New lab data published on September 21, 2026, shows exactly how the storage in Apple’s iPhone 18 Pro Max — launched that same month with a 6.9-inch display across four capacities, 256GB, 512GB, 1TB, and 2TB — slows down as the drive fills up, and the numbers go well beyond a simple benchmark footnote. According to testing lab HOMOLAB, cited by Tom’s Hardware, sustained write speeds on the 1TB iPhone 18 Pro Max can collapse to as low as 1.1 MB/s once the phone’s storage is roughly 60% full and its cache buffers run dry. That is a reported 38% drop compared with the TLC-based storage Apple uses in the 256GB and 512GB Pro models. The culprit is QLC NAND, a cheaper, denser flash memory type Apple introduced in the 1TB and 2TB tiers to hit higher capacities without redesigning the phone.
What makes this story different from a routine storage complaint is the mechanism behind it. Apple’s engineers built a caching system that hides QLC’s weaknesses during ordinary use, so most buyers will never see the phone crawl along at hard-drive speeds — the same phone Apple markets on battery claims of up to 30 hours of typical use and, per its October 2026 spec update, up to 45 hours of video playback, six hours more than the iPhone 17 Pro Max offered. But that protection depends entirely on how full the drive is, and the newest benchmarks show precisely when it breaks down, and by how much. Tech-Insider’s earlier report on the Notebookcheck benchmark first flagged the cache-exhaustion pattern; this piece breaks down what that pattern means for real creative workflows.
Don't miss new tech stories on Google
Add Tech Insider once in the Google app and our stories appear in your news suggestions.
What HOMOLAB’s Benchmarks Actually Measured
HOMOLAB ran sustained, heavy-load write tests against the 1TB iPhone 18 Pro Max, pushing large sequential writes at the storage controller until its fast cache layers were exhausted and the phone had to write directly to the underlying QLC flash. AppleInsider reported that on a near-empty unit, throughput fell from an initial burst above 3,000 MB/s down to an average of 79.4 MB/s once the cache ran out, with occasional dips to 25.6 MB/s. That alone is a steep decline, but it is not the headline number. The real shock came when the same test ran on a drive filled to roughly 60% of its capacity: average direct-to-QLC write speed dropped to about 45 MB/s, and the worst recorded moments hit just 1.1 MB/s.
For context, 1.1 MB/s is slower than a 2007-era USB flash drive and well below what most microSD cards deliver even in budget tiers. It is also the number Tom’s Hardware used to frame its September 21 report, describing the iPhone 18 Pro Max’s worst-case storage performance as capable of dropping “lower than a hard drive.” Sportskeeda Tech independently confirmed the same figures from the HOMOLAB data set, adding detail on why the fill level matters so much.
How the SLC Cache Collapses as the Drive Fills
Modern flash storage, including the QLC NAND in the iPhone 18 Pro Max, rarely writes data directly to its slowest, densest cells during normal use. Instead, controllers reserve a portion of the chip to behave like faster, less dense memory, a technique commonly called SLC caching. Writes land in this fast cache first, and the controller migrates the data to the denser QLC cells later, in the background, when the phone is idle.
The problem is that this cache is not fixed in size. Sportskeeda Tech’s coverage of the HOMOLAB results notes that the available SLC cache shrinks from around 250GB on a nearly empty 1TB drive down to only about 58GB once the phone reaches roughly 60% capacity. That is a drop of more than 75% in usable fast-cache headroom, and it happens exactly as users accumulate the photos, videos, and apps that make a 1TB or 2TB phone appealing in the first place. Once that shrunken cache fills during a long write, every subsequent byte has to go straight into QLC cells at their native, much slower speed.
There is also a secondary buffer layer in play, and outlet Shattered’s own September 2026 breakdown of the benchmark data corroborated the same three-stage architecture: an SLC cache, a simulated TLC-mode buffer, and finally raw QLC. Reporting on the same benchmark set, outlets tracking HOMOLAB’s data describe that middle tier as behaving like TLC-mode storage, writing at roughly 578 MB/s when the drive is empty. That middle tier also degrades as the phone fills, falling to around 396 MB/s at the same 60% capacity mark. In other words, the slowdown is not a single cliff. It is a multi-stage decline: fast SLC cache, then a slower TLC-mode buffer, then, only when both are exhausted, the raw QLC floor that can bottom out at 1.1 MB/s.
| Write Stage | Near-Empty Drive | ~60% Full Drive |
|---|---|---|
| Burst speed (SLC cache) | Up to 3,000 MB/s | Up to 3,000 MB/s (smaller window) |
| SLC cache capacity available | ~250 GB | ~58 GB |
| Middle buffer (TLC-mode) | ~578 MB/s | ~396 MB/s |
| Direct QLC average | 79.4 MB/s | ~45 MB/s |
| Direct QLC worst-case minimum | 25.6 MB/s | 1.1 MB/s |
Data in the table above is drawn from HOMOLAB’s testing as reported by Tom’s Hardware, AppleInsider, and Sportskeeda Tech on September 20 and 21, 2026.
QLC vs TLC NAND: Why Apple Made This Trade-Off
The root cause is a component decision Apple made months before launch. QLC, short for quad-level cell, stores four bits of data per memory cell instead of the three bits TLC stores. That extra bit per cell lets manufacturers pack more capacity into the same physical die area, which is exactly why it shows up in the highest-capacity tiers rather than the base models. Techtimes explained the trade-off plainly: QLC allows more data per wafer at lower cost, but at the expense of performance, particularly under sustained, heavy write loads. GSMArena’s October 2026 teardown corroborated the split independently, confirming that the 256GB and 512GB tiers use TLC NAND while the 1TB and 2TB tiers use QLC NAND, even though every capacity shares the same 12GB of LPDDR5X RAM, which rules out memory as the source of the slowdown.
Industry coverage of the decision goes back further than this week’s benchmarks. Telbb reported as early as July 2026 that QLC drives typically run 30% to 40% slower than TLC in sustained writes for large file transfers, and also carry lower write endurance over the life of the device. Wccftech was more pointed, framing Apple’s move as swapping “the faster TLC for slower QLC storage” in the 1TB and 2TB iPhone 18 Pro Duo models while continuing to charge premium prices for those tiers. That framing matters commercially: the 1TB and 2TB configurations are the most expensive iPhone 18 Pro Max variants, and buyers paying the most are the ones getting the storage technology with the weaker sustained-write ceiling. It is part of a broader pattern of memory-market tension this year, alongside the HBM4 memory shortage squeezing DRAM inventory and the DDR5 price spikes forcing AMD to extend its AM4 platform, both of which have pushed device makers toward cheaper, denser memory tiers wherever they can get away with it.
| Attribute | TLC NAND (256GB / 512GB tiers) | QLC NAND (1TB / 2TB tiers) |
|---|---|---|
| Bits stored per cell | 3 | 4 |
| Sustained write speed vs TLC | Baseline | 30-40% slower (Telbb); up to 38% slower under stress tests (Tom’s Hardware) |
| Write endurance | Higher, per industry reporting | Lower, per industry reporting |
| Manufacturing cost per GB | Higher | Lower |
| Storage tier used in | iPhone 18 Pro / Pro Max, 256GB and 512GB | iPhone 18 Pro / Pro Max, 1TB and 2TB |
Real-World Impact: ProRes and Long-Form Video
Benchmark numbers only matter if they touch something people actually do with the phone, and video is the clearest case. Standard 4K HEVC recording sits well under the reduced QLC ceiling even in the worst-case scenario, so casual video shooters are unlikely to notice anything. ProRes is a different story. Apple’s mobile ProRes profiles at 4K and above can run into the hundreds of megabits per second, and sustained recording on a phone that is already 60% full and mid-way through the QLC slowdown risks brushing up against that 45 MB/s average, with the 1.1 MB/s floor as a worst case during long, uninterrupted takes.
This is precisely the audience Apple markets the 1TB and 2TB tiers toward: creators who need the extra headroom for long 4K or 8K ProRes captures. The irony, as multiple outlets have pointed out, is that the storage tier built for that workload is the one most exposed to the slowdown once the phone accumulates enough footage to push past the 60% fill threshold.
RAW Photo Bursts and Buffer Clearing
High-resolution RAW photography is a similarly sustained-write workload, just delivered in short, repeated bursts rather than one continuous stream. Each RAW frame can run into tens of megabytes, and a long burst sequence generates the kind of multi-hundred-megabyte write spike that eats into cache quickly. While the SLC cache is intact, this is invisible to the user. Once it is depleted on a partially filled 1TB or 2TB drive, the phone has to fall back to the slower middle buffer and eventually direct QLC writes, which can visibly delay how quickly shots clear from the buffer and become available for review or editing.
Photographers who shoot fast-moving subjects, sports or wildlife being the obvious examples, are the ones most likely to notice shot-to-shot delays on a nearly full high-capacity model, even though the same phone felt perfectly responsive when it was new and mostly empty. Creators working around this limitation increasingly offload footage to external drives instead of relying on internal capacity; our comparison of the Corsair EX400U, Samsung T9, and SanDisk Extreme portable SSDs covers which external options can actually keep up with sustained 4K and RAW transfers.
Game Installs and iOS Updates on a Nearly-Full Phone
The slowdown is not limited to camera workflows. Installing a large mobile game, restoring apps from iCloud, or applying a full iOS system update all involve writing gigabytes of data in one continuous operation, which is exactly the pattern that exposes QLC’s weak point. A modern AAA-style mobile title can exceed 20GB to 30GB. At a sustained 45 MB/s, the average rate HOMOLAB recorded at 60% fill, installing a 30GB game works out to roughly 11 minutes of write time alone, before accounting for unpacking and verification. On a TLC-based 256GB or 512GB model maintaining several hundred megabytes per second, the same install finishes in a fraction of that time.
System updates carry the same risk. iOS updates rewrite large portions of the system partition in a short window, and on a heavily filled 1TB or 2TB Pro Max, that process can take noticeably longer than on a TLC-based tier, particularly if the update lands while the phone is already past the 60% capacity mark where the cache has shrunk to its smallest.
Timeline: How the QLC Story Broke Ahead of Launch
This week’s benchmarks are the conclusion of a story that has been building since before the iPhone 18 Pro Max even launched. Wccftech first reported in July 2026 that Apple was swapping TLC for QLC in the 1TB and 2TB tiers. Telbb followed with a deeper breakdown of the sustained-write and endurance trade-offs the same month. By mid-August, supply-chain leaks had been picked up by wider tech press, warning prospective buyers that the top storage tiers would run 30% to 40% slower once their caches filled.
Those warnings stayed largely theoretical until launch week. Apple held its keynote on September 9, with retail availability following on September 18. Independent lab testing did not surface hard numbers until HOMOLAB’s benchmarks began circulating around September 19 and 20, at which point Notebookcheck, Tom’s Hardware, AppleInsider, and Sportskeeda Tech all published detailed breakdowns within a 48-hour window. That is the sequence that turned a known trade-off into a specific, quotable number: 1.1 MB/s, 38% slower, and a cache that shrinks from roughly 250GB to 58GB.
Historical Precedent: QLC Backlash Is Not New
None of this is a first for the storage industry, even if it is a first for a flagship iPhone. QLC NAND has circulated in the PC and laptop SSD market for years, and it has consistently been positioned as a budget or high-capacity option rather than a performance one. Techtimes and Telbb both describe the same pattern in that market: QLC drives sell on capacity-per-dollar, not sustained throughput, and enthusiast buyers who assumed a “faster because newer” story were often disappointed once independent reviewers ran the same kind of sustained-write tests HOMOLAB just ran on the iPhone 18 Pro Max.
What is new is seeing that pattern applied to a phone marketed under Apple’s “Pro” branding, sold at the company’s highest price points, and aimed explicitly at creators who need to trust their storage under continuous professional workloads. Wccftech’s criticism of “sky-high” pricing for the affected tiers reflects that gap between positioning and performance directly.
Market Reaction: What Analysts and Outlets Are Telling Buyers
Concrete sales, return, or exchange figures tied to this story have not been published as of September 21, 2026. What has emerged instead is a fast-forming consensus in buying guidance. Coverage of the QLC decision has increasingly steered everyday buyers toward the 256GB and 512GB TLC-based tiers, reserving the 1TB and 2TB QLC models for users who specifically need the extra capacity and are prepared to manage the sustained-write trade-off, for example by keeping the drive well under the 60% fill threshold where the worst numbers appear. That is the same conclusion Tech-Insider’s coverage of the QLC storage row reached when it examined why the 512GB tier has become the pragmatic default recommendation rather than the entry-level compromise it used to be.
That kind of guidance tends to shift demand at the margins rather than trigger a wave of returns, since most iPhone buyers never come close to saturating a terabyte of storage with the kind of continuous, cache-busting writes HOMOLAB used to produce its worst-case numbers. Still, the volume of coverage across Tom’s Hardware, AppleInsider, Notebookcheck, Sportskeeda Tech, and international outlets within days of launch signals a level of scrutiny that is unusual for a storage spec most users would never otherwise think to test.
Apple’s Silence on the Benchmark Numbers
As of this report, Apple has not issued a public technical statement addressing the QLC-versus-TLC split, the 1.1 MB/s figure, or the 38% sustained-write gap cited by Tom’s Hardware. The company’s own marketing for the iPhone 18 Pro Max has focused on camera and general performance claims rather than the underlying storage architecture. That is not unusual: component-level trade-offs like NAND flash type are rarely part of a keynote, and Apple’s engineering approach, running QLC in a TLC-like caching mode for the vast majority of everyday use, is designed precisely so that most users never have a reason to ask the question.
The gap between that design intent and this week’s headlines is the story itself. Apple built a system that hides QLC’s weaknesses for ordinary use, and independent lab testing found, and published, exactly where that system breaks down.
What Comes Next: Five Predictions
- Expect third-party lab tests to multiply over the coming weeks, with more outlets replicating HOMOLAB-style sustained-write benchmarks across different fill levels and storage tiers to see how consistent the 1.1 MB/s floor is across units.
- Buying guides will likely continue steering mainstream buyers toward the 256GB and 512GB TLC tiers, treating the 1TB and 2TB QLC models as a specialist option rather than a default upgrade path.
- Apple may address the concern indirectly through software, such as clearer on-device storage warnings or smarter background garbage collection to keep the SLC cache larger for longer, rather than through a public statement about NAND types.
- Professional video and photo workflows on 1TB and 2TB models will increasingly come with informal guidance to keep usable capacity below the 60% mark where the worst-case numbers appear, similar to the long-standing SSD industry practice of leaving free space for performance.
- Pressure from this cycle of coverage makes it more likely Apple reconsiders QLC for high-capacity tiers in a future generation, following the same trajectory PC SSD makers eventually took as QLC’s reputation among performance-sensitive buyers hardened.
How to Check If This Affects Your iPhone 18 Pro Max
Only the 1TB and 2TB iPhone 18 Pro and Pro Max configurations use QLC NAND; the 256GB and 512GB tiers use TLC and are not affected by this specific slowdown pattern. It is a storage-only issue, distinct from processor performance covered in reporting on the next-generation A20 Pro chip, and distinct from the design trade-offs Apple made on the foldable iPhone Duo launched earlier this cycle. If you own or are considering a 1TB or 2TB model, the practical takeaway from HOMOLAB’s data is straightforward: performance stays strong for the overwhelming majority of everyday tasks, and the worst-case numbers only appear during long, continuous, heavy write operations on a drive that is already well over half full. Keeping meaningful free space, and avoiding back-to-back large transfers such as full iCloud restores or multiple game installs in a single session, keeps the phone operating inside the fast SLC and TLC-mode buffer zones rather than the slow QLC floor.
Frequently Asked Questions
What is QLC NAND and how is it different from TLC NAND?
QLC (quad-level cell) NAND stores four bits of data per memory cell, versus three bits for TLC (triple-level cell). That extra bit allows higher storage density and lower manufacturing cost per gigabyte, but according to Techtimes and Telbb, it comes with slower sustained write speeds and lower write endurance than TLC.
Which iPhone 18 Pro Max storage tiers use QLC NAND?
Based on supply-chain reporting and subsequent benchmark coverage, the 1TB and 2TB iPhone 18 Pro and Pro Max models use QLC NAND, while the 256GB and 512GB tiers use TLC NAND.
How slow can iPhone 18 Pro Max storage actually get?
According to HOMOLAB’s testing, cited by Tom’s Hardware and AppleInsider, direct QLC write speeds on a 1TB unit filled to roughly 60% capacity can drop to a worst-case minimum of 1.1 MB/s, with an average around 45 MB/s once both cache layers are exhausted.
Will normal iPhone 18 Pro Max users notice the QLC slowdown?
Most users are unlikely to notice, since Apple runs the QLC storage in a caching mode that behaves like faster TLC storage for typical, short bursts of activity. The worst-case numbers only appear during long, sustained writes on a drive that is already well over half full.
Does the QLC slowdown affect ProRes video recording?
It can, in specific conditions. Long, continuous ProRes recording at 4K or above on a 1TB or 2TB unit that is already over 60% full risks running into the reduced write-speed ceiling, since ProRes bitrates can exceed the 45 MB/s average HOMOLAB recorded at that fill level.
Has Apple responded to the QLC storage benchmarks?
As of September 21, 2026, Apple has not issued a public technical statement specifically addressing the QLC-versus-TLC split or the benchmark figures reported by Tom’s Hardware, AppleInsider, and other outlets.
Should I buy the 512GB or 1TB iPhone 18 Pro Max?
That depends on your workflow. Buying guidance circulating since these benchmarks generally recommends the 512GB TLC-based tier for most buyers, reserving the 1TB and 2TB QLC tiers for users who need the extra capacity and can manage keeping usable space below the roughly 60% fill point where sustained-write performance drops most sharply.
Is this the first time a flagship phone has used QLC NAND storage?
QLC NAND has a longer history in PC and laptop SSDs, where it has typically been positioned as a budget or high-capacity option rather than a performance one, according to Techtimes and Telbb. Its use in a flagship-priced smartphone storage tier is the newer and more scrutinized part of this story.
Related
- iPhone 18 Pro Max QLC Storage Slower Than SD Card [2026]
- iPhone 18 Pro Max QLC Row Pushes Buyers to 512GB [2026]
- HBM4 Crunch: DRAM Inventory Falls Below 10 Days [2026]
- Corsair EX400U vs Samsung T9 vs SanDisk Extreme: 4x Speed Gap [2026]
- Apple A20 Pro Leaked Specs: 4.90GHz CPU, 7-Core GPU Beats M3 [2026]
Related Coverage
- iPhone 18 Pro Max QLC Row Pushes Buyers to 512GB [2026]
- iPhone 18 Pro Max QLC Storage Slower Than SD Card [2026]
- Arctis Nova Pro vs Stealth 600 vs BlackShark V3 Pro: $250 Gap [2026]
- A20 Pro Runs 27B AI Model at 2x A19 Pro Speed [2026]
- Eero Max 7 vs Orbi 970 vs Deco BE95: $600 Price Gap [2026]
