Apple Patents Technology to Stop Hackers From Corrupting Your Computer's Memory
There's a well-known hardware attack that can corrupt or steal data by hammering the same memory location thousands of times per second. Apple's latest patent aims to make that attack essentially impossible by scrambling the address before the memory ever sees it.
How Apple's memory address encryption stops a silent attack
You're using your laptop or phone and everything looks fine, but deep inside the chip, a program is hitting the same tiny spot in memory over and over, thousands of times a second. That repeated pounding can cause nearby memory cells to flip their values, a trick attackers use to change data or break into secured areas without ever touching them directly. It's called a row hammer attack, and it's been a real-world threat for over a decade.
Apple's approach is to hide which physical memory location a program is actually using. When your device starts up, it generates a secret encryption key. Every time a program asks to read or write memory, the chip encrypts the address before sending it to the memory hardware. So even if malicious code tries to hammer a specific address, it hits a scrambled location that changes with each boot.
The result is that the physical row of memory cells that gets activated is essentially unpredictable to anyone without the key. No predictable target, no hammer attack.
… an encryption circuit configured to: obtain a particular one of the first set of cryptographic keys from the cryptographic key source circuit; receive the memory address of the given memory transaction; and encrypt, based on the particular cryptographic key, a portion of the memory address to form an encrypted memory address …
Translation: The system uses secret keys to scramble the digital addresses where your computer stores its sensitive information.
How the encryption circuit scrambles and redirects memory access
The patent describes a hardware system built into a chip that intercepts memory addresses before they reach the physical memory chips.
Here's the chain:
- At startup, a cryptographic key source circuit generates a set of secret keys (think of them as one-time padlocks created fresh every boot).
- Whenever a processor, GPU, or other component issues a memory request, an encryption circuit grabs one of those keys and encrypts part of the memory address.
- The encrypted address is what actually gets sent to the memory hardware, not the original address the program asked for.
- The memory system stores and retrieves data based on the encrypted address, so reads and writes still work correctly from the program's perspective.
The key insight is that a row hammer attack (where an attacker deliberately reads the same memory row thousands of times to flip bits in neighboring rows) depends on knowing which physical row to target. By encrypting the address mapping, Apple breaks that assumption: the attacker sees logical addresses, but the physical rows being activated are scrambled.
Because the keys are generated fresh at initialization, the mapping changes every time the device reboots. An attack pattern that worked one session would hit a completely different set of physical locations the next time.
The encrypted memory address corresponds to the row that is activated for the memory transaction, instead of the memory address generated by the source circuit.
Translation: Instead of using the real location of data, the computer points to a fake, encrypted address to confuse potential hackers.
What this means for the security of Apple's chips
Row hammer attacks are not theoretical. Researchers have used them in the real world to escalate privileges, break out of sandboxes, and compromise virtualized servers. They work because modern memory chips pack cells so tightly that electrical interference between rows is physically unavoidable. Most defenses so far have lived in software or memory firmware, which adds latency or depends on manufacturers updating their chips. A defense baked directly into the address path of Apple's own silicon sidesteps those dependencies entirely.
For you as a user, the payoff is the kind of attack you would never see coming and could never patch your way out of becomes much harder to pull off. Apple has been designing its own chips (the A-series in iPhones and M-series in Macs) for years, and this patent fits the pattern of using that control over hardware to close gaps that generic chips leave open. It sits alongside the broader wave of silicon-level security work that Big Tech patent news on chips and memory protection has been tracking across multiple companies.
Apple's 356th filing in our Apple coverage since May follows applications like the sliding charging-port cover and the predictive video pixel system.
Row hammer attacks work by hammering the same memory location repeatedly until neighboring data flips, silently corrupting files or handing a bad actor control of the machine. A user would never see it coming. Apple's approach scrambles the address each transaction uses, so an attacker cannot predict which physical location to target.
You would never notice this protection working. No slowdown, no prompt, no setting to enable. The benefit arrives entirely as an absence of something bad that would otherwise have been invisible anyway.
That quiet reliability is exactly what makes it matter. Silent data corruption and silent takeovers are the failures people discover only after serious damage is done, and a defense baked into the hardware Apple designs itself is far harder for an attacker to route around than anything running in software.
There are more where this came from
We read every patent application Big Tech publishes and send you the ones worth knowing. Plain English, free, every week.
The drawings
11 drawing sheets from US 2026/0252497 A1 · click any drawing to enlarge
Want this weekly breakdown for a company we don't cover? Patentlyze Pro →