Inside Your Devices, There Is a Safe Room. Here Is Why It Needs an Upgrade

Every time you tap your phone to pay for something, a small but significant event takes place inside the chip.

Your payment credentials do not travel through the phone's main operating system — the same environment where your apps, your browser, and your notifications all run. Instead, they are processed inside a separate, isolated area of the chip: a protected zone that runs independently of everything else on the device, with its own memory, its own security rules, and no access from the outside.

That protected zone is called a Trusted Execution Environment, or TEE.

Most people have never heard of it. Almost everyone has one in their pocket.

This is Part 3 of the FOCAL Knowledge Series. Parts 1 and 2 covered post-quantum cryptography and crypto-agility. This post introduces the third building block: the secure hardware environment where the most sensitive operations on your devices actually take place — and why that environment is not yet quantum-safe.

What is a Trusted Execution Environment?

A Trusted Execution Environment is a secure area within a processor that guarantees that code and data loaded inside it are protected with respect to confidentiality and integrity.

The most useful way to think about it is as a safe room inside a chip. The main operating system of your phone or computer is powerful and flexible, but it is also large and complex — and complex systems have vulnerabilities. The TEE sits alongside it as a separate, smaller, hardened environment. Even if an attacker compromises the main operating system entirely, the data inside the TEE remains isolated and inaccessible.

A TEE achieves this through four core mechanisms:

Isolated memory: The TEE has its own protected memory region. Data stored there is encrypted and inaccessible to everything running outside the TEE — including the operating system, the hypervisor, and other applications.

Secure boot: The system boots from an immutable root of trust in the hardware, verifying the integrity of the TEE and its software during startup. This prevents the TEE from being tampered with before it has even loaded.

Hardware enforcement: The isolation between the TEE and the rest of the system is enforced at the hardware level — not by software that could potentially be overridden, but by the physical architecture of the chip itself.

Attestation. A TEE can provide remote verification that it is running the correct, unmodified software and is in a trusted state. This allows a bank, a payment processor, or a cloud service to confirm that the device it is communicating with is genuinely secure before transmitting sensitive data.

Where TEEs are used today

TEEs are already embedded in billions of devices across multiple industries. The three most widely deployed implementations are:

ARM TrustZone — present in the overwhelming majority of smartphones and embedded devices running ARM processors, which includes most Android and iOS devices in circulation today. TrustZone divides the processor into two worlds: a Secure World for trusted operations and a Normal World for everything else.

Intel SGX (Software Guard Extensions) — used in server environments and cloud computing, SGX allows applications to create isolated enclaves within the processor where sensitive computations can take place, even in infrastructure that the application owner does not control.

AMD SEV (Secure Encrypted Virtualization) — designed for virtualised cloud environments, AMD SEV encrypts entire virtual machines in memory so that cloud providers cannot access the contents of a customer's workload, even at the infrastructure level.

Apple's Secure Enclave — the dedicated chip that protects Face ID data, fingerprints, and device encryption keys on iPhones and iPads — is a well-known example of the same principle applied to consumer hardware.

These implementations share a common architectural goal, as GlobalPlatform, which coordinates TEE standardisation, has articulated: to provide a safe area of the device to protect assets and execute trusted code. In 2012, GlobalPlatform and the Trusted Computing Group published the first standard specifications for TEE architecture, interfaces, and security requirements — the technical foundations that implementations like TrustZone and SGX are built on.

How large is the market?

TEEs are not a niche technology. They sit at the centre of the rapidly growing confidential computing sector — the broader category of technologies that protect data while it is being processed, rather than only when it is stored or in transit.

According to MarketsandMarkets, the confidential computing market is expected to grow from $5.3 billion in 2023 to $59.4 billion by 2028, at a compound annual growth rate of 62.1%. TEEs represent one of the largest application segments within that market.

The drivers are straightforward: data breaches continue to rise in frequency and cost. The average global cost of a data breach in 2024 was $4.88 million — a 10% increase from the previous year and the largest single-year jump since the pandemic, according to IBM's 2024 Cost of a Data Breach Report. In healthcare, that figure rose to $9.77 million per breach. Organisations that process sensitive data in untrusted environments — cloud computing, shared infrastructure, third-party devices — are increasingly turning to TEEs as a way to reduce that exposure.

The problem: TEEs are not quantum-safe

Despite their strengths, TEEs have a significant blind spot.

The cryptography that TEEs depend on — the algorithms used for attestation, for key management, for securing the communication between the TEE and the outside world — is the same classical cryptography that post-quantum computing will eventually be able to break. The implementations are hardware-enforced and physically isolated, but the mathematical foundations of the security they provide are not resistant to quantum attack.

Side-channel attacks on current TEEs — such as Foreshadow, which was demonstrated against Intel SGX, and SEVered, which targeted AMD SEV — have already shown that the boundary between the TEE and the outside world is not impenetrable under all conditions. These attacks exploit physical properties of the hardware rather than breaking the cryptography directly. A future adversary with quantum capability would have an additional and more powerful avenue: breaking the classical cryptographic algorithms that protect the TEE's attestation and key exchange protocols.

This means the safe room has a lock that the next generation of computers will be able to open.

 The standards landscape 

The standards bodies responsible for TEE specifications have begun to respond.

The Confidential Computing Consortium — a Linux Foundation project whose members include AMD, ARM, Google, IBM, Intel, Microsoft, and NVIDIA — has published whitepapers on the security properties of TEEs and the requirements for confidential computing environments. While the CCC's published work to date has focused on the properties of existing implementations, quantum resistance is an active area of discussion within the consortium's technical advisory councils.

GlobalPlatform, which maintains the formal TEE specifications underpinning most commercial implementations, has a technical committee actively reviewing how TEE specifications will need to evolve to accommodate post-quantum cryptographic primitives.

Where FOCAL comes in

FOCAL's Work Package 3, led by the University of Lübeck, focuses on PQC-enabled applications — and TEEs are a central part of that work. As the FOCAL partners page confirms: "Activities in FOCAL align with recent studies on improving security for commercial trusted execution environments and on the secure transition and deployment of post-quantum cryptosystems."

The research is examining how TEE attestation protocols can be adapted to use post-quantum cryptographic algorithms, how the transition from classical to quantum-safe TEE security can be managed without disrupting the systems that depend on TEEs today, and how TEEs can be made flexible enough to update their own cryptographic foundations as standards evolve.

These are research goals currently in development, not finished products. But they address a gap that no existing commercial TEE implementation has yet solved.

The bottom line

There is almost certainly a TEE protecting your most sensitive data right now. It is doing its job well against the threats that exist today.

The question is whether it will continue to do so in a world where quantum computers can break the mathematical locks that TEE security depends on. That is the problem FOCAL is working to solve — and it sits at the intersection of hardware, cryptography, and the very foundations of how we trust the devices we rely on.

This is Part 3 of the FOCAL Knowledge Series. Part 4 covers edge and large-scale networked ecosystems: why securing billions of small, distributed devices is a fundamentally different challenge from securing a data centre.

Subscribe to the FOCAL newsletter at focal-pqc.eu to receive each post as it is published.