The MacBook Air receives its first major chip upgrade in nearly 19 months, moving from the M2 to the new M5.
According to Apple’s latest testing data, the M5-powered MacBook Air pushes the boundaries of performance and efficiency. AI workloads are processed up to 4 times faster than the previous M4 model, while overall performance speeds show an impressive 9.5-fold increase compared to the original M1 model. These numbers suggest a significant leap forward in capabilities for both everyday users and professionals requiring high processing power.
Storage capabilities have also seen significant enhancements. The base model now starts at 512GB, alongside a promised 2x increase in read/write performance over the previous generation.
Despite these advancements in processing and storage, Apple’s latest pricing strategy has drawn criticism from industry observers. While the M4 MBA received a price cut, dropping to $999, the M5 model starts at $1,099.
This $100 price increase raises questions about the value proposition of the M5 MBA, particularly given the impressive performance capabilities of the more affordable M4 model. While the M5 offers clear performance upgrades, the premium price tag may deter some potential buyers.
As Apple continues to push the boundaries of technology with the M5 MacBook Air, the debate over its value in the market is sure to continue. Whether the performance benefits justify the increased cost will ultimately depend on individual user needs and priorities.
Hold on: WHO actually needs 4x faster AI processing?
Let’s start with the number Apple most wants you to repeat at dinner parties: AI workloads processed 4x faster than the M4. Sounds enormous. But I noticed, reading through Apple’s methodology footnotes on previous chip launches, that “AI workloads” in their testing environment tends to mean highly optimized benchmark tasks, image segmentation, on-device translation — not the chaotic, real-world mix of browser tabs, Slack notifications, and a half-finished Figma file that describes how most people actually use a laptop. The 4x figure is a ceiling, not a floor. Treat it accordingly.
The 9.5x performance improvement over M1 is genuinely frustrating to parse. That comparison is doing a lot of heavy lifting, it spans nearly four chip generations and conveniently makes the M5 sound more dramatic than a straight M4-to-M5 comparison warrants. Honestly, if your M1 MacBook Air is still running fine, this number is marketing math dressed up as engineering progress.
During our testing of the M4 Air last cycle, the storage read/write gains Apple promised were real; in sustained synthetic benchmarks. Real. Not in typical user sessions involving lots of small file writes, where thermal throttling on a fanless chassis quietly eats your gains like a memory leak nobody bothered to fix.
Here’s the counter-argument nobody is resolving: a Dell XPS 13 with Snapdragon X Elite starts at $999 and delivers competitive AI inference performance on Qualcomm’s NPU, with a fan that actually lets it sustain those speeds. Fanless is elegant. It is also a thermal constraint that physics doesn’t negotiate with.
Genuine doubt, stated plainly: I am not confident the 512GB base storage bump meaningfully changes the purchase calculus for users who were already buying the 256GB M4 at $999 and upgrading storage separately. The math might still favor the older machine depending on your configuration.
Apple walked the M4 Air’s price down to $999, then immediately walked the M5 back up $100. That’s not a value improvement. That’s a price reset wearing a spec sheet as a disguise.
So here’s the rhetorical question that should follow you to the checkout page: if the M4 was good enough to justify a price cut, what exactly changed in 19 months that now justifies paying more for the privilege of owning the newer one?
M5 MacBook air: the $100 premium nobody asked for, examined without mercy
Here is the uncomfortable arithmetic Apple hopes you skip: the M5 MacBook Air costs $1,099, exactly $100 more than the M4 model that Apple itself validated by cutting to $999. A price cut is a confession. It says: this is what the machine is worth. Walking that back 19 months later with a new chip number is a pricing maneuver, not a value argument.
Now bind the jargon to reality. The 4x AI processing gain over M4 sounds transformative until you ask what workload you are actually running. Apple’s benchmark methodology on previous chip launches has consistently optimized for controlled, single-task inference, image segmentation, on-device translation – not the entropy of real usage. In practice, that 4x ceiling collapses fast under a browser with 22 tabs, a video call, and background syncing. The number is real in a lab. Labs are not your desk.
The 9.5x performance improvement over M1 is the stat doing the heaviest lifting here. That gap spans nearly four chip generations and close to four years of silicon development. Comparing M5 to M1 is designed to make the delta feel urgent. It isn’t — not if your M1 machine still closes Figma files without complaint.
Storage read/write performance — promised at 2x faster than the previous generation; is where the fanless chassis becomes a genuine technical liability. When actually tested under sustained write loads, a passive cooling system with no fan imposes a hard thermal ceiling. Physics does not issue firmware patches. The 512GB base storage is a meaningful upgrade from 256GB, but 2x read/write gains in synthetic benchmarks do not survive sustained real-world sessions on a machine that manages heat through aluminum and prayer.
From what I’ve seen across multiple fanless laptop generations, sustained storage throughput on passive-cooled machines degrades measurably under continuous load, the kind of load that matters for video editors and developers who were the implicit audience for those 2x storage numbers. The M5 Air will throttle. The only question is when.
The competitive pressure matters here too. A Dell XPS 13 at $999 with Snapdragon X Elite offers active cooling, meaning its AI inference performance, competitive on Qualcomm’s NPU — can actually sustain. The M5’s 4x AI workload advantage over M4 is real at peak. Sustained That gap narrows every time the aluminum gets warm.
The conditional recommendation: Buy the M5 Air at $1,099 if you are upgrading from an M1 machine, where that 9.5x generational gap is genuine and felt. If you own an M4 and were waiting for a reason to upgrade, the 4x AI figure does not constitute one for most workflows. And if you are buying new on a budget, the M4 at $999 remains defensible until Apple discontinues it.
Watch for two things: whether Apple’s M4 pricing holds after M5 launch volumes stabilize, and whether third-party benchmarks confirm the 2x storage performance figure survives sustained thermal load, or quietly disappears into the footnotes like it has before.
Is the $100 price increase over the M4 justified by the performance gains?
That depends entirely on your workload. The M5 delivers AI processing 4x faster than the M4 under Apple’s benchmark conditions, and overall performance is 9.5x faster than the M1. But if you are not running sustained AI inference tasks, those numbers do not translate into a felt $100 difference at your desk.
Should M1 MacBook air owners upgrade to the M5?
This is the one case where the data actually supports upgrading. The 9.5x performance gap between M5 and M1 reflects nearly four generations of silicon development, and the base storage jumping to 512GB from what M1 owners typically configured is a real daily improvement. The $1,099 price is easier to swallow when the machine you are replacing is genuinely long in the tooth.
Does the fanless design hurt real-world storage performance on the M5?
Apple promises 2x read/write performance over the previous generation, and synthetic benchmarks will likely confirm that figure. The problem is that a passive-cooled chassis with no fan cannot sustain peak throughput indefinitely, thermal constraints are a physics problem, not a software one, and they affect every fanless machine regardless of chip generation.
How does the M5 air compare to windows alternatives at the same price?
The most direct competitor, the Dell XPS 13 with Snapdragon X Elite, starts at $999; a full $100 less than the M5 Air. The Snapdragon machine includes active cooling, which allows it to sustain AI inference speeds that a fanless design cannot maintain under prolonged load. The M5’s 4x AI advantage over M4 is a peak number; sustained comparisons against actively cooled Windows machines will be narrower.
Is the M4 MacBook air still worth buying now that the M5 exists?
At $999, the M4 Air remains a technically sound machine, and Apple’s own pricing history validated that number when they cut to it. The M5’s improvements – particularly the 4x AI processing gain and doubled storage baseline, are real, but not universally relevant. If the M4 is still available at $999, it represents better value for anyone whose workload does not specifically stress AI inference or sustained storage throughput.
Compiled from multiple sources and direct observation. Editorial perspective reflects our independent analysis.
