
Imagine a bank vault manufacturer who runs a public challenge: 'Here's one of our vaults, locked. Crack it and we'll pay you.' A year goes by and nobody claims the prize. The manufacturer starts marketing the vault as unbreakable. Then two locksmiths show up, not for the bounty, but to publish exactly how the lock leaks — because every library and hospital in the country just bought one. That's what Tattikota and Immler have done with Raspberry Pi's RP2350 microcontroller. The RP2350 Challenge 2 presented a hardened AES implementation and invited the security community to break it. For over a year, nobody publicly did. This paper demonstrates comprehensive leakage detection, plaintext dependence, and full key extraction from the challenge implementation — contradicting the challenge authors' claims about its resistance to side-channel attacks. Side-channel analysis (SCA) exploits unintentional information leakage from physical implementations — power consumption, electromagnetic emanations, timing variations — rather than attacking the math of the cipher itself. AES is mathematically sound; the question is always whether the silicon running it whispers secrets through its power pins. The RP2350's challenge implementation was supposed to be hardened against exactly this class of attack. The paper shows it wasn't. The authors present a methodical three-stage attack. First, leakage detection confirms that exploitable side-channel signals exist in the implementation's physical emissions. Second, plaintext dependence analysis demonstrates that the leakage correlates with known inputs in a way that narrows the key search space. Third, full key extraction recovers the secret key. The paper spans 7 pages with 12 figures documenting each stage — this is a compact, evidence-heavy result, not a theoretical argument. What makes this publication notable is its independence from the bug bounty structure. The authors explicitly frame their work as free from third-party obligations, arguing that the RP2350's widespread adoption in embedded systems makes independent security assessment a public interest matter. Bug bounty programs can carry restrictive terms — NDAs, vendor approval of publications, timing constraints. By publishing independently, the authors ensure the security community can scrutinize and build on the findings without gatekeeping. The paper's core tension is with Raspberry Pi's own security claims. The challenge was implicitly a marketing tool: an unsolved challenge suggests a secure product. The authors' demonstration of 'concerning weaknesses' directly undermines that narrative. For a microcontroller deployed in everything from hobbyist projects to commercial embedded systems, the gap between marketed security posture and demonstrated vulnerability is the real story. The broader lesson here is familiar but important: hardened implementations of well-understood ciphers still fail in the physical domain. AES itself isn't broken — the RP2350's implementation of AES is. This distinction matters because it means every deployment is a fresh attack surface. The authors' call for further public scrutiny is effectively a statement that one team's analysis shouldn't be the last word — the community needs to pile on.
The analogy
Think of a bank vault manufacturer who runs a public 'crack this vault' challenge. Nobody claims the prize for a year, so the marketing team calls it unbreakable. Then two independent locksmiths publish exactly how the lock leaks heat signatures that reveal the combination — not for the bounty, but because thousands of buildings already installed these vaults.
Vault heat leak
A lock that leaks thermal signatures revealing internal state to anyone with an infrared camera
RP2350 SCA leak
A hardened AES implementation that leaks power/EM signals correlating with the secret key to anyone with an oscilloscope
Honest assessment
This is a clean, compact proof-of-concept that breaks a year-old public challenge on a widely deployed chip. The SCA techniques are standard — the contribution is the target, not the method. The paper's real weight comes from its independence and its direct contradiction of vendor claims. It's short on statistical depth (7 pages constrains trace-count reporting and reproducibility detail), but the physical-experiment basis is strong. Hardware security engineers deploying RP2350 in security-sensitive applications should take notice immediately; the broader embedded security community should wait for replication before drawing systemic conclusions.
What Is CLAIMS?
67
/ 100
BedrockNews applies a structured analytical framework called CLAIMS to academic papers and technical research releases.
How CLAIMS worksClaim
4/5
Ladder
3/5
Architecture
3/5
Integrity
4/5
Milestone
3/5
Successor
3/5
What is this paper actually committing to?
Genuine result: full key extraction from a year-unsolved hardened AES challenge on a widely deployed microcontroller. Not a new SCA technique — but a new target broken, contradicting vendor claims.
The field fight
Should hardware security research be channeled through vendor bug bounty programs, or does independent publication serve the public better?
Side A
Coordinated disclosure through vendor programs produces better outcomes — vendors patch before publication, users are protected
Held by: Raspberry Pi, HackerOne, most vendor security teams
Side B
Independent publication without vendor gatekeeping is essential — bounty terms can suppress findings, unsolved challenges create false confidence
Held by: This paper, academic SCA community, independent security researchers
Milestone tracker
Unlocks: Definitive evidence forcing vendor response — firmware patch, updated security guidance, or revised marketing claims
Current
1
implementation broken
RP2350 Challenge 2
Next
3+
independent replications
community validation
~3-6 months if other SCA labs attempt replication
Verified Facts (6)
The RP2350 side-channel analysis challenge remained unsolved for more than a year
The paper demonstrates comprehensive leakage detection, plaintext dependence, and key extraction from the challenge implementation
The authors state their analysis contradicts the challenge authors' claims
The paper is 7 pages with 12 figures
The paper is categorized under cs.CR (Cryptography and Security)
Authors are Nikhil Tattikota and Vincent Immler
Practical Implications (3)
Designers using RP2350 for security-critical applications should not rely on the AES hardening as demonstrated in Challenge 2 without independent validation
Bug bounty program operators should note that unsolved challenges are not evidence of security — they may reflect insufficient incentive or restrictive terms rather than implementation strength
Organizations deploying RP2350 in physical-access threat models (IoT, embedded devices) should assess side-channel exposure as a concrete risk, not a theoretical one
More latest news
Read all