A storage test published today revealed that the 1TB iPhone 18 Pro Max can write data slower than a cheap microSD card once its cache runs dry. Testing lab HOMOLAB pushed the phone through a sustained, heavy-load write test and watched throughput collapse from a burst of over 3,000 MB/s down to an average of 79.4 MB/s, with speeds dipping as low as 25.6 MB/s during the worst stretches. For a phone that starts at $1,899 in its 1TB configuration, that is not a small footnote. It is the first hard data confirming what supply-chain leaks warned about back in July: Apple swapped in QLC NAND flash for its highest-capacity iPhone 18 Pro models, and under load, that decision has a real cost.
This matters beyond spec-sheet trivia. The iPhone 18 Pro Max is Apple’s flagship camera phone, marketed heavily around ProRes video capture and long-form 4K recording, the exact workloads that hammer storage the hardest. If sustained writes really do fall into double-digit megabyte-per-second territory during a long recording session, that is a functional problem for the people paying the most for the biggest phone in the lineup.
What HOMOLAB Found in the 1TB iPhone 18 Pro Max Test
HOMOLAB’s methodology was straightforward: hammer the drive with continuous writes long enough to exhaust every layer of caching Apple builds into the storage stack, then measure what is left. The lab ran the test against a 1TB iPhone 18 Pro Max and, for reference, against a TLC-equipped comparison unit and Sony’s Xperia 1 V (256GB), according to a report on the results carried by Sina Finance. In a weighted 4K read benchmark, the QLC-based 1TB unit scored 46,045 points against 51,282 for the TLC comparison device, a modest but measurable gap. The real story showed up in mixed read/write workloads: at low queue depth, the TLC unit led by roughly 38%, and even under high-load conditions that had narrowed only to about 12%.
None of this is exotic lab trickery. It is the kind of sustained-write stress that happens the moment someone records a long video, restores a large backup, or transfers a big media library onto the phone. Once the fast caching layers run out, the phone has nowhere left to hide the underlying NAND’s limits.
The Numbers: SLC Cache, TLC Buffer, and Raw QLC Speeds
Apple’s storage stack on the 1TB iPhone 18 Pro Max runs through three effective tiers before hitting raw QLC. First comes a fast SLC cache for burst writes, then a simulated TLC secondary buffer, and only after both are saturated does data land directly on QLC cells. HOMOLAB’s numbers trace that collapse stage by stage.
Peak Speed While Cache Is Available
With the SLC cache intact and the drive near-empty, HOMOLAB recorded peak write speeds above 3,000 MB/s, in line with what you would expect from a UFS 4.0-class flagship during a short burst. That number is genuinely fast, and it is the number most casual buyers will experience most of the time, since most phone writes are short bursts rather than sustained streams.
What Happens Once the Cache Runs Out
The burst speed only lasts as long as the cache does. Once it is exhausted on a near-empty 1TB unit, raw QLC writes averaged 79.4 MB/s, with a floor of 25.6 MB/s, according to HOMOLAB’s results. That average already sits below the write speeds of many current mid-range microSD cards, and the floor is worse than some of the slowest cards sold today.
| Test Stage | Drive Fill Level | Active Storage Layer | Avg. Write Speed | Min. Write Speed |
|---|---|---|---|---|
| Burst writes | Near-empty | SLC cache | > 3,000 MB/s | — |
| Available SLC cache | Near-empty | SLC cache capacity | ~250 GB | — |
| Cache exhausted | Near-empty | Raw QLC (direct) | 79.4 MB/s | 25.6 MB/s |
| Buffered writes | ~60% full | Simulated TLC buffer | 396 MB/s | — |
| Available SLC cache | ~60% full | SLC cache capacity | ~58 GB | — |
| Cache exhausted | ~60% full | Raw QLC (direct) | 45 MB/s | 1.1 MB/s |
HOMOLAB sustained-write test, iPhone 18 Pro Max 1TB (near-empty drive)
Stage 1, SLC cache burst: > 3,000 MB/s
Stage 2, raw QLC after cache exhausted:
average: 79.4 MB/s
minimum: 25.6 MB/s
HOMOLAB sustained-write test, iPhone 18 Pro Max 1TB (~60% full)
SLC cache capacity available: 58 GB (down from ~250 GB near-empty)
Simulated TLC buffer, average: 396 MB/s
Raw QLC after cache exhausted:
average: 45 MB/s
minimum: 1.1 MB/sWhy Storage Gets Slower as You Fill It Up
The near-empty numbers are the best-case scenario, and they still are not flattering. The picture gets worse as the drive fills, because the SLC cache Apple reserves for fast writes is not a fixed slice of the drive. It shrinks as usable capacity fills up. HOMOLAB measured available SLC cache dropping from roughly 250 GB on a nearly empty 1TB drive down to just 58 GB once the phone reaches about 60% capacity.
That shrinking cache changes everything downstream. At 60% fill, the simulated TLC buffer averaged 396 MB/s, itself a step down from the emptier-drive figures. Once that buffer saturates too, raw QLC writes averaged just 45 MB/s, with a measured minimum of 1.1 MB/s in the worst stretches of the test. A phone that cost close to $2,000 briefly writing data at a rate slower than a USB 2.0 flash drive from a decade ago is the kind of result that is easy to reproduce and hard to explain away.
Practically, this means the buyers most likely to fill a 1TB phone, people who shoot a lot of video or carry large offline media libraries, are also the buyers most likely to run into the slowest tier of performance. The people paying the most for capacity get the least consistent throughput once that capacity starts filling up.
QLC vs TLC NAND: What’s Actually Different
NAND flash stores data by packing electrical charge into cells, and the difference between QLC and TLC comes down to how much each cell is asked to hold. TLC (triple-level cell) NAND stores 3 bits per cell. QLC (quad-level cell) stores 4. That extra bit means more data fits in the same physical space, which is exactly why QLC is cheaper per gigabyte to manufacture, per Micron’s overview of NAND flash technology. The tradeoff is that packing more states into each cell makes those states harder to read and write reliably, which is why QLC has historically shown slower sustained writes and lower write endurance than TLC.
Apple has tried to paper over that gap with software. Reports describe Apple running the QLC cells in a TLC-like access pattern and layering a large dynamic SLC cache on top, an approach that gets the phone to the same headline burst speeds as a TLC device in short tests. The catch is that a dynamic cache can only mask the underlying media for so long. Sustained, heavy writes are exactly the condition that strips that mask away, which is what HOMOLAB’s test demonstrated directly.
| iPhone 18 Pro Max Storage Tier | NAND Type | Launch Price (US) |
|---|---|---|
| 256GB | TLC | $1,299 |
| 512GB | TLC | $1,499 |
| 1TB | Mixed TLC/QLC | $1,899 |
| 2TB | QLC (SK hynix enterprise-grade) | $2,499 |
Pricing above comes from PCMag’s iPhone 18 Pro Max review. The jump from 512GB to 1TB costs an extra $400, and that upgrade is also the point where the storage architecture quietly changes underneath the buyer.
Which iPhone 18 Pro Models Are Actually Affected
The slowdown is not universal across the iPhone 18 Pro lineup. Apple’s own published specifications list 256GB, 512GB, 1TB, and 2TB as the four storage tiers, without disclosing NAND type. The 256GB and 512GB tiers continue to use TLC NAND, which does not exhibit the same collapse under sustained writes. It is specifically the 1TB and 2TB tiers, marketed as the premium storage options for power users, where QLC shows up. According to reporting on the supply chain, the 1TB models use a mix of TLC and QLC flash, while the 2TB variant leans on enterprise-grade QLC supplied by SK hynix.
Supply-chain accounts differ on the exact part numbers involved. Some outlets, including Japan’s Gazlog and China’s Sina.cn, have pointed to Kioxia’s 218-layer BiCS8 QLC as the chip found in teardown analysis of early 1TB units, with Samsung’s 3DV8 TLC cited elsewhere as a secondary, less common source mixed into the same capacity tier. That kind of multi-sourcing is normal in high-volume phone manufacturing, but it also means two buyers of the same 1TB model, bought in different production runs, could plausibly own phones with meaningfully different sustained write behavior.
How This Compares to a Cheap microSD Card
The comparison to microSD cards is not a throwaway line, it is the core of why this story is spreading. Even budget microSD cards rated for basic video recording typically sustain write speeds in the 20 to 90 MB/s range depending on speed class, and better cards clear 100 MB/s or more without much effort. HOMOLAB’s 79.4 MB/s average and 25.6 MB/s floor on a near-empty 1TB iPhone 18 Pro Max sit right in that same band, and the 45 MB/s average with a 1.1 MB/s floor at 60% capacity falls well below it.
That is a striking result for a phone with no card slot at all, built entirely around the idea that internal storage should outperform any removable card you could buy separately. It is also a reminder that headline burst numbers, the ones usually quoted in marketing and short synthetic benchmarks, do not tell the full story once a device is asked to sustain heavy writes for more than a few seconds.
The Supply Chain Behind Apple’s QLC Decision
Apple did not stumble into this. Supply-chain reporting on the iPhone 18 Pro’s storage flagged the TLC-to-QLC shift months before launch. WCCFTech reported in July that Apple was swapping in QLC for the 1TB and 2TB models while keeping list prices high, calling it “a stealth downgrade, especially as QLC NAND typically entails more erratic and slower write speeds.” The economics are simple enough: QLC packs more bits per cell, so Apple gets more storage capacity out of the same die area and pays less per gigabyte to its suppliers, while charging the buyer the same premium the high-capacity tier has always carried.
That tradeoff lands at a difficult moment for the broader memory market. NAND and DRAM prices have both climbed through 2026, a trend covered in shattered.io’s reporting on the memory chip shortage, and it has pushed device makers across the industry toward exactly this kind of cost-saving substitution, as seen in coverage of memory costs eating into budget phone and laptop margins. Using cheaper QLC in the highest-capacity, highest-margin iPhone tier lets Apple protect its bill of materials without touching the sticker price, but it does so by quietly changing what “1TB iPhone” actually means on the inside.
Historical Context: Apple Has Managed NAND Tiers Before
This is not the first time Apple has varied storage hardware across otherwise identical-looking SKUs. Base and mid-tier iPhones have long used slower NAND controllers and fewer parallel flash dies than Pro models, a difference most buyers never notice because everyday app launches and photo saves rarely stress the drive hard enough to expose it. What is different this time is the direction of the change. Historically, storage differences tracked price tier, cheaper phones got slower storage. With the 1TB and 2TB iPhone 18 Pro Max, the most expensive configurations are the ones carrying the compromise, which inverts the usual logic buyers use when deciding whether to pay up for more capacity.
The broader NAND industry has also been sliding toward QLC for years as SSD makers chased capacity and cost, first in bulk data-center drives, then in consumer laptops, and now in flagship phones. Apple adopting QLC in the iPhone is less a surprise and more a milestone in that longer transition, one that arrived at a phone price point high enough to draw real scrutiny.
Market Impact: What This Means for Apple and Buyers
The immediate impact is reputational rather than financial. Apple has not commented publicly on the HOMOLAB results as of this writing, and there is no indication of a recall, refund program, or software fix in the works. What has changed is the burden of proof for the 1TB and 2TB tiers. Reviewers and prospective buyers now have hard sustained-write numbers to point to, rather than relying on supply-chain rumor, and that data is unflattering enough that it will likely shape purchasing advice going into the holiday buying season.
For Apple, the calculus is about margin protection during a memory-constrained market. Component costs for NAND and DRAM have risen sharply industry-wide in 2026, squeezing every phone and laptop maker’s bill of materials, a dynamic Intel’s own leadership has warned will not ease before 2028. Choosing QLC for the priciest storage tier is a defensible cost decision on paper. Whether it survives customer scrutiny once it is this well documented is a separate question, and one that will play out in reviews and returns over the next few months rather than in a single news cycle.
Competitive Comparison: How Rival Flagships Handle High-Capacity Storage
Android flagship makers have faced the same NAND cost pressure Apple is navigating, and most have so far kept TLC NAND across their storage tiers rather than mixing in QLC at the high end, according to device teardown trackers that have followed the broader QLC-in-mobile trend. That is not a guarantee those brands will hold the line forever, given that NAND pricing pressure is industry-wide, not Apple-specific. It does mean that, for now, a buyer chasing the fastest sustained storage performance in a flagship phone has reason to look past capacity alone and ask what flash type sits behind that number, something that was rarely a meaningful question in phones before this generation.
The comparison also cuts across product categories. Apple’s own Apple Silicon Macs and iPads still generally use TLC-class storage in their higher tiers, which is part of why the iPhone 18 Pro Max result reads as an outlier within Apple’s own lineup rather than a company-wide policy shift, at least for now.
What It Means If You Shoot ProRes or Long 4K Video
The iPhone 18 Pro Max is sold heavily on its camera and its ability to record ProRes and high-bitrate 4K footage directly to internal storage, workloads that involve exactly the kind of sustained, continuous writes HOMOLAB’s test simulated. A dropped frame or a recording that stutters mid-take because the drive fell into its slowest QLC tier is a materially different problem than a slightly lower synthetic benchmark score. Reporting from techpulse.press has also connected the QLC tiers to worse battery life during sustained ProRes recording, estimating a 10 to 15% hit compared with the TLC-equipped 512GB model, attributed to storage throttling and higher controller power draw once the cache is under stress.
For professional and semi-professional video shooters, who are exactly the audience Apple targets with the Pro Max branding and its ProRes feature set, this is the scenario where the HOMOLAB numbers stop being an abstract benchmark and start being a workflow risk.
Predictions: What Happens Next
- Expect independent reviewers and YouTube teardown channels to replicate HOMOLAB’s sustained-write methodology on their own 1TB and 2TB units within the next few weeks, since the test is simple to reproduce and the results are dramatic enough to draw an audience.
- Expect Apple to stay quiet publicly rather than issue a detailed rebuttal, following the pattern it has used with prior component controversies, unless the story escalates into a formal complaint or regulatory inquiry.
- Expect storage-tier buying advice to shift, with more outlets recommending the 512GB TLC tier over the 1TB QLC tier for buyers who record a lot of video, even though 1TB nominally offers more room.
- Expect the QLC-vs-TLC question to become a standard line item in flagship phone reviews going forward, the way RAM and chipset generation already are, now that a documented real-world performance gap exists.
- Expect other phone makers to watch the backlash closely before deciding whether to follow Apple into QLC at their own high-capacity tiers, given how much attention this test has drawn less than three weeks after the iPhone 18 Pro Max went on sale.
How to Check Your Own iPhone 18 Pro Max Storage Type
Apple does not expose a simple menu toggle showing whether a given unit’s storage is TLC or QLC, so there is no fully reliable way for an average owner to check this from Settings alone. There are still a few practical steps worth taking if you are concerned about sustained write performance on a 1TB or 2TB unit.
- Record a long continuous 4K or ProRes video, 10 minutes or more, and watch for stutters, dropped frames, or a recording warning near the end, which can indicate the drive has fallen into its slowest write tier.
- Keep meaningful free space on a 1TB or 2TB unit rather than filling it close to capacity, since HOMOLAB’s results show the available SLC cache shrinks substantially as the drive fills.
- Run a large file transfer, such as restoring several gigabytes of photos from a backup, and time how long it takes relative to the file size to get a rough sense of real-world sustained speed.
- Watch for aftermarket benchmarking apps on the App Store that specifically test sustained sequential writes rather than short bursts, since burst-only tests will not reveal this issue.
- If performance concerns you and you are still within Apple’s return window, consider the 512GB TLC tier instead, especially if your workflow involves frequent long-form video capture.
Frequently Asked Questions
What did the HOMOLAB test actually find?
HOMOLAB tested a 1TB iPhone 18 Pro Max under sustained, heavy write load and found that once the phone’s cache layers are exhausted, raw QLC write speeds averaged 79.4 MB/s on a near-empty drive, with a minimum of 25.6 MB/s, and fell further to a 45 MB/s average with a 1.1 MB/s minimum once the drive reached about 60% capacity.
Which iPhone 18 Pro Max storage tiers use QLC NAND?
Reports point to the 1TB tier using a mix of TLC and QLC flash, while the 2TB tier uses enterprise-grade QLC supplied by SK hynix. The 256GB and 512GB tiers continue to use TLC NAND.
Is QLC storage actually slower than a microSD card?
Under sustained, cache-exhausted writes, yes, according to HOMOLAB’s numbers, which land in or below the range of many entry-level microSD cards. Under short bursts, which cover most everyday phone use, the SLC cache keeps speeds well above what any microSD card can sustain.
Will this affect normal, everyday iPhone use?
Most everyday tasks, like opening apps, saving photos, and installing small updates, are short bursts that stay within the fast SLC cache and are unlikely to expose the slowdown. The risk grows with sustained workloads such as long video recording, large file transfers, or restoring big backups, especially as the drive fills up.
Does QLC storage wear out faster than TLC?
QLC NAND generally has lower write endurance than TLC because it packs more bits into each cell, per Micron’s technical overview of NAND flash. Apple’s dynamic caching and controller optimizations are designed to offset some of that wear, but the underlying physical limitation remains.
Should I avoid buying the 1TB or 2TB iPhone 18 Pro Max?
That depends on your workload. If you frequently shoot long-form 4K or ProRes video, transfer large media libraries, or expect to fill the drive close to capacity, the documented sustained-write slowdown is worth factoring into your decision, and the 512GB TLC tier may be a safer choice. For buyers whose usage stays mostly in short bursts, the practical difference is likely to be far smaller.
Has Apple responded to the HOMOLAB results?
As of this writing, Apple has not issued a public response to the HOMOLAB test results. There is no confirmed recall, refund program, or firmware fix tied to this issue.
How does this compare to storage on the iPhone 18 Pro (non-Max) or other Apple devices?
Supply-chain reporting indicates the same 1TB and 2TB QLC storage arrangement applies across the iPhone 18 Pro lineup, not just the Pro Max, since both share the same internal storage options. Apple’s other product lines, including current Apple Silicon Macs, such as those benchmarked in recent M-series GPU testing and chips covered in A20 Pro benchmark leaks, have generally continued to rely on TLC-class storage at their higher capacity tiers, at least for now.
