for someone. That someone is Samsung

by

The great supply crisis in 2026 is starring memoirs. Samsung, SK Hynix and Micron control 90% of global DRAM production, and can currently only cover about 60% of projected demand. This is bad news for consumers, and excellent news for giants that cannot keep up with selling memory. Tell Samsung. Samsung Electronics has published its financial report corresponding to the first quarter of 2026. The company recorded revenue of 133.9 trillion won and, so that we understand each other better, this is its all-time quarterly high, with a 43% increase compared to the previous quarter. Memories, memories, memories. The figure is even more surprising if we look closely at the Device Solutions Division, in which its memory business is located. It recorded a sales increase of 86% compared to the previous quarter, with a historical record in operating profit. Samsung itself details that this boom comes from the hand of much higher demand and, to no one’s surprise, a sales price that has increased in the industry. It is not something isolated. Sales related to memories and semiconductors will continue with strong demand throughout the second quarter, Samsung predicts. The company wants to continue capitalizing on demand for GPUs, CPUs and DRAMexpecting advances in agentic AI to continue accelerating demand growth. Why is it important. Samsung’s results are not only good news for the company’s shareholders: they are a reflection of a change in the industry that is here to stay. The RAM crisis will change forever the price of the products we buywill make companies that have never participated in the manufacture of memories have to start considering doing so, like teslaand positions manufacturers like Samsung in a position of power that they have not had for years. The new Samsung. Samsung has always been relevant in semiconductors and memories, but currently this division accounts for 94% of the company’s total operating profit. Virtually every won Samsung earns comes from its device solutions business (RAM and chips). And what about mobile phones?. Although Samsung’s near future will be led by a single division, the company gives enough clues about its future in a territory that touches the average user very closely: mobile phones. Its DX division (in which smartphones are found) grew 19%, with more sales and more profit compared to the previous quarter. Samsung expects a slight drop in revenue next quarter, although it will continue to focus on three clear pillars: high end, folding and series A. In Xataka | There is a company that has grown 3,000% in the stock market, even beating the performance of Nvidia: Sandisk

They say things get worse before they get better. The RAM crisis teaches us that they can always get worse

by

The current situation in which hyperscalers have made all the hardware manufacturers produce almost exclusively for them is leading us to a curious scenario. Apart from the huge RAM and SSD crisis that affects everything –and everyone– Changing from one technology to a newer one no longer depends so much on the needs of a company but on what is barely available on the market. AND The Elec points to a movement by Samsung that represents a new thrust for mobile phones, computers and everything that has soldered RAM. No more LPDDR4 modules. LPDDR4 LPDDR5. They stand for Low-Power Double Data Rate, the low-power version of the RAM tablets that we can buy when building a PC, for example. Unlike conventional RAM pickups, LPDDR memory is soldered to the boardachieving very high speeds with a minuscule energy cost. That is why it is the preferred one for smartphones, tablets and ultrabooks, but it is also ideal for some miniPCs that have become popular in recent months. The downside is that it cannot be expanded or replaced, but its features make it the only option for certain devices. The most powerful versions mount LPDDR5 and LPDDR5Xbut there are still many devices that have the fourth generation versions for cost savings reasons. The turn comes when, according to the South Korean media The Elec, Samsung has begun to cut off the supply of LPDDR4 and LPDDR4X modules to its customers. Translation. Although they are memories with a decade behind them, mid-range and entry-level mobile phones, as well as many other devices, continue to have these versions to keep prices low. At a time when the market is more volatile than evermaintaining those competitive prices by mounting memories that are still interesting in certain ranges was a strategy that made a lot of sense. However, as the media points out, Samsung seems to want to focus on the production of LPDDR5 and LPDDR5X memories. By converting the output lines of the LPDDR4 memories, they will be able to manufacture more new generation RAM, but the price to pay will be that mobile manufacturers will have to switch to that LPDDR5. Mid-range and entry-level smartphones will be faster, but also more expensive. The price to pay. A few weeks ago we already said that the impact was evident. memory represented 20% of the bill of manufacturing an entry-level mobile phone, being one of the most expensive components. And, at that time, the figure was expected to reach 40% by the middle of this year. With this reconversion of Samsung’s lines, we will see where the percentage increase is in a year in which it is already estimating a drop of more than 10% in mobile shipments. The calculations They are tremendous: In the entry range – increments of 30 dollars per unit. In the middle range – from 60 to 80 dollars per unit. In the premium – from 100 and 150 dollars per unit. Samsung itself is not spared. Here you can think that Samsung has a lever to eat the mobile market. That is to say, if it is one of the three that controls the memory production segment and, in addition, has its line of mobile phones and tablets, it can give preferential treatment to its ‘brothers’ to maintain the price in the midst of the crisis. Well no. They already commented that this was not going to happen and, furthermore, it is already flirting with the idea thats Galaxy A17 be an example of this movement. The company’s entry-level mobile has the Exynos 1330 SoC that supports both LPDDR4X and LPDDR5 memories. When the supply of LPDDR4X runs out, they will move to the new generation, which will mean that there will be two different A17s, one of them being one with 50% faster memory than the other. They go direct with HBM. But, as two pieces of news are better understood together, at the same time that the abandonment of the LPDDR4 production lines is pointed out, we have confirmation that Samsung is going to press ahead with the development of HBM memories. These are high-bandwidth memories that are packaged in AI training and inference platforms, and have been reported that Samsung has managed to reduce the HBM memory development cycle from two years to one. It’s a necessary boost to continue being both NVIDIA and AMD’s preferred choice for AI hyperscalers. Shortage. Putting all this together, the result is that there is a RAM crisis for a while. The bottleneck of the industry is enormous and that only three companies –SK Hynix, Micron and Samsung– are the ones who call the shots, and all have opted to satisfy the demands of AI, does not help the situation return to normal. Although there is Chinese companies that may have their opportunity In this scenario, the reality is that the estimate is that all the production of the large It will barely cover 60% of the memory demand until 2027. These companies, of course, are doing great. An example is that, in three months of 2026, Samsung earned more than in all of 2025. But for users and the consumer industry itself, the reality is different. And the worst thing is that there is no realistic date when we will start to see a recovery. NVIDIA has taken the lead, AMD tooand it is no longer just the US and China that need memory: Europe also wants its share of the pie. There are voices that They aim for 2028 as the year of recoverybut other forecasts they go above 2030. What is clear is that there is a crisis ahead In Xataka | TSMC’s only problem was that it was in Taiwan. So the United States has decided to get her out of there

there is no RAM for everyone and hyperscalers have absolute priority

by

When Gabe Newell, head of Valve, asked for help a few weeks ago to find RAM memory anyway and be able to take out his Steam Machine, the comment was half joking… half serious. It was planted in the same GDC in which NVIDIA took advantage of its technology of artificial intelligence to beg beg for some RAM. It was not a situation that caught us by surprise, since we have had quite negative news since 2025 regarding memory supply. RAM, SSD, hard drives and any element that the monster data centers to function. But the wheel does not stop, devices must continue to be launched and the problem is that, beyond the initial ‘run’, no one knows very well if they will be able to continue selling the hardware. And it is a problem that concerns even Apple. Neither one of the main clients of a giant like TSMC is above the needs of the hyperscalers. And he is already suffering the consequences: They have stopped selling Mac models with high amounts of RAM in a coup, especially for professional users who need all the memory possible on their computers. There are no Macs with a lot of RAM left, kid, only Maxibon A few hours ago, media like 9to5Mac and Macrumors They echoed the problem. If you went to Apple’s settings page and tried to purchase a Mac Mini or a mac studio with the largest amount of RAM available (64 GB and 256 GB respectively), the warning was not the one that usually appears on occasions of “it will take x weeks”, but rather a “not available”. That already made us suspect that something was wrong with the supply, but it is also not an isolated problem in the Apple Store in the United States. If right now we go to the Spanish page to configure either of the two Macs with the maximum amount of RAM, it directly sends the same message that it is not available. This goes beyond the classic “reserve and we ship it on a certain date”, it implies that Apple does not accept orders for those specific models. And it is not an anecdote. As pointed out MacRumorsa few weeks ago Apple quietly removed the option to configure Mac Studio with 512 GB, which already indicated that something was up. Other configurations had delivery times of one to five months, and the fact that both models cannot now be configured with the maximum versions of RAM suggests that they will probably also end up disappearing from the store. For most users, 16 and even 32 GB of RAM is more than enough, but those who configure a Mac Mini and Mac Studio model with 64 or 256 GB of RAM do so because it is necessary. It is no longer so much the extra price as it is knowing that that amount of RAM is needed for professional tasks, and eliminating the option (just when Apple has killed the Mac Pro) can be a problem for a niche of users who precisely need these features. For now, if we are going to configure a MacBook Pro, we can choose the maximum amount of RAM without a problem (beyond the longer waiting times than when there was no global supply problem), but in the background there is a much bigger problem. The estimates They point out that RAM producers will increase their production by 16% year-on-year, a figure very far from market needs. Analysts predict that this shortage will last until 2027 or 2028, but also there are more negative estimates They point to 2029 or 2030 to begin to see the market recovery. As soon as possible. And, although we focus on Apple because it is always the most striking case, we have already commented the case of the Steam Machine which cannot be launched because Valve does not have RAM… and there are more cases from manufacturers such as Dell, Lenovo or Asus turning to the Chinese market to be able to launch teams. Computex is one of the big annual events for PC manufacturers and is just around the corner in the middle of an unprecedented crisis. Because there is no RAM, there is no SSD, hard drives are in danger and even graphics cards are not secured. The truth is that it will be interesting to cover the event because the Manufacturers and sellers are going to have to juggle with the little that is available to them. And in Apple’s own upcoming calendar, we will have WWDC where, supposedly, there will be new professional processors. And, no matter how much there are new Apple Silicon, if these professionals cannot configure their computers with large amounts of RAM… it will not be worth much. It could also be a movement in anticipation of the renewal of the equipment, but they eliminated the 512 GB in March and now this is strange. In short, and as has been happening for months, everything is wrong with the RAM market. In Xataka | “We buy anything”: there are stores in Japan so desperate for PC components that anything is worth it

With the RAM market in crisis, an unexpected winner appears: China

by

The saying goes that, in a troubled river, fishermen gain. In the case of the RAM crisisto a troubled market, manufacturers profit. All devices need NAND chips. They are the ones that go into the RAM memory or the storage that is used from the mobile phone to the car, the router and the SD memories and Samsung, SK Hynix and Micron are the ones that control the majority of the production. The data centers need a huge amount of memorywhich has caused everything other than producing for them to be missing out on a large portion of the pie, which is why the three companies have thrown themselves into it. And, since their most important factories cannot do more, they have made the decision to inject a lot of money into China, which is not their favorite scenario, but what gives immediate relief. And all the extra RAM they make… it’s not going to be for us. Exploited. A few weeks ago we said that Jensen Huang, boss of NVIDIA, had met with senior officials from the Asian technology industry, including executives from TSMC and Samsung. He told the first ones to get their act together because NVIDIA was going to need a lot of wafers this year. In the seconds, more of the same, but with HBM4 memory new generation. Shortly after, it was Lisa Su, AMD’s boss, who visited Samsung’s offices in South Korea to reach a deal for HBM4 memory of South Koreans for AMD’s new platform focused on artificial intelligence. Everything moves to the tune of AI training and inference. We are talking about Samsung, but SK Hynix is ​​also developing new generation memory and the objective is the same: to produce everything possible because, although as users we cannot buy RAM or SSD and Valve can’t make the Steam Machinethey are doing great. Wons galore. The problem is that, although the numbers come out, the production lines can’t take it anymore. There are very few companies to create RAM that supplies a brutal demand, and that means that they either expand… or they don’t arrive. And that is precisely what they are doing, but looking at the industrial fabric that can serve as support: what they have been manufacturing in China. In SCMP we can read that Samsung is going to intensify its investment in its Xi’an plant. Specifically, 67.5% compared to the previous year. This will bring the investment to 465.4 billion won -about 264 million euros- in the Chinese plant. This is Samsung’s only plant abroad, and also one of the company’s most important because it is estimated to produce 40% of South Koreans’ NAND memory. The million-dollar investment comes after a few years of hiatus, but they are not the only ones. SK Hynix is ​​also going to inject 581.1 billion won -331 million euros- into its Dalian plant. It is 52% more than in the previous period and the largest disbursement since they acquired the factory in 2022. Immediate relief. The information They point out that it is not so much to produce more, but to satisfy the demand for cutting-edge memories. Recently, Samsung began mass manufacturing the HBM4 memory and SK Hynix the fastest DDR5 memory, and this strategy is focused on the two plants manufacturing that advanced memory instead of the rest of the factories having to adapt to the cutting-edge memory creation processes in order to continue dedicating themselves to other types of NAND chips. It also responds to a more pragmatic vision. Setting up a memory factory is not cheap, but above all, it is not fast. It takes about four or five years to build, polish the clean rooms and optimize the operational lines. It is much faster to adapt existing factories to obtain a much faster response. The reason is that they need wafers, and they need them now. From SK warned that the global shortage of wafers exceeds 20% and, probably, the situation will continue until 2030. Not very favorable weather. The curious thing is that this increase in investments occurs when the situation between China and the United States continues to be very turbulent. Although they have been relaxing, the United States imposed export controls on advanced chip manufacturing equipment destined for China. As much as Samsung is moving money and advanced machines to Samsung, it is in China and that means they must obey Washington’s order. There is licenses and annual permits and, both Samsung and SK Hynix, have a deadline to be able to send tools to their facilities, which are the ones they are taking advantage of because it is estimated that China represents 40% of Samsung’s NAND production and between 40% and 45% of SK Hynix’s. In fact, the company has another plant in Wuxi from which 30% of its NAND chips come out. China, from chill. Whether there is an upsurge in export orders or not remains to be seen. What is on the table at this moment is that China, “without doing anything” (and this with many quotes) is emerging as a very important player in this playing field. It is not only that Samsung and SK Hynix, the two most powerful in the sector, have greatly increased investment in their territory, but that their own RAM companies can see in this scenario the boost they needed to place themselves in the global conversation. One of the largest manufacturers in the country is CXMT and not only have they been polishing their manufacturing process in recent months to create 8,000 MHz DDR5 memories, but they have scaled their production capacity to reach a global market share of between 11% and 13%. Together with the manufacturer YMTC, they are emerging as an opportunity for brands like Lenovo, Dell or Asus, which need RAM to continue selling computers, have available without drastically increasing the price of their equipment. But hey, as we have said more than once in recent weeks, all the extra RAM they manufacture is … Read more

Data centers have eaten up the world’s RAM. Now they threaten to eat the batteries

by

If the question is “what are data centers hungry for,” the answer is a simple “yes.” We hadn’t talked about the RAM memory crisis not because would have finishedbut because it was nonsense keep repeating it. The summary is that things are still as bad as they were a few weeks ago and, although the machines are at full capacity to create more, everything is going to the same place: the AI ​​platforms of the data centers. But it is no longer that they have broken the market for RAM, SSD, hard drives and everything that has to do with chips: it is that they are now going after batteries. The Panasonic case. The Japanese giant advertisement a few hours ago its plan to triple its lithium-ion cell production capacity. They are going to expand their facilities dedicated to this, but they will also adapt some of their manufacturing plants for elements for the automotive industry to manufacture more batteries. All the extra batteries they can make will be few, to the point that they not only propose the change for Japanese plants: also for foreign ones like the one in Kansas. Because? The short answer is because of AI. The long answer is that AI can’t stop working for even a second, and that’s why computers need backup power sources. That energy comes from batteries that are installed between the racks and which, in the event of any outage or specific peak, they ‘pull’ in order to continue operating. And since the equipment requires an insane amount of energy to operate, many, many backup batteries must be made. They are still modules with hundreds of “stacks” that are embedded in the racks All sold. The forecast is such that the Japanese company estimates that, for the next fiscal year, it can sell batteries worth 800,000 million yen, about 5,000 million dollars. It would quadruple its current sales and that implies something else: everything is sold. Its customers have already bought 80% of Panasonic’s output, leaving non-customers to fight for just a fifth of the volume. That will increase prices, generate shortages and cause the same thing that is happening with RAM and other components: there are no units, prices skyrocket, companies see that there is demand and allocate their production to creating that product and the consumer market suffers the consequences. It’s exactly the same thing we’ve seen with HDDs, with Seagate and Western Digital pointing out that what they were going to produce during the next few months was already sold. And it has also happened with RAM. The situation with them became so desperate that the main manufacturers have begun to ask for payments three years in advance. Because as the boss of SMIC – one of the largest foundries in China – pointed out a few days ago, everyone wants to have the infrastructure of the next decade by… yesterday. Supercapacitors. Aside from the “bad” news, Panasonic is also working on something new. Compared to traditional capacitors, the Japanese company is developing supercapacitors for data centers. These are capacitors that can store more energy, but also deliver it more slowly. They are denser than batteries and are expected to be high-fidelity elements to support data center equipment during outages or peak loads. They wait have them ready by 2027. The renewables. In the end, these Panasonic batteries (and other manufacturers) are simple safety elements to ensure that uninterrupted flow of power in the hyperscalers’ racks. How does it affect us? Well, because the capacitors and equipment manufactured by Panasonic are also found in consumer hardware and if they now focus on data centers, the same thing will happen as with NAND chips and everything that uses a memory chip. And, in the background, there are also the most conventional batteries to store a large amount of energy from renewables. Because we have already mentioned that data centers consume a lot, so much so that even has turned to coal, gas is a common resource and there are companies that are opening its nuclear power plants. But if you opt for renewables, it will be necessary to equip data centers with tens of hundreds of batteries capable of absorbing the energy blow. In fact, there are already car battery manufacturers that they are converting. In short: everything bad… except for companies that manufacture those components. Images | panasonic In Xataka | If you were thinking about setting up a NAS to create your own cloud, we have bad news: AI has other plans

Micron knew that the RAM crisis was going to be great for them. The reality that has gone even better

by

As it could not be otherwise, the companies that are benefiting the most from the RAM crisis They are precisely those that have the product and, therefore, they are the ones that set the price. Micron is one of those few companies that is profiting from the excessive demand of this key component for any gadget, a demand caused by the AI ​​fever. The figures from its latest financial report have even exceeded expectations. Although there are some nuances to comment on. Let’s go to trouble. What has happened? Micron just published the results of its second fiscal quarter with numbers that have left analysts speechless. Its revenues have almost tripled those of the previous year, reaching $23.9 billion, well above Wall Street estimateswho expected about 20,000 million. Earnings per share have skyrocketed to $12.20, compared to the $9 projected. And for the third quarter, the company anticipates revenue of approximately $33.5 billion, almost ten points above what the market expected. Those who share the benefit. Artificial intelligence has changed everything in the memory market. The data centers that power AI models require massive amounts of high-performance memory, and the available supply cannot meet that demand. Micron, together with Samsung and SK Hynix, forms the trio that controls practically the entire supply world of high-bandwidth memory (HBM) chips, which are basically one of the key components to run the long-awaited NVIDIA GPUs. Those who buy at any price. Micron’s own CEO, Sanjay Mehrotra, counted to CNBC that the company can only cover between 50% and two-thirds of what its main clients need. Put another way: there is a queue of buyers willing to pay whatever it takes, and Micron simply doesn’t have RAM for everyone. According to SK Group President Chey Tae-won, the global shortage could last another four to five years due to structural bottlenecks in semiconductor production. What’s coming Aware that what is happening now will not last forever, Micron is investing at a speed that has made the market nervous. The company plans to exceed $25 billion in capital spending in 2026 alone, and has already announced that in 2027 that number will rise another $10 billion. Among other operations, it has closed purchasing a plant of Taiwanese Powerchip for $1.8 billion, which will begin producing DRAM wafers in the second half of 2027. The company has also started mass shipments of its new HBM4 memory of 12 layers, which will be directed to the new Vera Rubin platform from NVIDIA. Precisely how much NVIDIA will depend on Micron for this new generation compared to its rivals is the big open question for all investors. Everything is going well for them, but the shares are going down. There has been a bit of a cold reaction in the stock market, as shares have fallen around 5% in the session after the results, despite the fact that the numbers have beaten all forecasts. The reason is the same thing that happened with NVIDIA a few weeks ago: When expectations are very high, even good results can disappoint. From Goldman Sachs they counted that the value could move in a narrow range in the short term after a “very solid quarter with guidance well above consensus, in a context of already elevated expectations.” That has not prevented banks like Wells Fargo or Barclays from updating their upward forecasts to $550 and $670 per share, respectively. The big photo. Micron has accumulated a revaluation of more than 60% so far this year, and has become the most profitable value on the PHLX (Philadelphia Semiconductor Index). Mehrotra affirms that Micron is “the invisible layer that powers AI today.” But it seems that the company is slowly losing that cloak of invisibility. In Xataka | NVIDIA has been pining for months to sell its H200 to China: it just received the news it was waiting for

With the RAM market completely destroyed, Valve has a message to create the Steam Machine: “help”

by

Valve is not having any luck in the hardware world. If with software it is the undisputed queen of the PC ecosystem thanks to Steamwhen they try to launch a console things don’t go so smoothly. More than a decade ago they already tried it with some first Steam Machines that they had no identity. Now they have returned to the fray with a Steam Machine that it looks very goodbut it comes at the worst time. And in the middle of the RAM memory crisisValve only has one thing to say. Aid. The crisis. At Xataka we have been covering the RAM memory news because, although it seems that it is a crisis of a specific component, it is really something that It is affecting the entire industry of semiconductors… and consumption. If in 2020 it was a perfect storm What caused the semiconductor crisis is now the enormous demand for RAM by AI companies. They are all building gigantic data centers and there is a problem: there is only three big RAM manufacturers (plus a fourth that is emerging in China) and all of them have focused on creating RAM for data center equipment. The consequence is that there is no RAM for anyone else. And this not only affects RAM as such: it affects the price of cell phones, computers, cars and even to the router. And, of course, to the Steam Machine. Hey kid, do you have RAM? Valve announced its new machine at the end of last year and they targeted early 2026 to give a release date and price. The problem is that the days were passing, the price of RAM was rising and the question arose: What about Steam Machine? Well what happens is that Valve is desperate. They have already said that it will be released this year (in principle it was going to be spring), it seems that it will be expensive and, in addition, they have pointed directly to the crisis in the supply problems they are having with his other console, the Steam Deck. With this panorama, Valve has appeared at the GDC fair to explain its vision of the console/PC and, in an environment full of manufacturers and professionals, launched a request to the public: If you have access to a large amount of RAM, we are in the market and we would like to buy it.” Complicated. It is a humorous comment, but also somewhat symptomatic. Valve has the money as punishment, but it is not even close to being a premium customer of those few foundries capable of creating RAM. If even Apple can have a bad time, being the second client of the giant TSMCValve does not even enter the annex in the memory request sheet. There are analyzes of all kinds about the consequences of this crisis. In it mobile market will feel a strong bloweven targeting manufacturers that will have to stop launching devices due to market conditions. But on PC, things are more or less the same, with global shipments forecast to be 11% less than the previous year. Captain after the fact. It’s no longer that the Steam Machine may or may not come out, it’s that if it does come out, it would be very expensive. It is something similar to what would happen with the rumored PlayStation 6 that could have seen the light this year and about which we already know that we will not have news in the short term. And here the big question may arise: why didn’t Valve release the Steam Machine when they announced it? Obviously, there were units prepared because they were shown to the press and, furthermore, it is not cutting-edge hardware, so it would have been easy to have it on the market in November 2025. But of course, the situation escalated at a dizzying pace and launching a console at X price and two months later raising it by 200 euros due to the price of RAM or, even worse, stopping selling it because you don’t have units available would have been a tremendous blow. Not so much to the coffers, which in the end with Steam they get a good pinch, but to the reputation. And it is clear that a second disastrous launch of a Steam Machine is something that Valve cannot afford. Now we just have to wait to see when they will be able to launch the machine and, above all, if the price corresponds to components that have already been available for five months. they seemed somewhat fair to us for the most demanding games. Images | DOTA2 The International In Xataka | The price of RAM has skyrocketed and the best example to see the debacle is a 100 euro PC: the Raspberry Pi

Without helium there are no chips or RAM. And the largest producers are in the eye of the Iran war

by

Think of the world as if it were a puppet. It is supported by threads that move, but when one of those threads breaks, the whole wobbles. If several strings break at once, the puppet falls apart. In the technological world, 2026 has started on the wrong foot. The main RAM memory companies They have turned to producing memory for AI, leaving the consumer market. This has caused an unprecedented increase of prices that affects consumers, but also companies. Right now, it’s impossible to guess when it will return to normal because each party involved thinks one thing. And, for a few days now, we have another of those threads that I talked about at the beginning: the iran war. The consequence We are already seeing it immediately: the Strait of Hormuz boilingthe barrel of crude oilreaching stratospheric prices and a gasoline –dieselabove all- through the clouds. But since everything that goes wrong can get worse, now there is another crisis knocking at the door: that of helium. And it is the perfect union between RAM crisis and the war in Iran because without helium… well, without helium there are many things that do not work. Neither does artificial intelligence. RAM crisis + Iran war = no helium For many, helium is that gas that gives us such a funny voice and allows us to inflate balloons that float. For the semiconductor industry, helium is a critical and irreplaceable element in the manufacturing process. Being a noble gas, it does not chemically interfere with the materials of the silicon crystal growth process. inside the huge machines that companies use to create the wafers that are later used to make chips. They prevent materials from reacting with oxygen or other contaminants, so the results are purer. They are like a shield, but helium is also essential to dissipate the heat of the extreme lithography machinesto eliminate waste after each manufacturing cycle and even to identify any leaks in one of these machines. Its particles are among the smallest that exist and are what reveal even the smallest leaks in manufacturing chambers that must be under vacuum. Come on, it is not an element that can be easily replaced. There are two companies that right now have such a deep dependency that any variation in supply would be fatal. What companies? Exactly: Samsung and SK Hynix, the same ones that have dedicated themselves to AI and the same ones that do not plan to lift a finger to alleviate the crisis of RAM memory prices (and therefore of SSDs and any device that has a NAND chip). Both are involved in the manufacturing process of the sophisticated HBM4 memoryand both need helium. The problem is that helium is a byproduct in natural gas production, and some of the world’s largest refineries are in the Middle East. With the war in Iran, it is clear that the civilian targets are data centers and energy producers. If these infrastructures are attacked, the rest of the West is paralyzed, and they have begun to launch kamikaze drones against them. There is the oil company Ras Tanurabut also that of Ras Laffan, from QatarEnergy. It is one of the whales in the production of natural gas and, therefore, in the production of helium. And if the refineries close and the ships do not arrive, the smelters’ reserves begin to run out. There are already voices that they point to problems in the medium term if the situation persists. SK Hynix claims that they have a “diversified supply chain and sufficient helium inventory”something similar to what has commented another of the large chip manufacturers: TSMC. The problem is that these guarantees are short-term. If the situation continues with a prolonged closure of Hormuz, more than 25% of the world’s helium supply will be affected. This will cause the companies that ‘use’ gas the most to begin to see that their reserves are depleted at a faster rate than they are replenished. the market, always so unstablehas already reacted and actions Both Samsung and SK Hynix have fallen in recent hours due to supply concerns. Because we are no longer talking about a price of RAM and runaway gasolinewe are talking about helium being necessary for the manufacturing of any advanced chip, but also in quantum computing or for the numerous space launches. And as Hormuz continues, there will be many entities fighting for an essential, irreplaceable and very valuable good. Faced with SK Hynix’s moderate optimism, more pessimistic voices are already seeing echoes of the component crisis of 2020. Images | VALGO, ASML In Xataka | ‘Focus: The ASML Way’: the book that reveals the secrets of the most powerful European company in the chip industry

Log In

Forgot password?

Forgot password?

Enter your account data and we will send you a link to reset your password.

Your password reset link appears to be invalid or expired.

Log in

Privacy Policy

Add to Collection

No Collections

Here you'll find all collections you've created before.