Apple’s newest iPhone is running into a problem no software update can fix: there isn’t enough memory to finish building it. A billion dollars worth of A20 Pro processors are reportedly sitting at TSMC, fully manufactured but unable to move to the packaging stage because Apple can’t secure enough mobile DRAM, according to semiconductor analyst Tim Culpan and reporting from 9to5Mac and MacRumors. The culprit isn’t a factory fire or a trade dispute. It’s AI. Samsung, SK Hynix and Micron are steering their memory output toward data centers building large language models, and phones are losing out.
The timing could hardly be worse. iPhone 18 Pro, iPhone 18 Pro Max and Apple’s first foldable device are expected within weeks, and the same memory crunch that’s driving HBM4 shortages for Nvidia’s Rubin platform is now showing up on Apple’s balance sheet. Forbes contributor Jon Markman reported on August 6 that memory makers are prioritizing high-margin server DRAM for AI, leaving consumer devices short. Six weeks later, that shortage hasn’t eased. It has, if anything, gotten more expensive.
A Billion Dollars in Chips, Stuck in Limbo at TSMC
According to Culpan’s reporting, cited by both 9to5Mac and MacRumors on August 6, TSMC has produced A20 Pro wafers on its N2 node with good volume and solid yields. The problem sits one step later in the process. Before a processor can be packaged and shipped to Apple’s assembly partners, it needs to be paired with mobile DRAM, and that DRAM simply isn’t showing up in the quantities Apple ordered. MacRumors put a dollar figure on the backlog: roughly $1 billion worth of unpackaged Apple processors, waiting on memory chips that haven’t arrived.
Forbes’ David Phelan added more detail on August 8, reporting that the shortage threatens not just the standard lineup but also Apple’s rumored foldable device, sometimes referred to in supply-chain circles as “iPhone Ultra.” Phelan’s reporting pointed to the same root cause Markman identified two days earlier: memory suppliers redirecting capacity toward AI infrastructure, leaving smartphone makers scrambling for what’s left.
Why a New Chip Design Made Apple More Exposed
Part of what makes this shortage bite harder than past ones is Apple’s own engineering choice. The A20 Pro reportedly uses an integrated design that places memory closer to the processor on the same wafer, a shift meant to boost performance and efficiency. That approach also means Apple can’t simply swap in whatever DRAM happens to be available. It needs specific, qualified memory that meets the new packaging spec, and qualifying alternate suppliers takes time Apple doesn’t have with a launch six weeks out. Remio.ai’s supply-chain writeup framed it plainly: Apple lacks enough qualified DRAM at the right time, and the shortage reflects a capacity contest between consumer devices and AI infrastructure.
None of this comes from an official Apple, TSMC or supplier disclosure. Kursiv Media’s August 7 report noted explicitly that Apple has not publicly confirmed the shortage or detailed its procurement plans, even as the company works with suppliers behind the scenes to close the gap. That’s worth keeping in mind: everything specific about wafer counts and dollar figures traces back to analyst sourcing, not Apple’s own statements.
The Analysts Behind the Warning
Two names keep surfacing in this story. Tim Culpan, the semiconductor analyst whose reporting anchors the 9to5Mac and MacRumors pieces, first flagged the stalled wafers and used the phrase “scrambling for DRAM” to describe Apple’s position. Ming-Chi Kuo, the TF International Securities analyst with a long track record on Apple’s supply chain, went further in a note picked up by ETNow on August 12: Apple is scaling back its 2026 device shipment plans across the board because of the global DRAM shortage, not just for iPhone 18 but for its wider hardware lineup.
Kuo’s shipment-cut claim matters because it moves the story beyond a single product line. If accurate, it suggests Apple is making company-wide production decisions in response to memory costs and availability, not just adjusting one phone’s launch quantities. A separate Weibo leak, discussed on The MacRumors Show back in May, claimed the standard iPhone 18 (as opposed to the Pro models) was already being trimmed on display and chip specifications as a cost-control response to memory prices. That leak predates the August reporting by three months, suggesting Apple may have seen this coming well before the wafers started piling up at TSMC.
DRAM Prices, By the Numbers
The scale of the price movement behind this shortage is unusual even by chip-industry standards. TrendForce data cited by The Register in January showed Samsung and SK Hynix pitching first-quarter 2026 server DRAM prices up 60 to 70 percent versus the fourth quarter of 2025 to major cloud clients including Google and Microsoft. That wasn’t limited to server buyers. PC and smartphone DRAM customers saw comparable increases in the same window, according to the same reporting.
NetworkWorld’s coverage of Samsung’s pricing moves adds a concrete example: 32GB DDR5 modules that sold for $149 in September 2025 were priced at $239 by early 2026, a roughly 60 percent jump. Contract pricing for DDR5 moved even further, surging past 100 percent to reach $19.50 per unit compared with around $7 earlier in 2025. Gartner’s own forecast, also reported by NetworkWorld, puts 2026’s overall DRAM price increase at roughly 47 percent, driven by what the firm describes as significant undersupply in both current-generation and legacy DRAM.
| Metric | Baseline (2025) | 2026 Figure | Change | Source |
|---|---|---|---|---|
| Server DRAM, hyperscaler pricing | Q4 2025 level | Q1 2026, up to +70% | +60% to +70% | The Register / TrendForce |
| Conventional DRAM contract price | Q4 2025 level | Q1 2026 forecast | +55% to +60% QoQ | The Register / TrendForce |
| Samsung 32GB DDR5 module | $149 | $239 | +60% | NetworkWorld |
| DDR5 contract price, per unit | ~$7 | $19.50 | +179% | NetworkWorld |
| Samsung DRAM revenue per bit | $0.36 | $0.79 (forecast) | +116% | S&P Global Market Intelligence |
| SK Hynix DRAM revenue per bit | $0.40 | $0.70 (forecast) | +78% | S&P Global Market Intelligence |
| Micron DRAM average selling price | lower base | $1.06 (forecast) | +54% | S&P Global Market Intelligence |
| Gartner 2026 DRAM price forecast | — | Full-year 2026 | +47% | Gartner / NetworkWorld |
Samsung, SK Hynix and Micron Are Choosing AI Over Phones
This isn’t a shortage caused by a broken factory or a natural disaster. It’s a deliberate allocation decision by three companies that control most of the world’s DRAM supply. NetworkWorld reported that SK Hynix told investors on an earnings call that its HBM, DRAM and NAND capacity is “essentially sold out” for 2026, a statement tied directly to AI-driven demand rather than any supply disruption. The company posted record quarterly operating profit exceeding $8 billion on the back of that demand, according to the same report.
S&P Global Market Intelligence’s January analysis lays out just how lucrative that reallocation has become for memory makers. Samsung’s revenue per bit from traditional DRAM is forecast to rise 116 percent year-over-year to $0.79 in 2026, up from $0.36. SK Hynix is projected to see a 78 percent increase to $0.70, and Micron’s average selling price is expected to climb 54 percent to $1.06. Tech-Insider’s coverage adds that Micron’s own guidance points to continued tight DRAM supply through at least its fiscal fourth quarter of 2026, with no clear date for relief. When the world’s three biggest DRAM suppliers are all telling investors that AI customers are locking down capacity at premium prices, phone makers like Apple are, by definition, standing at the back of the line.
What Tim Cook Has Said About the Memory Crunch
Apple’s public comments on the shortage have come almost entirely through Tim Cook’s remarks on quarterly earnings calls, not through a dedicated announcement. Speaking on Apple’s July 2026 earnings call, Cook told analysts, “We’re seeing some very significant constraints currently with limited flexibility in the supply chain to remedy it,” as TechCrunch reported on July 30. That’s about as directly as Apple’s chief executive has acknowledged the problem in public.
Cook had signaled the pressure was building for months. On Apple’s January 2026 earnings call, he told investors the company continued to see market pricing for memory increasing significantly, per Reuters. In February, asked how Apple planned to respond, Cook said there were different levers the company could pull, without specifying which ones or how effective they’d be, according to Reuters. By May, per CNBC’s reporting, Cook had grown more direct, telling investors Apple believed memory costs would drive an increasing impact on its business.
None of this has come with a corresponding drop in demand. In earlier remarks reported by Yahoo Finance, Cook said, “We were excited by the feedback from customers regarding the latest iPhone series. It far surpassed our expectations, and iPhone sales grew by 23%.” That combination, strong sell-through on the existing lineup plus a tightening memory supply for the next one, is exactly the kind of setup analysts say points to a fast sellout once iPhone 18 actually reaches shelves.
The Nvidia Connection: One Shortage, Two Very Different Victims
Apple’s DRAM problem and Nvidia’s HBM supply constraints trace back to the same root cause, even though they show up in different products. SK Hynix, the company that told investors its 2026 capacity is essentially sold out, is also one of the primary suppliers of high-bandwidth memory for Nvidia’s GPUs, including the chips powering the AI server price increases shattered.io covered earlier this year. When a memory maker prioritizes HBM for Nvidia’s data-center customers, it has less traditional DRAM capacity left over for laptops and phones. That’s not a coincidence. It’s the same pool of fab capacity being split between two very different buyers, one of whom is willing to pay far more per gigabyte.
The knock-on effects reach Apple’s own hardware lineup beyond the iPhone. Rising DRAM and NAND costs have already pushed up pricing across the industry, a dynamic that shows up in the RTX 5090’s own price surge as GPU makers pass memory costs on to buyers. Apple’s Mac Studio, Mac mini and MacBook Air configurations, which shattered.io covered when Apple’s M6 and M5 Ultra chips launched earlier this year, rely on the same DRAM supply chain as the iPhone. A squeeze that starts in mobile memory doesn’t stay contained to phones for long.
| Comparison Point | Apple / Mobile DRAM Squeeze | Nvidia / HBM Squeeze |
|---|---|---|
| Primary suppliers affected | Samsung, SK Hynix | SK Hynix, Samsung, Micron |
| What’s constrained | Mobile DDR/LPDDR for A20 Pro packaging | HBM for GPU/accelerator packaging |
| 2026 capacity status | LTA secured for H1 2026 only, per Notebookcheck | “Essentially sold out” for 2026, per SK Hynix earnings call |
| Buyer leverage | Competing against AI hyperscalers for allocation | Direct AI hyperscaler demand, premium pricing |
| Visible symptom | Stalled A20 Pro packaging, potential iPhone 18 shortages | Higher AI server prices, extended GPU lead times |
2021 vs. 2026: Why This Chip Shortage Feels Different
Apple lived through a chip shortage before, in 2021, and it’s worth asking whether this is a repeat. The two episodes share a symptom, tight component supply, but the underlying mechanics differ. The 2021 shortage traced back to pandemic-driven factory shutdowns, a snap-back in demand that caught automakers and electronics firms under-ordering parts, and general logistics chaos across shipping and fabrication. It hit automotive microcontrollers and GPUs hardest, and it eased over roughly two years as fabs caught up and demand normalized.
The 2026 squeeze looks structurally different. Nobody’s factory shut down and nobody under-ordered. Instead, memory makers are actively steering supply toward the highest-margin customer, which right now is the AI industry rather than smartphone makers. That’s a strategic reallocation, not an accident, and it explains why relief doesn’t have an obvious end date the way the 2021 shortage eventually did. Micron’s own guidance, per Tech-Insider, points to tight supply persisting through at least its fiscal fourth quarter of 2026. Nothing in the current reporting suggests memory makers have a reason to change course as long as AI infrastructure spending keeps climbing.
What It Means for iPhone 18 Price and Availability
The most immediate consequence, according to every analyst cited in this story, is a rough launch window for anyone trying to buy an iPhone 18 Pro, Pro Max or the rumored foldable at release. Kuo’s note, picked up by ETNow, frames Apple’s 2026 shipment cuts as a company-wide response, meaning constrained supply likely extends past the flagship models. Notebookcheck’s reporting on Apple’s DRAM long-term agreement situation adds another wrinkle: Apple, sourcing from Samsung and SK Hynix, only secured a supply commitment for the first half of 2026 rather than its usual full-year deal, according to ZDNet Korea. That leaves the back half of the year, which covers most of the iPhone 18 production cycle, considerably less certain.
Price increases are the other likely outcome. Notebookcheck’s reporting explicitly flags an expected iPhone 18 price hike tied to the DRAM situation, and Tech-Insider’s coverage cites a roughly 29 percent rise in memory-linked costs feeding into Apple’s pricing decisions. Given that DRAM contract prices have already more than doubled in some segments since early 2025, a component cost increase of that size is not a stretch. Whether Apple absorbs the hit or passes it to buyers is the open question heading into launch.
How Rivals Are Positioned in the Same Squeeze
Apple isn’t the only phone maker buying from Samsung and SK Hynix, and it isn’t the only one facing higher component bills. But Apple’s position differs in one important way: its integrated memory design on the A20 Pro requires specifically qualified DRAM, which narrows the pool of usable suppliers compared with rivals running more conventional memory configurations. Samsung, notably, sits on both sides of this story. It’s simultaneously one of the DRAM makers raising prices and a smartphone manufacturer that will feel those same price increases on its own Galaxy lineup. That dual position gives Samsung more room to manage its own margins than Apple, which has to buy DRAM from the same companies it competes against in the phone market.
Chinese phone makers reliant on the same DRAM suppliers face a similar squeeze, though with less public reporting available on their specific allocation agreements. What’s consistent across the industry is the incentive structure: any manufacturer without a locked-in, full-year supply agreement is exposed to the same pricing spikes hitting Apple’s cost sheet right now.
Five Predictions for the Rest of 2026
- iPhone 18 Pro models sell out within days of launch. With Kuo, Culpan and multiple outlets all pointing to constrained packaging capacity, initial allocation is likely to fall short of pre-order demand in most regions.
- Apple raises prices on at least one iPhone 18 configuration. Given the scale of DRAM cost increases already documented by NetworkWorld and Tech-Insider, absorbing the full hit without any price adjustment looks unlikely.
- DRAM prices stay elevated through at least mid-2027. Micron’s own guidance offers no relief date within fiscal 2026, and SK Hynix’s sold-out capacity signals the crunch extends into next year’s planning cycles too.
- Mac and iPad pricing sees knock-on pressure. Since Apple’s other product lines draw from the same DRAM pool, expect similar cost pressure to show up in Mac Studio, Mac mini and iPad pricing conversations before the end of 2026.
- Memory makers keep favoring AI customers over consumer electronics. As long as hyperscalers are willing to pay premium prices for HBM and server DRAM, Samsung, SK Hynix and Micron have little financial incentive to redirect capacity back toward phone makers.
The Bigger Picture: Consumer Tech Is Competing With Data Centers for the Same Chips
Strip away the specific numbers and what’s left is a pattern that keeps repeating across the tech industry in 2026: consumer hardware and AI infrastructure are drawing from the same limited pool of advanced memory, and AI keeps winning the allocation fight. It happened with high-bandwidth memory for GPUs, it’s happening now with mobile DRAM for phones, and there’s little in current supplier guidance suggesting that dynamic reverses soon. Apple’s situation is a visible, well-documented example of a squeeze that’s quietly reshaping cost structures across laptops, tablets, gaming hardware and enterprise servers alike. The company that once set the pace for chip supply chains is now, on this particular component, waiting in line behind the AI industry it also supplies chips to through its own silicon business.
Frequently Asked Questions
What exactly is causing the iPhone 18 DRAM shortage?
Memory makers Samsung, SK Hynix and Micron are directing a growing share of their DRAM and HBM production toward AI data centers, which pay premium prices for capacity. That leaves less mobile DRAM available for smartphone makers like Apple, according to reporting from Forbes and Remio.ai.
Is the iPhone 18 launch actually going to be delayed?
No delay has been reported. The concern, according to Culpan’s reporting cited by 9to5Mac and MacRumors, is limited launch-day availability rather than a pushed-back release date.
Will iPhone 18 cost more because of the DRAM shortage?
Analysts cited by Notebookcheck and Tech-Insider expect a price increase tied to rising memory costs, though Apple has not confirmed specific pricing.
Does this affect Macs and iPads too?
Apple’s other product lines draw from the same DRAM supply chain, so cost pressure from the shortage is likely to extend beyond the iPhone over time, even though the most acute reporting so far centers on iPhone 18.
How is this different from the 2021 chip shortage?
The 2021 shortage stemmed from pandemic-related factory shutdowns and a demand snap-back. The 2026 squeeze comes from memory makers deliberately prioritizing AI customers over consumer electronics, which is a structural shift rather than a temporary disruption.
Has Apple officially confirmed the DRAM shortage?
Not directly. Apple has not issued a dedicated statement on the shortage. Tim Cook has acknowledged memory-related supply constraints on quarterly earnings calls, as reported by TechCrunch and Reuters, but the specific wafer and dollar figures come from analyst sourcing, not Apple disclosures.
When will DRAM supply and pricing normalize?
There’s no confirmed date. Micron’s guidance points to tight supply persisting through at least its fiscal fourth quarter of 2026, and SK Hynix has described its 2026 memory capacity as essentially sold out.
Are Samsung and other Android phone makers facing the same squeeze?
Yes, though Samsung’s position is more complex since it is both a DRAM supplier raising prices and a phone maker that will absorb those same cost increases on its own devices.
