The surge in spending on artificial intelligence infrastructure is starting to bite into the supply of the more ordinary chips that power laptops and smartphones. Samsung Electronics and SK Hynix, two of the world’s largest memory makers, are warning that customers in the PC and mobile markets are already struggling to secure enough components as capacity is pulled toward AI data centers. Their message is blunt: the AI boom is turning into a squeeze on conventional memory, with potential consequences for device prices, product launches, and the pace of consumer upgrades.
What looks like a technical supply chain issue is, in practice, a reshuffling of who gets priority in the semiconductor economy. High bandwidth memory for AI servers now commands the highest margins, and that is forcing a rethink of how much DRAM and NAND can be reserved for more price sensitive products such as budget Android phones or entry level Windows laptops. I see this as the first real stress test of whether the AI build out can coexist with healthy consumer hardware markets, or whether one will be sacrificed to feed the other.
AI servers are crowding out everyday devices
Samsung Electronics and SK Hynix have both told investors that PC and mobile customers are finding it harder to lock in memory orders as demand for AI infrastructure soaks up capacity. In their latest briefing, executives said that PC and mobile are being directly and indirectly affected as more wafers are diverted to lucrative AI chips. That is a striking admission from companies that built their dominance on supplying DRAM and NAND for consumer electronics, and it signals that the balance of power inside their own product portfolios has shifted decisively toward data center parts.
The race to build AI infrastructure has pushed chipmakers to prioritize high bandwidth memory, or HBM, which is essential for training and running large language models. Industry executives say capacity is being reallocated toward HBM for AI, leaving less room for conventional DRAM used in mainstream laptops and phones. From my perspective, that is not just a short term bottleneck but a structural repricing of memory, because the most profitable customers are now hyperscale cloud providers rather than PC OEMs or smartphone brands.
Samsung and SK Hynix lean into AI, even as they warn
Samsung Electronics and SK Hynix are not passive victims of this shift, they are active participants who are reshaping their own businesses around AI demand. In Seoul, Jan reporting highlighted that Samsung and Hynix are two of the world’s largest memory chipmakers and are now channeling more investment into advanced AI products. At the same time, they are cautioning that this pivot will leave fewer chips available for lower margin segments, especially budget smartphones and PCs that cannot easily absorb higher component costs.
Samsung has been explicit that it sees an acute chip shortage persisting even as its own profits improve. The company told analysts that it expects a continued shortfall and warned of a headwind for mobiles after its profit roughly tripled, with AI related demand remaining strong. I read that as a signal that, even with more money coming in, the company is not rushing to expand low end capacity, because the returns are simply better in AI.
Smartphone and PC markets face a double squeeze
The timing of this supply crunch is particularly awkward for consumer hardware makers, because demand for new devices is already fragile. Research firms IDC and Counterpoint now expect global smartphone sales to shrink at least 2 percent this year, reversing earlier forecasts for growth, according to IDC and Counterpoint. That means phone makers are being hit from both sides, with weaker consumer appetite and rising component costs that are hardest to pass on in mid range and entry level models.
Analysts at IDC have already flagged that the AI build out is leaving less DRAM available for consumer devices, which is exacerbating price pressure in what they describe as a tight market. Their assessment is that this dynamic is reversing years of progress in bringing affordable smartphones to more markets. From my vantage point, that suggests we could see fewer aggressively priced devices like the Samsung Galaxy A15 or Xiaomi Redmi series with generous memory configurations, as brands either trim specs or accept lower margins.
Inside the DRAM bottleneck and rising costs
At the heart of this crunch is DRAM, the workhorse memory that sits inside everything from gaming PCs to low cost Chromebooks. The world’s two largest DRAM chip manufacturers, Samsung Electronics and, have warned that strong demand for AI related products is creating shortages in more conventional chips. Park Joon Deok, head of one of their memory businesses, has pointed to rising memory chip costs as a direct consequence of this tightness, which is already feeding through to contract prices for PC and phone makers.
Industry analysis of the global memory shortage notes that the shift of capacity toward AI has left less DRAM available for consumer devices, which is tightening supply and pushing up prices. I expect that to show up first in higher end configurations, such as 32 GB gaming laptops or 512 GB flagship phones, where manufacturers may either raise prices or delay broader rollouts until they can secure enough inventory at acceptable costs.
Strategic bets, Nvidia ties, and what comes next
Samsung is not just reallocating capacity, it is also trying to catch up with SK Hynix in the most profitable AI segments. Analysts expect some of its advanced memory chips to be initially shipped to Nvidia, which has become the anchor customer for many AI memory suppliers. That relationship helps explain why Samsung has been trying to catch up with its cross town rival SK Hyni in the lucrative HBM segment, even as it acknowledges that the resulting shortages are a headwind for its own Galaxy smartphone line.
Both Samsung and SK Hynix have issued what one report described as a stark warning that AI chip demand is squeezing consumer supply, with Samsung Electronics and both emphasizing the strain on conventional products. A related analysis of the same warning underlined that the world’s two largest DRAM manufacturers are effectively choosing to serve AI first, then allocate what is left to PCs and phones. From where I sit, that hierarchy is unlikely to change as long as AI customers are willing to pay a premium, which means consumers should brace for a period in which the latest AI breakthroughs are powered by chips that their own devices struggle to obtain.
For buyers, the practical impact may be subtle at first, such as fewer discounts on high memory variants of laptops like the Lenovo Legion series or delayed launches of mid tier 5G phones in emerging markets. For the industry, the bigger question is whether this AI centric allocation of silicon becomes the new normal. If it does, then the warnings from Samsung and SK about squeezed supplies will look less like a temporary crunch and more like the opening chapter of a longer realignment of the chip economy, one where AI servers sit firmly at the front of the line.
