By using this site, you agree to the Privacy Policy and Terms of Use.
Accept

GadgetBond

  • Latest
  • How-to
  • Tech
    • AI
    • Amazon
    • Apple
    • CES
    • Computing
    • Creators
    • Google
    • Meta
    • Microsoft
    • Mobile
    • Samsung
    • Security
    • Xbox
  • Transportation
    • Audi
    • BMW
    • Cadillac
    • E-Bike
    • Ferrari
    • Ford
    • Honda Prelude
    • Lamborghini
    • McLaren W1
    • Mercedes
    • Porsche
    • Rivian
    • Tesla
  • Culture
    • Apple TV
    • Disney
    • Gaming
    • Hulu
    • Marvel
    • HBO Max
    • Netflix
    • Paramount
    • SHOWTIME
    • Star Wars
    • Streaming
Add GadgetBond as a preferred source to see more of our stories on Google.
Font ResizerAa
GadgetBondGadgetBond
  • Latest
  • Tech
  • AI
  • Deals
  • How-to
  • Apps
  • Mobile
  • Gaming
  • Streaming
  • Transportation
Search
  • Latest
  • Deals
  • How-to
  • Tech
    • Amazon
    • Apple
    • CES
    • Computing
    • Creators
    • Google
    • Meta
    • Microsoft
    • Mobile
    • Samsung
    • Security
    • Xbox
  • AI
    • Anthropic
    • ChatGPT
    • ChatGPT Atlas
    • Gemini AI (formerly Bard)
    • Google DeepMind
    • Grok AI
    • Meta AI
    • Microsoft Copilot
    • OpenAI
    • Perplexity
    • xAI
  • Transportation
    • Audi
    • BMW
    • Cadillac
    • E-Bike
    • Ferrari
    • Ford
    • Honda Prelude
    • Lamborghini
    • McLaren W1
    • Mercedes
    • Porsche
    • Rivian
    • Tesla
  • Culture
    • Apple TV
    • Disney
    • Gaming
    • Hulu
    • Marvel
    • HBO Max
    • Netflix
    • Paramount
    • SHOWTIME
    • Star Wars
    • Streaming
Follow US
AppleComputingMacTech

New “GoFetch” flaw breaks encryption on Apple’s M1, M2, M3 chips

Apple has a big problem: "GoFetch" is an unfixable flaw in M1, M2, and M3 silicon that allows extracting crypto keys, cracking encryption. The only available mitigations are software-based.

By
Shubham Sawarkar
Shubham Sawarkar's avatar
ByShubham Sawarkar
Editor-in-Chief
I’m a tech enthusiast who loves exploring gadgets, trends, and innovations. With certifications in CISCO Routing & Switching and Windows Server Administration, I bring a sharp...
Follow:
- Editor-in-Chief
Mar 22, 2024, 6:34 AM EDT
Share
We may get a commission from retail offers. Learn more
New "GoFetch" flaw breaks encryption on Apple's M1, M2, M3 chips
Photo by David Paul Morris/Bloomberg via Getty Images
SHARE

A team of researchers has uncovered a critical security vulnerability that affects Apple‘s custom M1, M2, and M3 processors, allowing malicious software to extract cryptographic keys and other sensitive information from the chip’s cache. Dubbed the “GoFetch” vulnerability, the flaw exploits a design oversight in Apple Silicon’s state-of-the-art data memory-dependent prefetcher (DMP), a cutting-edge feature shared with Intel’s latest Raptor Lake CPUs.

The DMP is designed to improve performance by intelligently loading data into the CPU’s cache before it is needed, leveraging advanced prediction algorithms. However, the researchers discovered that the prefetcher could be tricked into loading sensitive cryptographic key material into the cache, where it becomes vulnerable to theft by an attacker’s code.

At the heart of the issue is the DMP’s ability to fetch data based on pointer values, which are often used to access other memory locations. In a critical oversight, the prefetcher can sometimes confuse these pointers with the actual data they point to, resulting in the unintended loading of sensitive information into the cache.

This vulnerability completely undermines the security guarantees provided by constant-time programming, a widely used technique for mitigating side-channel attacks on encryption algorithms. By exploiting GoFetch, malicious applications can effectively bypass these protections, gaining access to even the strongest encryption keys, including those designed to withstand attacks from future quantum computers.

The implications of this flaw are far-reaching, as it affects a wide range of encryption algorithms and protocols used to secure everything from online communications to financial transactions and government secrets.

Disturbingly, the GoFetch vulnerability is baked into the silicon of Apple’s M-series chips, making it impossible to fully patch at the hardware level. The only recourse for Apple and its software partners is to implement software-based mitigations, which will inevitably come at the cost of reduced performance for encryption and decryption operations.

One potential workaround is to restrict encryption tasks to the M-series chips’ energy-efficient E-cores, which lack the vulnerable DMP. However, this approach also carries a significant performance penalty, as the E-cores are less powerful than the main performance cores.

Apple has reportedly incorporated a “switch” in the M3 chip to disable the DMP, but the performance impact of this mitigation remains unknown. It’s possible that turning off the prefetcher could result in performance losses on par with software-based mitigations.

Interestingly, Intel’s Raptor Lake CPUs, which share the same DMP design as Apple Silicon, do not appear to be affected by the GoFetch vulnerability. The reasons for this discrepancy are unclear, but it suggests that the flaw can be addressed at the hardware level, potentially in future iterations of Apple’s M-series chips.

For now, Apple has not provided any timeline for an official fix, but given the severity of the vulnerability, a patch or mitigation is expected within the year.

The GoFetch vulnerability was discovered by a team of researchers from several prestigious institutions, including the University of Illinois Urbana-Champaign, the University of Texas at Austin, the Georgia Institute of Technology, the University of California, Berkeley, the University of Washington, and Carnegie Mellon University.


Discover more from GadgetBond

Subscribe to get the latest posts sent to your email.

Topic:Apple M1Apple M1 MaxApple M1 ProApple M1 UltraApple M2Apple M2 MaxApple M2 ProApple M2 UltraApple M3 chipApple M3 Max chipApple M3 Pro chipApple silicon
Most Popular

The $19 Apple polishing cloth supports iPhone 17, Air, Pro, and 17e

Apple MacBook Neo: big power, surprising price, one clear target — Windows

Everything Nothing announced on March 5: Headphone (a), Phone (4a), and Phone (4a) Pro

OpenAI’s GPT-5.4 is coming — and it’s sooner than you think

MacBook Neo and external monitors: it’s complicated

Also Read
A simple illustration shows a large black computer mouse cursor pointing toward a white central hub with five connected nodes on an orange background.

Claude Marketplace lets you use one AI commitment across multiple tools

Perplexity Computer promotional banner featuring a glowing glass orb with a laptop icon floating above a field of wildflowers against a gray background, with the text "perplexity computer works" in the center and a vertical list of action words — sends, creates, schedules, researches, orchestrates, remembers, deploys, connects — displayed in fading gray text on the right side.

Perplexity Computer is the AI that actually does your work

99ONE Rogue 102321

99ONE Rogue wants to kill the ugly helmet comms box forever

TACT Dial 01 tactile desk instrument

TACT Dial 01: turn it, press it, focus — that’s literally it

Close-up of a person holding the Google Pixel 10 Pro Fold in Moonstone gray with both hands, rear-facing triple camera array and Google "G" logo prominently visible, worn against a silver knit top and blue jacket with a poolside background.

Pixel Care+ makes owning a Pixel a lot less scary — here’s why

Woman with blonde curly hair sitting outside in a lush park, holding a blue Google Pixel 10 and smiling at the screen.

Pixel 10a, Pixel 10, Pixel 10 Pro: one winner for every buyer

Google Search AI Mode showing Canvas in action, with a split-screen view of a conversational AI chat on the left and an "EE Opportunity Tracker" scholarship and grant tracking dashboard on the right, displaying a total funding secured amount of $5,000, scholarship cards with deadlines, and status labels including "To Apply" and "Awarded."

Google’s Canvas AI Mode rolls out to everyone in the U.S.

Google NotebookLM app listing on the Apple App Store displayed on an iPhone screen, showing the app icon, tagline "Understand anything," a Get button with In-App Purchases noted, 1.9K ratings, age rating 4+, and a chart ranking of No. 36 in Productivity.

NotebookLM Cinematic Video Overviews are live — here’s what’s new

Company Info
  • Homepage
  • Support my work
  • Latest stories
  • Company updates
  • GDB Recommends
  • Daily newsletters
  • About us
  • Contact us
  • Write for us
  • Editorial guidelines
Legal
  • Privacy Policy
  • Cookies Policy
  • Terms & Conditions
  • DMCA
  • Disclaimer
  • Accessibility Policy
  • Security Policy
  • Do Not Sell or Share My Personal Information
Socials
Follow US

Disclosure: We love the products we feature and hope you’ll love them too. If you purchase through a link on our site, we may receive compensation at no additional cost to you. Read our ethics statement. Please note that pricing and availability are subject to change.

Copyright © 2026 GadgetBond. All Rights Reserved. Use of this site constitutes acceptance of our Terms of Use and Privacy Policy | Do Not Sell/Share My Personal Information.