AMD Quietly Stripped Memory Encryption from Consumer CPUs, Users Push Back

AMD has quietly removed Transparent Secure Memory Encryption (TSME) from its consumer CPU lineup, according to Ars Technica, triggering a wave of complaints from users and privacy advocates who had come to depend on the feature as a hardware-level safeguard against cold-boot attacks and physical memory exploits.

How the Removal Was Discovered

The removal came to light in April 2026 when Ben Kilpatrick, a privacy-conscious Linux user, was installing a new operating system on a machine running an AMD Ryzen 7 9700X processor from the Zen 5 architecture. Kilpatrick ran Host Security ID (HSI), an auditing tool that evaluates firmware and hardware security configurations, and found that TSME was listed as "not supported"—despite having enabled it in BIOS settings previously. The same HSI output showed that TSME had previously displayed as "encrypted" on the same machine, indicating that the feature had been present and functional on an earlier firmware version.

Kilpatrick's investigation revealed a critical technical detail: when using an older firmware version available through AMD's Generic Encapsulated Software Architecture (AGESA), consumer-grade Ryzen CPUs on MSI and Gigabyte motherboards had TSME enabled. However, when a newer AGESA version—specifically 1.2.7.0—ran instead, TSME showed as "not supported". PRO versions of the same Ryzen CPU, by contrast, supported TSME across both older and newer motherboard firmware versions.

After MSI engineers conducted controlled testing, they provided Kilpatrick with critical findings. Testing on an Asus X870E motherboard showed tsme_status = 1 on a PRO processor and tsme_status = 0 on a consumer Ryzen 9800X3D with the same board and BIOS. Additionally, memory capture analysis from the AMD Boot Loader (ABL) showed that an internal AGESA flag called DfIsTsmeEnabled returned FALSE for consumer processors regardless of whether TSME was set to AUTO or ENABLED in BIOS, while the same flag returned TRUE for PRO and EPYC processors when TSME was enabled.

AMD's Opacity and Contradictory History

Throughout Kilpatrick's investigation, AMD declined to provide clear explanations. AMD declined to answer questions sent by email other than to say TSME "is a security feature only applied to PRO CPUs as part of AMD PRO Technologies," marking the first known time the chipmaker explicitly made this restriction public. Two AMD engineers—Tom Lendacky, an AMD fellow software engineer, and Mario Limonciello, AMD senior principal software engineer and maintainer of the fwupd version of HSI—initially suggested troubleshooting steps but ultimately provided no substantive answer.

The company's current position contradicts its own historical record. In 2020, Tom Lendacky had confirmed that the Ryzen 3700X, a consumer CPU, "should support TSME". In a 2025 comment, Lendacky had recommended using TSME on consumer chips, stating "I recommend using TSME (Transparent SME), but it is a BIOS option that needs to be exposed by your BIOS provider". The feature worked reliably on consumer Ryzen chips for years, conditioning users to treat it as a standard part of the hardware package.

When Kilpatrick reported MSI's findings back to AMD's engineers and asked directly whether the DfIsTsmeEnabled flag being set to FALSE on consumer SKUs was "a silicon-level limitation, or is it a firmware policy decision within AGESA," Mario Limonciello promptly replied: "My apologies; but I don't have any more information to share on this topic", ending the discussion entirely.

The Intentional-vs.-Accidental Question

Kilpatrick frames the core ambiguity: "The big outstanding question is whether this is a deliberate policy decision by AMD to restrict TSME to PRO chips, or an unintentional regression that was introduced in AGESA 1.2.7.0". He noted: "The reason that distinction matters is that if it is deliberate policy, AMD made a conscious decision to remove a working feature from consumer hardware and restrict it to enterprise customers. If it is an accidental regression, it is a firmware bug that AMD should fix. Either way the silicon is capable, either way the change happened in AGESA, and either way AMD has declined to explain it".

The technical facts suggest deliberation rather than accident. The DfIsTsmeEnabled flag shows different states depending on whether the CPU is classified as consumer or PRO—a distinction enforced at the firmware level rather than dictated by silicon. If the removal were a firmware regression, both consumer and PRO versions would likely show the same flag state. The targeted nature of the change—affecting consumer SKUs while leaving PRO SKUs intact—indicates that someone made a specific choice to restrict the feature.

Security and Market Implications

TSME encrypts the entire contents stored in memory, protecting against cold boot attacks and other types of physical exploits that siphon sensitive data out of connected memory chips. When active, it provides protection against physical attacks, including cold boot exploits, DRAM interface snooping, and memory module removal, and it activates silently when enabled in the BIOS, making it the more practically useful of the two protections.

The removal represents a security regression for millions of users. Silicon-level security expert Joe FitzPatrick noted the gravity of AMD's position: "They could have not realized they did it leading to their cagey responses, or they could have done it intentionally and tried to get away with it, leading to the same cagey responses. But I really feel like an explanation should be in order, even if it was 'TSME was never supposed to be supported. We did ship some firmwares that erroneously enabled it, but you shouldn't use them since we can't guarantee it'll work properly'".

The episode illustrates a deeper tension in chip design: which security features justify the silicon area, power budget, and validation cost required to ship them at scale. Memory encryption is not a zero-cost addition—it requires dedicated cryptographic logic, adds latency to memory transactions, and demands substantial firmware and driver support across a fragmented ecosystem of motherboard vendors and operating systems. The cost-benefit calculation differs sharply between enterprise and consumer segments. AMD's EPYC server processors retain full memory encryption support, reflecting the market reality that enterprise buyers treat security features as table-stakes for data centre procurement, while consumer buyers prioritize cost and performance per dollar.

The bifurcation has real consequences. Home users and small businesses running consumer chips in contexts that would benefit from hardware-enforced memory protection—such as privacy-focused Linux systems or machines handling sensitive data—now face an unexpected capability gap between the hardware they believed they had purchased and what AMD is now claiming it never supported.

What Remains Unclear

AMD has not confirmed whether TSME will return to future consumer CPU generations. There's no indication that AMD ever advertised or marketed TSME as being available in consumer CPUs—a technical truth that AMD now emphasizes—but the years of functional availability and explicit engineer endorsements created a reasonable user expectation that the feature was supported. Whether AMD's roadmap for Ryzen desktop and mobile processors will include TSME restoration, or whether the removal represents a permanent product segmentation decision, remains to be seen.

The lack of transparency itself has become part of the story. Users discovered the change only through reverse engineering and persistence; AMD provided no release notes, no advance warning, and no public explanation. For a company selling security-focused hardware to privacy-conscious customers, that approach has eroded trust and raised questions about what other undocumented changes may have occurred in the firmware ecosystem.