A Hosting Provider Lost Its Customers' Data, and Told Them to Use Their Own Backups
BlackFog's The State of Ransomware: September 2026 report, last updated on 5 October 2026, carries an entry that should worry every small Indian business with a website. Indian hosting provider HostDZire confirmed a ransomware attack on 5 August 2026 after multiple VMware ESXi virtualisation nodes were compromised. The attack reached its infrastructure in India and several nodes in the Netherlands and the US. Virtual disks on the affected servers were encrypted beyond recovery, and BlackFog records the outcome as widespread service outages and permanent customer data loss.
HostDZire rebuilt the affected systems from scratch, launched a forensic investigation, and advised customers to restore their services from external backups where they had them. The ransomware group was not publicly identified.
Read that advice again from the customer side. The provider did not restore customer data, because the data no longer existed in any form it could recover. Every website, mailbox, database and application that lived only on those virtual disks was gone, and the only route back ran through a copy the customer had made and stored somewhere else.
Why this is an insurance question, not only an IT one
Thousands of Indian SMEs, D2C brands, chartered accountancy and law practices, clinics and small manufacturers run their storefronts, email and client portals on budget hosts. Many of them also bought the host's backup add-on, which sits on the same provider. When the host is the victim, the customer's own cyber policy is the first place the owner looks, and that policy was usually written on the assumption that the insured's systems are the ones attacked. This post walks through what actually responds when the breach happens at your provider, and what it takes to make the cover work before the next incident.
The Scale of the Exposure Is Hard to See Because Most Victims Are Small
The HostDZire incident is one data point, but it sits inside a larger pattern. The ransomware.live tracker's India map, viewed in October 2026, lists 549 Indian victims claimed on ransomware leak sites. Most are unlisted private firms, and the entries are claims made by the criminal groups, not confirmed incidents. Even with that caveat, the number tells you where the attacks land: on firms that rarely make the news and rarely have an in-house security team.
A hosting provider multiplies that exposure. One compromised virtualisation cluster can take down hundreds of customer sites at once, whatever security those customers kept on their own side. A business can patch every laptop and enforce strong passwords and still lose its data.
The same problem with a physical cause
Third-party provider failure is not only a ransomware story. In June 2026, Reuters-based reports carried by Business Today and w.media described a lithium-battery fire at an STT Global Data Centres India facility in New Delhi that forced Google to shut down networking equipment. Matrix Cellular said it might have lost access to more than 20 years of data, and ISP R2 Net estimated about US$2 million in outage losses.
The cause was different, a fire rather than an attacker, but the customer's position was identical: their data and their revenue depended on infrastructure they did not own, could not inspect, and could not insure directly. For a broader treatment of hyperscaler outages, see our post on cloud dependency and business interruption for SaaS firms. The HostDZire case is narrower and harsher: not an outage that ends, but a permanent loss at a small provider.
What Your Cyber Policy Pays When the Provider, Not You, Is Breached
Most SME cyber policies sold in India are built around first-party sections that respond to a security event affecting the insured's computer system. The definition of computer system is where the HostDZire scenario either fits or falls out.
Some wordings define the insured's computer system to include systems operated by a third party under a written contract for the insured's benefit, which would bring hosted servers inside the definition. Others limit it to hardware and software owned, leased or controlled by the insured. Under the narrower definition, an attack on a hosting provider's ESXi cluster is not an attack on the insured's system at all, and the first-party sections may not be triggered.
The sections that matter
When a provider is breached, three first-party sections usually carry the claim:
- Data restoration (sometimes called data recovery or digital asset replacement): pays the cost of restoring, recreating or recollecting data and software that has been damaged, destroyed or encrypted.
- Business interruption for the insured's own system: pays loss of net profit and increased cost of working when the insured's system goes down.
- Dependent or contingent business interruption: pays the same heads of loss when the outage originates at a third-party provider the insured relies on.
Incident response costs, such as forensic support and legal advice on notification, may also respond, but for a hosted-site customer the bulk of the money sits in restoration and lost revenue. Our post on cyber business interruption cover for Indian corporates covers how the main BI section is calculated.
Dependent Business Interruption: Triggers, Waiting Periods and Sub-Limits
Dependent business interruption is the section designed for a HostDZire-type event, and it is also the section most often missing, sub-limited or narrowly triggered in SME cyber policies.
How the trigger is usually worded
Dependent BI typically responds to an interruption of the insured's business caused by a security event or, in broader wordings, a system failure at a dependent business or service provider. Three variations matter in practice:
- Named versus unnamed providers. Some policies cover any IT service provider the insured contracts with. Others cover only providers listed on a schedule, which means an unlisted host is outside cover.
- Security event only versus system failure. A ransomware attack on the host is a security event under most wordings. A battery fire like the New Delhi incident is not, and would need a system failure or non-malicious outage extension, which many SME policies do not carry for dependent providers.
- Sub-limit. Dependent BI is commonly sub-limited to a fraction of the main BI limit. A policy with a headline limit of INR 1 crore may offer a much smaller figure for dependent BI. Check the schedule, not the brochure.
The waiting period
Cyber BI sections apply a waiting period, a time-based deductible that must elapse before loss starts to count. Dependent BI waiting periods are often longer than those on the insured's own system. For an outage that lasts a weekend, that may wipe out the claim. For HostDZire customers it should not, because the disks were never recovered and the interruption ran until each customer restored or rebuilt its services. The harder question is the indemnity period: whether the policy pays until the site is restored, or stops at a fixed number of days regardless of how long the rebuild takes.
Data Restoration Cover Pays to Restore From Backups That Exist
This is the point most SME owners get wrong. Data restoration cover pays the cost of restoring or recreating data. It does not conjure data back into existence.
If the customer has a clean, independent backup, the section pays for the technical work of rebuilding the site on a new host, reinstalling the application and loading the backup. That is a modest, well-defined claim.
If the customer has no backup, most wordings pay the reasonable cost of recreating data from other sources, such as re-entering product catalogues from supplier sheets or rebuilding client records from paper files and email archives, and they stop there. Many policies also say that if data cannot be restored or recreated, the insurer pays only the cost incurred in reaching that conclusion. They do not pay for the value of the data itself, the years of content, the SEO history, or the customer list that has simply disappeared.
The underwriting proposal form already asked
SME cyber proposal forms routinely ask whether the business keeps regular offline or offsite backups. An owner who ticked yes because the host's dashboard showed a backup button may face a utmost good faith argument when the claim arrives and the backups turn out to have lived on the same encrypted cluster. Answer that question accurately, and fix the answer before renewal if it is wrong.
The Hosting Contract Caps What You Can Recover From the Provider
When the cyber policy does not respond, or responds only in part, the obvious next step is to claim against the host. The hosting contract usually blocks that route.
Budget hosting terms of service typically contain a limitation-of-liability clause capping the provider's total liability at the fees paid by the customer over a defined period, often the preceding month or year. For an SME paying a few thousand rupees a year for shared or VPS hosting, that cap is a fraction of a single day's lost sales. The same terms usually exclude indirect and consequential loss, such as lost profits and lost data, and place responsibility for backups on the customer.
Why this matters for the insurer too
If your cyber insurer pays a dependent BI or restoration claim, it will look at subrogation against the provider. A liability cap at fees paid leaves almost nothing to recover, which is one reason insurers price dependent BI cautiously and sub-limit it hard. It also means you should not sign anything in the claims process that waives rights against the host without reading the policy's subrogation condition first, because prejudicing the insurer's recovery rights can affect the claim even when the recoverable amount is small.
The practical conclusion is simple. For an SME on budget hosting, the contract with the provider is not a risk transfer instrument. The cyber policy and your own backups are the only protection that will pay at the size of the loss.
A Checklist Before the Next Provider Breach
In a HostDZire-type event, the customers who come through best are the ones who have already done a few unglamorous things. Use this list at the next renewal, or sooner.
Independent backups and a vendor schedule
- Keep at least one backup outside the hosting provider's infrastructure. A copy on a different cloud account, a different provider, or offline storage you control. Test a restore at least once a year, because a backup that has never been restored is an assumption, not a control.
- Ask your broker for a vendor schedule on the cyber policy. List the hosting provider, email provider, payment gateway and any SaaS platform that runs revenue. Confirm that dependent BI applies to each, whether the trigger covers system failure as well as security events, and what the sub-limit and waiting period are.
- Check the definition of computer system to see whether hosted infrastructure under contract is included in first-party cover.
- Match the indemnity period to a realistic rebuild time. For a D2C brand that needs to rebuild a storefront, migrate payments and recover search rankings, a short indemnity period may end before trading recovers.
Documentation to prove a dependent BI loss
A dependent BI claim depends on evidence the insured has to assemble:
- The provider's incident notices, status page captures and support tickets, dated and saved.
- Daily sales, order and lead data for the weeks before the outage, to set the baseline.
- Invoices for emergency hosting, developers and data re-entry, which support increased cost of working and restoration.
- A timeline showing when the site went down, when restoration began and when normal trading resumed.
Notify the insurer as soon as you learn of the provider breach, not when your own losses are tallied. Our post on vendor breaches, leak sites and cyber claim notification covers the notification timing issues in more detail.
The Real Gap Is the Assumption, Not the Policy
Most Indian SMEs that buy cyber insurance do so assuming the policy covers whatever goes wrong with their digital business. The HostDZire attack shows where that assumption breaks. When the provider is the victim, cover depends on how the policy defines the insured's system, whether dependent BI is included and what triggers it, how far the sub-limit and waiting period cut the loss, and whether a usable backup exists for restoration cover to pay against.
None of those questions are exotic. They sit on the schedule and in the definitions section of a standard SME wording. Uptake of cyber cover among Indian MSMEs is still low, as we covered in our MSME cyber insurance uptake post, and the firms that do buy often buy the cheapest version without checking these terms.
BlackFog's September 2026 report will not be the last to list an Indian hosting provider. The fix costs less than the premium: an independent backup, a vendor schedule, and an hour with the policy wording before renewal.
