Risk Management Strategies

eMudhra's Cracked CA Keys and Chrome's 27 October Distrust: Is a Certificate Authority Failure a Cyber Loss Your Policy Pays?

Eight eMudhra intermediate CA keys were trivially weak, Mozilla has distrusted its ECC root and Chrome is expected to follow on 27 October. Here is what Indian businesses holding eMudhra certificates should do now, and why most cyber policies will not pay for a non-malicious trust-provider failure.

Tarun Kumar Singh
Tarun Kumar SinghStrategic Risk & Compliance SpecialistAIII · CRICP · CIAFP
9 min read

Listen to this article

Audio version • 9 min read

cyber insurancecertificate authoritydependent business interruptiontech E&Ovendor riskeMudhra

Last reviewed: November 2026

What happened between 17 and 22 October

The sequence ran fast. According to Help Net Security (21 October 2026), the problem was reported on 17 October 2026: eight intermediate certificates issued under eMudhra's emSign hierarchy had private keys smaller than 256 bits. Security researcher Scott Helme recovered those private keys from the public certificates in under a second. Six of the eight intermediates issue TLS certificates; two are for code signing.

The response from the ecosystem was equally quick. eMudhra revoked the intermediates on 18 October. Apple blocked them on 19 October and Microsoft on 20 October, the same day Mozilla disclosed the incident. eMudhra attributed the weak keys to a faulty script used during a key ceremony in April 2025, although its own statement says the root cause is not yet confirmed.

Then the root itself came under pressure. The Hacker News (22 October 2026) reported that Mozilla decided on 20 October to stop trusting the emSign ECC Root CA - G3, and that Chrome's distrust of that root is expected on 27 October. eMudhra has said it will exit its global SSL/TLS business, which it puts at about 1% of revenue.

Downstream, CyberScoop (22 October 2026) reported that Go, Node.js, Python certifi and several Linux distributions shipped updates, that CERT-In said it had not seen malicious use of the keys, and that eMudhra shares fell 40% in the two days after disclosure.

The numbers do not agree, and you should not force them to

  • eMudhra's official statement (first published 20 October, updated 22 October) says root keys were not affected, seven of the eight intermediates were never used in production, the one in production issued no certificates, and the distrust affects one US customer ecosystem.
  • The Hacker News reports eMudhra's own estimate of about 30,000 certificates needing replacement.
  • Chrome counted about 30,800 unexpired certificates chaining to the ECC G3 root, per The Hacker News.

These figures describe different things (compromised intermediates versus a distrusted root) and come from parties with different incentives. For a risk manager the practical answer is the same either way: check your own inventory rather than relying on anyone's aggregate count.

Why a certificate authority is a critical dependency

Most Indian vendor risk registers list cloud providers, payment gateways and core banking or ERP vendors. Few list the certificate authority. That is an oversight, because a CA failure does not need an attacker to stop your business. When a browser or operating system stops trusting a chain, every certificate under it fails validation on that platform, whether or not anyone has misused a key.

Trust-service providers touch three separate operational layers:

  1. TLS certificates that let customers, partners and APIs reach your web properties and endpoints without warnings or hard failures.
  2. Code-signing certificates that let your software, drivers or updates install and run on customer machines without being blocked.
  3. Electronic signature services (eSign, signing platforms, certificate lifecycle tools) that underpin contracts, KYC journeys and statutory filings.

eMudhra states that its eSign, emSigner and CertiNext products are unaffected. That matters for Indian users of those services, but it does not change the general lesson: each layer is a single point of failure that sits outside your own security controls. A failure at the trust layer can be entirely non-malicious and still take a revenue channel offline.

This is the same accumulation problem we described in AI provider concentration: the named vendor looks diversified on paper while the exposure sits in shared infrastructure underneath it.

What to do now if you hold eMudhra TLS or code-signing certificates

This post is published on 23 October and Chrome's distrust is expected on 27 October. That leaves a few working days, so sequence the work by impact.

  1. Inventory first. Pull every certificate in your estate and identify anything chaining to an emSign intermediate or to the emSign ECC Root CA - G3. Include load balancers, API gateways, mobile app pinning configurations, IoT devices and partner-facing endpoints, not just the main website.
  2. Replace TLS certificates on customer-facing endpoints before 27 October. eMudhra says it is offering Sectigo-issued certificates as replacements; you may also move to another CA of your choice.
  3. Review code-signing. If any build was signed using a chain under the revoked intermediates, test whether it still installs and runs on current Windows and macOS versions, given that Microsoft and Apple have blocked the intermediates. Plan a re-sign and re-release where needed.
  4. Check runtime trust stores. Go, Node.js, Python certifi and several Linux distributions have shipped updates. Services that call out to partners using certificates under the affected chain may start failing after routine dependency upgrades.
  5. Record everything. Log the date you learned of the issue, the steps taken, staff hours, vendor invoices and any revenue or transaction dips. That record is what an insurer will ask for if you later notify a claim.

Does cyber business interruption respond to a non-malicious CA failure?

This is the core coverage question, and the answer depends on how your cyber insurance wording defines the trigger for business interruption.

Security failure versus system failure

Most cyber wordings in the Indian market split the business interruption trigger in two:

  • A security failure trigger responds to unauthorised access, malicious code, denial-of-service or similar attacks. On CERT-In's statement that it has not seen malicious use of the keys, an outage caused by browser distrust is hard to fit here. Nobody attacked you.
  • A system failure trigger responds to an unintentional, unplanned outage of computer systems, often including human error or programming error. Some wordings limit this to your own systems; some extend it to systems operated by a provider on your behalf.

A browser deciding to stop trusting a root is not an outage of anyone's computer system in the ordinary sense. Your servers are up. The CA's systems are up. What has changed is a trust decision made by third parties. Whether that is a "system failure" turns on the exact policy wording, and many wordings will not stretch to it.

Dependent or outsourced-provider BI

Where a policy offers dependent business interruption, check whether the listed or unnamed providers include a certificate authority or trust-service provider, and whether the dependent trigger includes system failure or only security failure at the provider. A dependent extension written for security failure only will not respond to a key-generation error at a CA that nobody exploited. We worked through the same split for network outages in the Azure gateway outage analysis.

Waiting periods narrow the question further. Cyber BI typically carries a waiting period (often expressed in hours) before loss is counted. A well-run certificate migration completed before 27 October may cause no outage at all, or one shorter than the waiting period. The cover question becomes live only for firms that miss the window or discover a forgotten endpoint or signed binary after the fact.

Are forced migration costs covered, or are they betterment?

Even if no revenue is lost, many firms will spend money: emergency procurement of replacement certificates, engineering time to redeploy them, re-signing and re-releasing software, and possibly external consultants to audit the estate.

Insurers commonly treat these costs as one of three things:

  • Incident response costs, which a cyber policy may pay when they follow a covered event. The difficulty is the same as the BI trigger: if there is no security or system failure as defined, there is no covered event to respond to.
  • Remediation or betterment, meaning the cost of bringing your systems to the state they should have been in, or improving them. Most cyber wordings carry an exclusion for upgrading or improving systems, and replacing certificates with a new CA's certificates looks a lot like routine maintenance.
  • Mitigation costs incurred to prevent a covered BI loss. Some wordings will pay reasonable costs that reduce a loss that would otherwise be covered. This is the most promising route, but it still depends on the underlying BI loss being covered in the first place.

Treat the most likely outcome as uninsured. The cost of an orderly migration is usually small next to the cost of an outage, so budget for it as an operational expense and pursue recovery from the vendor contract where possible.

Contract terms to require from certificate and trust-service vendors

If insurance is unlikely to pay for a non-malicious CA failure, the vendor contract carries more weight. Our broader framework for third-party cyber vendor risk applies here, with some CA-specific terms worth adding at the next renewal or procurement:

  1. Distrust and revocation notice. A defined notice period and channel for any revocation, browser root-programme action or compliance incident affecting your certificates.
  2. Replacement at vendor cost. Free reissuance from an alternative trusted chain, including a named fallback CA, if the vendor's hierarchy is revoked or distrusted.
  3. Reimbursement of migration costs. A specific, capped obligation to pay reasonable internal and external costs of forced migration, rather than relying on a general liability clause that will usually exclude consequential loss.
  4. Service exit commitments. Minimum notice and transition support if the vendor exits a product line, as eMudhra has now announced for global SSL/TLS.
  5. Evidence of insurance. A requirement that the vendor carries technology errors and omissions cover at a stated limit, with a certificate of insurance delivered annually.
  6. Key ceremony and audit disclosure. Access to the vendor's audit reports and a duty to disclose ceremony or key-generation failures, since eMudhra has attributed the weak keys, pending confirmation of the root cause, to a script error in an April 2025 ceremony that surfaced only in October 2026.

Diversification also helps. Holding certificates from two CAs, with automated issuance tooling ready to switch, turns a multi-day migration into a configuration change.

What listed trust-service and tech vendors should check in tech E&O and D&O

eMudhra is a listed company that faced a public CA key incident, saw its shares fall 40% in the two days after disclosure and announced a product-line exit within the same week. Any Indian listed technology vendor should treat that combination as a test case for its own programme.

Tech E&O

Customers who incur migration costs or outage losses may look to the vendor. A professional indemnity or tech E&O policy is the natural respondent, and the questions are familiar from the Azure credential leak case: whether a customer's cost of replacing a product counts as damages or as a refund of fees (often excluded), whether contractual liability assumed beyond the law is covered, and whether the E&O and cyber policies of the vendor sit with the same insurer or will argue about which one responds.

D&O

A sharp share fall after a security disclosure is the classic setup for shareholder or regulatory scrutiny of what directors knew and when. Check that the directors and officers liability policy covers securities-related claims and regulatory investigations, that any cyber or data exclusion carves back cover for management liability claims, and that the timeline of internal knowledge (here, report on 17 October and public disclosure on 20 October) is documented. A product-line exit adds a further question: whether disclosure about the exit and its revenue impact was accurate.

Notify both insurers of circumstances early. Late notice is the easiest way to lose cover you paid for.

About the Author

Tarun Kumar Singh

Tarun Kumar Singh

Strategic Risk & Compliance Specialist

  • AIII
  • CRICP
  • CIAFP
  • Board Advisor, Finexure Consulting
  • Developer of the Behavioural Underinsurance Risk Index (BURI)

Tarun Kumar Singh is a seasoned risk management and insurance professional based in Bengaluru. He serves as Board Advisor at Finexure Consulting, where he advises insurance, fintech, and regulated firms on governance, growth, and trust. His work spans insurance broker regulatory frameworks across India, UAE, and ASEAN, IRDAI compliance and Corporate Agency model reform, VC governance in insurtech, and MSME insurance gap analysis. He is the developer of the Behavioural Underinsurance Risk Index (BURI), a framework applying behavioural economics to underinsurance and insurance fraud risk.

Frequently Asked Questions

We hold an eMudhra TLS certificate. Will our website stop working on 27 October?
Only if the certificate chains to the emSign ECC Root CA - G3 or to one of the revoked intermediates, and only on platforms that apply the distrust. Chrome's distrust is expected on 27 October; Mozilla, Apple and Microsoft have already acted on the root or intermediates. Check the chain on every endpoint and replace affected certificates before that date. eMudhra is offering Sectigo-issued replacements.
Is a certificate authority distrust a cyber attack for insurance purposes?
Usually not. CERT-In has said it had not seen malicious use of the keys, and the distrust is a decision by browser and operating system vendors. A security-failure trigger generally needs unauthorised access, malicious code or a similar attack, so most claims would have to rely on a system-failure or dependent-provider trigger, and those wordings vary.
Will our cyber policy pay for replacing certificates?
Probably not on its own. Insurers tend to treat replacement certificates and redeployment work as remediation or betterment. Some wordings pay mitigation costs incurred to reduce a covered business interruption loss, so the answer depends on whether an underlying BI loss would be covered. Ask your insurer in writing before spending.
Should we notify our insurer if we have suffered no loss?
Yes. Notifying the issue as a circumstance protects your position if a forgotten endpoint, partner integration or signed software build fails after 27 October, and it gets the insurer's position on migration costs on record early.
What should we ask certificate vendors for at renewal?
A defined notice period for revocation or distrust events, free reissuance from a named fallback CA, a capped obligation to reimburse forced migration costs, minimum notice before any product-line exit, disclosure of key ceremony failures, and an annual certificate of insurance showing tech E&O cover.

Related Glossary Terms

Related Insurance Types

Related Industries

Related Articles

Sarvada Intelligence

Ready to see Sarvada in action?

Explore the platform workflow or start a product conversation with our underwriting automation team.

Explore the platform