Skip to content
New DelhiFree OSINT tools and field notes for investigators in India · Newsletter · Contact
IGRS logoIGRSIntel Grid Research Station
PayU's Rs 25.67 crore payout API breach: how the money moved and what investigators should ask for

PayU's Rs 25.67 crore payout API breach: how the money moved and what investigators should ask for

Attackers reached three PayU cloud servers, used the payout API to add beneficiaries and pushed out 265 unauthorised transfers to 156 bank accounts. The first arrests are not hackers but two men who rented out a bank account.

What happened

Gurugram Cyber Crime Police have arrested two residents of Ludhiana, Punjab, in a case where payment aggregator PayU lost about Rs 25.67 crore. According to MediaNama, attackers gained access to three of the company's cloud servers and manipulated its payout API, the interface a payment aggregator uses to send money out to merchants and other beneficiaries. They added new beneficiary accounts and moved funds without the company's approval.

MediaNama reports 265 unauthorised transactions into 156 bank accounts. PayU's security team noticed unusual API activity at about 11:30 pm on 5 November, and its cloud service provider also flagged suspicious access, but by then the transfers had gone through. PayU filed its complaint on 9 November. The420.in reports that the FIR invokes Section 318(4) of the Bharatiya Nyaya Sanhita (cheating).

The two men were arrested in Ludhiana on 15 November and remanded to police custody, according to The Pioneer's PTI-based report. Police say they did not hack anything. One allegedly let his account receive Rs 6.20 lakh of the stolen money for a 2% commission and handed it to the second man, who passed it on to a further handler for another 2%. Police say about Rs 1.5 crore has been frozen so far. How the attackers first got in is still unknown: police are waiting for an audit report, and The420.in says the compromised credentials could not be identified during detection. PayU had made no public statement at the time of the reports.

We covered this in one paragraph in this morning's India Cyber Brief. This piece looks at the mechanics and the investigation.

Why this is not an isolated case

On 4 November CERT-In and CSIRT-Fin issued a critical advisory, CIAD-2026-0058, on an AI-assisted threat actor they call CRAFTY FALCON that targets banks, NBFCs, digital lenders, payment aggregators and crypto exchanges. It updates an October advisory (CIAD-2026-0046) on the exploitation of payment and disbursement APIs. The advisory describes cloud management tools being abused to run commands, internet-facing test and staging systems used as a way into production, tampered mobile apps making API calls the real app never would, and Rs 1 test transfers to confirm a route before large amounts move. It also says attacks were timed for weekends, holidays and low-monitoring hours.

MediaNama notes that investigators have not linked the PayU case to that advisory, and we do not either. In the same week, City Union Bank told the stock exchanges that it found an unauthorised NEFT transfer from a customer account on 16 November and closed the NEFT and RTGS gateway for that account while external experts investigate, according to Elets BFSI. The bank did not disclose the amount or the cause.

Analysis: The common thread is that the weak point is no longer a customer's OTP but the machine-to-machine plumbing behind payouts. Once an attacker can call a payout or settlement API with valid-looking authority, the money leaves at software speed, and the only remaining brake is how fast the receiving accounts can be frozen.

How a payout API fraud works

Based on the reporting and the CERT-In advisory, the pattern has four stages. We describe them at the level a defender or investigator needs, not as a recipe.

  1. Foothold in the cloud. The attacker obtains some form of access to servers or cloud accounts that can talk to the payout system. CERT-In lists exposed non-production systems, shared credentials and tokens between test and production, and hardcoded secrets as common routes.
  2. Beneficiary manipulation. Instead of stealing card data, the attacker adds new beneficiaries or edits existing ones, so the system itself pays out. In the PayU case, police say new beneficiary accounts were added through commands on the compromised servers.
  3. Rapid fan-out. Money is split across many accounts (156 in the first layer here) so that no single transfer looks large and freezing one account recovers little.
  4. Layering through mules. Account holders rent out their accounts for a small cut, and the money is withdrawn, moved again or converted to crypto. The420.in reports that some account providers were paid as little as Rs 10,000 in cash.

Analysis: The first arrests in such cases are almost always at stage 4, because bank KYC leads straight to the account holder. Reaching the people at stage 1 depends on cloud and API logs, which is why the evidence request matters so much.

Red flags

WhereIndicator
Payout and settlement APIsBeneficiaries added or edited outside normal workflows, especially in batches or at night
TransactionsRs 1 or other micro test transfers to new beneficiaries followed soon after by large payouts
TimingActivity spikes on weekends, public holidays or late at night, as CERT-In describes
Cloud audit logsRemote command execution, new access keys or roles, changes to serverless functions in production
Test and stagingAny connection from a non-production system into production
Receiving accountsNewly opened or long-dormant savings accounts suddenly receiving business-sized credits and withdrawing them within hours

What to do

If you are an ordinary user

  1. Never rent, sell or "lend" your bank account, UPI ID or SIM for a commission. The men arrested in this case are accused for exactly that. Offers dressed up as part-time work or "account verification" jobs are the usual bait.
  2. If money you did not expect lands in your account, do not withdraw or forward it. Tell your bank in writing and keep the acknowledgement.
  3. If you lose money to unauthorised transfers, call 1930 or report at cybercrime.gov.in at once. Quick reporting is what makes freezes like the Rs 1.5 crore here possible.
  4. If your email or phone number appears in a breach notice from a payments company, check it with IGRS Breach Check and change passwords you have reused elsewhere.

If you are an investigator

  1. Report bulk fraudulent transactions to I4C through the Citizen Financial Cyber Fraud Reporting and Management System (CFCFRMS) so that onward layers are flagged across banks, as the CERT-In advisory asks.
  2. Send a notice under Section 94 BNSS to the victim company for the payout API logs, beneficiary creation and change logs, admin and service-account login records and cloud provider audit logs for the whole window, not just the day of the transfers.
  3. Ask the cloud provider to preserve its logs for the affected accounts quickly; retention periods can be short.
  4. Map the money trail layer by layer from bank statements and KYC of the first-layer accounts. Treat account holders as both suspects and witnesses to reach the handlers above them.
  5. For source IP addresses in the API or cloud logs, use IP Lookup to identify the network and VPN and proxy checker to see whether it is a VPN, hosting or anonymising exit. CERT-In warns that this actor rotates infrastructure within hours to days, so record timestamps carefully.
  6. Obtain a Section 63 BSA certificate for every electronic record you rely on, from the company and from banks.
  7. Regulated entities that suffer such an incident should report it to CERT-In ([email protected]) within the mandated window, with logs.

If you run a fintech, NBFC or aggregator

CERT-In's countermeasures include server-side validation and maker-checker for settlement instructions, cooling periods and velocity rules for new beneficiaries, signed API requests, no shared credentials between test and production, short-lived tokens and secret scanning in code repositories. The420.in reports that the RBI has also asked banks and NBFCs to tighten API and third-party controls, though no circular number has been published yet.

Applicable law

ProvisionWhat it coversRelevance here
S.318(4) BNS 2023 (was S.420 IPC)Cheating and dishonestly inducing delivery of propertyCited in the FIR, per The420.in
S.61 BNS 2023Criminal conspiracyLinks hackers, aggregators and account providers
S.317 BNS 2023Stolen propertyReceiving or retaining the diverted money
S.43 and S.66 IT Act 2000Unauthorised access and computer-related offencesAccess to cloud servers and API abuse
S.66C IT Act 2000Fraudulent use of passwords, keys or other unique identifiersIf stolen credentials or API keys are established
S.70B IT Act and CERT-In Directions, 2022Mandatory incident reportingReporting by the affected entity
S.94 BNSS 2023 (was S.91 CrPC)Summons to produce documents or recordsLogs from the company, cloud provider and banks
S.106 BNSS 2023Seizure of propertyFreezing accounts holding proceeds
S.63 BSA 2023 (was S.65B IEA)Admissibility of electronic recordsCertificate for logs and statements
S.8(5) and S.8(6) DPDP Act 2023Security safeguards and breach intimationOnly if personal data was also affected; none has been reported here
PMLA 2002, S.3 and S.4Money launderingLayering of proceeds through mule accounts

Analysis: Sections beyond S.318(4) BNS are our reading of what may apply as the investigation develops, not charges reported by police.

What to watch next

The audit report on how the servers were breached is the key document still pending. If it shows stolen API credentials or a path from a test environment, the case will fit squarely into the pattern CERT-In has been warning about since October. We will update this page if PayU or the police publish findings.

Sources

Frequently Asked Questions

What is a payout API?
It is the interface a payment aggregator or lender uses to send money out to merchants, borrowers or other beneficiaries automatically. If an attacker can add beneficiaries or issue instructions through it, the system pays the attacker as if it were a normal payout.
Did hackers steal customer card or UPI data in the PayU case?
No such loss of customer payment data has been reported. The reported loss is PayU's own funds, about Rs 25.67 crore, sent through its payout system to 156 bank accounts.
Why were account holders arrested and not the hackers?
Bank KYC leads police directly to the people whose accounts received the money. Police say the two arrested men rented out an account for a commission. Identifying the original attackers depends on the pending audit and log analysis.
What should I do if unknown money appears in my bank account?
Do not withdraw or forward it. Inform your bank in writing and, if you suspect fraud, report it on 1930 or cybercrime.gov.in. Using such money can make you part of the case.