TL;DR: July 2026 crypto losses exceeded $240 million across 29 incidents, with private key, entropy, and hot-wallet compromise accounting for about $149.3 million and a firmware flaw in Coldcard hardware wallets contributing roughly $115 million, according to Quantstamp. The month reinforces that key custody and randomness quality are now operational risk controls, not implementation details.
At a glance
What this is: Quantstamp’s July Security Beat says crypto losses topped $240 million across 29 incidents, with key compromise, weak randomness, and signing-path failures driving most of the damage.
Why it matters: IAM and security teams should treat key custody, entropy quality, and delegated signing paths as governance issues because the same failure mode now cuts across onchain infrastructure and adjacent enterprise systems.
Context
July’s loss pattern is a governance problem as much as a technical one. When attackers can win by stealing keys, abusing signing paths, or exploiting weak randomness, traditional controls that focus only on code defects miss the mechanism that actually creates loss. The same risk pattern now spans blockchain infrastructure, Web2 accounts, and third-party access paths.
For identity and access practitioners, the important signal is that authentication strength alone is not enough if key material can be predicted, copied, or reused. The article shows that custody, lifecycle management, and trust in external dependencies are now part of the attack surface, especially where systems rely on hot signers, bridge keys, or third-party platforms.
Key questions
Q: What breaks when key generation falls back to predictable inputs?
A: The trust boundary breaks immediately because the resulting keys are no longer exclusive or unguessable. Attackers can reproduce candidate seeds, test them against public or observable data, and recover access without exploiting the application itself. In high-value systems, entropy failure turns a cryptographic control into a recoverable identity event.
Q: Why do compromised signing keys create outsized crypto losses?
A: Because the signer is the authority boundary. Once an attacker can produce valid signatures, they can authorize withdrawals, governance actions, or oracle updates that the system will accept as legitimate. That makes key compromise a direct path to control, not just a confidentiality issue.
Q: What are the signs that signing workflows are too trusted?
A: Look for workflows where one credential can approve multiple high-value actions, where bridge or oracle keys lack explicit ownership, or where revocation would be difficult to execute quickly. Those patterns indicate that the environment is trusting the signer more than the policy around it.
Q: What should security teams do when a wallet library may have exposed seed material or private keys?
A: Immediately assume the affected secrets are no longer trustworthy. Replace the compromised package, invalidate any keys or seeds that passed through it, and migrate funds or privileges to new wallets created from clean material. Then search telemetry for the malicious domain and confirm whether any systems transmitted key data externally.
Technical breakdown
Why key custody failures dominate loss events
A private key is not just a credential. It is the root of trust for whatever system accepts it, which means compromise usually turns into immediate authority rather than incremental access. In July, multiple incidents were driven by compromised bridge keys, oracle keys, and hot-wallet signing keys, showing that once the key is captured the attacker no longer needs to break application logic. This is why key custody must be treated as a control plane, not a storage problem.
Practical implication: separate high-value keys from routine operational workflows and treat signer access as a privileged control boundary.
How weak entropy becomes a takeover path
Randomness failures are dangerous because they collapse the search space an attacker must explore. If key generation falls back to predictable inputs, the resulting private keys can be derived by brute force or candidate regeneration, even when the underlying algorithm is sound. The Coldcard incident illustrates this clearly: a firmware defect shifted key generation away from dedicated hardware randomness toward predictable software inputs, shrinking the seed space enough to make recovery by attackers feasible.
Practical implication: validate entropy sources and fail closed whenever a device cannot prove it is generating keys from approved randomness.
Why signing-path abuse is more damaging than transaction fraud
Signing-path abuse differs from ordinary fraud because the attacker does not merely trick a user into approving a transaction. Instead, they obtain authority over the mechanism that authorizes the transaction in the first place. That can happen through compromised bridge keys, malformed governance execution, or oracle manipulation that lets an attacker mint false state into a protocol. Once the signing or verification layer is abused, the downstream system often accepts malicious actions as legitimate.
Practical implication: place verification, policy checks, and privilege separation around signing workflows, not just around the assets those workflows control.
Threat narrative
Attacker objective: The attacker’s objective is to obtain trusted authority over key material or signing workflows so that transfers, withdrawals, or state changes are accepted as legitimate.
- Entry occurred through compromise of signing infrastructure, weak randomness in device firmware, or an already exposed employee or third-party account.
- Credential or key material was then abused to generate valid signatures, derive recoverable wallet seeds, or authorize malicious protocol actions.
- The attacker escalated by using trusted signing paths, oracle feeds, bridge keys, or governance execution to move funds or falsify state.
- Impact was mass theft, unauthorized withdrawals, or large-scale data exposure across both crypto-native and adjacent Web2 environments.
Breaches seen in the wild
- Caesars Entertainment breach 2023: Social engineering of an IT support vendor let attackers copy Caesars loyalty database; about $15 million was reportedly paid.
Read and download The State of NHI & AI Agent Breach Report 2026, covering 200+ breaches impacting Non-Human Identities including AI Agents.
NHI Mgmt Group analysis
Key custody is now an operational risk control, not a backend detail. When a signing key or wallet seed can be recovered, copied, or reused, the security boundary collapses immediately. That is why key custody belongs alongside access governance, not only cryptography. Practitioners should treat signer stewardship as part of identity and privilege management, not as a narrow engineering task.
Weak randomness creates a governance failure, not just a technical defect. If a device can silently fall back from hardware entropy to predictable software inputs, the organisation has no meaningful assurance that the generated identity is unique or unguessable. This is a direct challenge to any control model that assumes key generation is trustworthy by default. Practitioners should require provable entropy assurance for every system that emits durable secrets.
Signing-path abuse is the real control gap behind many July losses. The breach pattern is not simply that keys were stolen, but that trusted paths existed for attackers to turn stolen or derived material into valid authority. That is a governance assumption failure: the environment assumed the signer, oracle, or bridge would remain trustworthy. Practitioners should review where trust is granted to signing workflows and who can influence them.
Cloud and Web2 incidents are converging on the same identity problem. The Bank of Baroda email compromise and Stadler Rail supplier breach point to the same lesson visible in crypto losses: one compromised credential or third-party access path can become the root cause of large-scale loss. The boundary between traditional IAM and machine or signing identity is narrowing, so the control model has to follow the authority path, not the asset label.
Randomness quality and key lifecycle management now deserve the same scrutiny as privilege review. Monthly loss aggregation is showing that the highest-cost failures are often the ones where governance was weakest upstream. The field should stop treating key generation, custody, and offboarding as implementation chores and start treating them as part of access assurance. Practitioners should align reviews to where authority is created, not only where it is used.
What this signals
Key custody has become a governance boundary. July’s losses show that when signing authority is concentrated in a small number of keys, the failure blast radius is immediate and large. That is why access governance for high-value signers now belongs in the same review cycle as privileged accounts and recovery controls.
Randomness quality is now a security control worth auditing directly. A device that can fall back to predictable inputs turns key generation into a repeatable attack surface rather than a source of uniqueness. Security teams should treat entropy assurance as part of secret creation governance, especially where wallets or hardware signers protect material value.
Signing-path trust creates the next exposure tier. Bridge keys, oracle keys, and delegated access paths are increasingly acting like machine identities with high authority but limited lifecycle oversight. The practical response is to map who or what can create valid authority, then reduce standing trust where that authority can be abused.
For practitioners
- Audit key generation entropy paths Verify that hardware wallets, signers, and key management services never silently fall back to predictable software inputs when hardware randomness is unavailable.
- Treat bridge and oracle signers as privileged identities Inventory every signing key that can authorize withdrawals, governance actions, or oracle updates, then assign explicit owners, rotation rules, and revocation procedures.
- Separate signing authority from routine operations Require policy checks, multi-party approval, or dedicated controls for any workflow that can move funds or change state, especially where hot signers are used.
- Review third-party access paths for identity collapse Map supplier platforms, employee accounts, and delegated access paths to the business assets they can influence, then remove standing access that can authorize high-impact actions.
Key takeaways
- July’s biggest losses were driven less by clever code exploits than by failures in key custody, entropy, and trusted signing paths.
- A single compromised or derivable authority path was enough to turn wallets, bridges, oracles, and accounts into immediate loss channels.
- Teams should harden entropy sources, narrow signer authority, and govern high-value keys as privileged identities rather than implementation details.
Standards & Framework Alignment
This section maps relevant standards and security frameworks to the operational risks and controls described in this guidance.
OWASP Non-Human Identity Top 10 and MITRE ATT&CK address the attack and risk surface, while NIST SP 800-53 Rev 5 and CIS Controls v8 set the governance and control requirements practitioners need to meet.
| Framework | Control / Reference | Relevance |
|---|---|---|
| OWASP Non-Human Identity Top 10 | NHI-02 — Secret Leakage | The article centres on exposed or recoverable key material across wallets and signing workflows. |
| NHI-07 — Long-Lived Secrets | July's losses show how persistent signing authority becomes catastrophic when compromise occurs. | |
| NHI-05 — Overprivileged NHI | Bridge, oracle, and wallet signers held more authority than their tasks should require. | |
| Recommendation — Scan for exposed NHI secrets and revoke any key material that could still authorize value transfer. Reduce the lifetime of high-value signing secrets and rotate authority before compromise windows widen. Limit signer scope so no single machine identity can authorize every high-impact action. | ||
| MITRE ATT&CK | TA0006;TA0008 — Credential Access; Lateral Movement | The incidents show attackers abusing keys and trust paths to move from access to authorization. |
| Recommendation — Map key compromise to credential access and lateral movement techniques in detection and response. | ||
| NIST SP 800-53 Rev 5 | IA-5 — Authenticator Management | Credential lifecycle control is central where signing keys and wallet secrets are the attack surface. |
| Recommendation — Apply authenticator management to rotate, revoke, and inventory high-value signing credentials. | ||
| CIS Controls v8 | CIS-5 — Account Management | The article shows why ownership and lifecycle of privileged access paths must be tightly controlled. |
| Recommendation — Maintain authoritative account inventories for signers, operators, and delegated access paths. | ||
Key terms
- Entropy Failure: A weakness in randomness generation that makes supposedly unique keys predictable or reproducible. In practice, entropy failure turns key creation into an attacker-searchable space and can undermine the security of wallets, certificates, and other secret-bearing systems.
- Signing Path: The workflow or system path that turns a secret, key, or approval into an accepted transaction or state change. When signing paths are overtrusted or insufficiently isolated, attackers can abuse them to authorize actions that appear legitimate to downstream systems.
- Hot Wallet: A hot wallet is a cryptocurrency wallet connected to the internet and used for active transactions. Because it is online, it can move funds quickly but also exposes signing authority to compromise if approvals, keys, or connected systems are abused.
- Machine Identity: The digital identity of a machine, device, or workload, such as a server, container, or VM, used to authenticate it within a network. Sometimes used interchangeably with NHI, though NHI is the broader category.
Deepen your knowledge
NHI Mgmt Group covers identity security, NHI governance, and agentic AI through independent research, practitioner guides, and the NHI Foundation Level course, the industry's only accredited NHI security programme. It is suitable for practitioners who need to connect identity governance to operational control.
Published by the NHIMG editorial team on August 18, 2026.
Updated on October 12, 2026.
NHI Mgmt Group, the independent authority on Non-Human Identity, IAM, and Agentic AI security. nhimg.org