A closed cooling valve set off days of disruption
Singtel has been fined S$1 million after a maintenance lapse, missing monitoring and problems at a second network facility disrupted mobile services for about 720,000 consumer and business subscribers in Singapore on March 16, 18 and 19, 2026.
Contents
- A closed cooling valve set off days of disruption
- What happened on each date?
- Why the cooling system failed
- Why moving traffic caused further problems
- How many customers were affected?
- How IMDA decided the penalty
- What Singtel has changed
- The one hour recovery requirement
- Another S$1 million fine after a landline failure
- Key Points
The Infocomm Media Development Authority (IMDA), announcing the penalty on October 20, said all three incidents were within Singtel's control to prevent. Its investigation found that the disruptions were not caused by a cyberattack.
The first failure began with a manual valve that a vendor did not reopen after scheduled maintenance. Without the required cooling water, mobile network equipment overheated and shut down. Two further disruptions followed when Singtel moved 5G traffic to another exchange that had a disabled network packet optimiser and a software bug.
The consequences reached beyond interrupted browsing. Online payments, ride hailing and food delivery were disrupted, showing how a fault inside a telecommunications facility can interfere with everyday purchases, journeys and business transactions.
Singtel Singapore chief executive Ng Tian Chong said the company accepted the regulator's ruling and financial penalty. He apologised to customers:
"We deeply regret the disruption and the inconvenience caused to our customers," he said.
IMDA considered both the extensive disruption and Singtel's response when setting the penalty under the Telecommunications Act. The company cooperated with investigators, restored services in stages and provided goodwill rebates totalling about S$8 million to consumer and business subscribers.
What happened on each date?
March 16 brought the longest disruption, lasting about nine hours. Customers began experiencing connectivity problems at around 10.30am. Singtel restored 4G before bringing its 5G network back progressively.
The sequence shows how the initial cooling failure was followed by difficulties handling traffic elsewhere in the network:
- March 16: Connectivity problems began at about 10.30am. 4G services returned around 1.30pm, while progressive restoration of 5G began at 2.45pm. Full 5G service returned at about 8pm.
- March 18: Some subscribers experienced intermittent connectivity during the evening. Singtel said the disruption was resolved within an hour.
- March 19: Some subscribers again experienced intermittent connectivity during the evening after traffic had been redirected to the other exchange.
- April 7: Minister for Digital Development and Information Josephine Teo addressed service reliability requirements in a parliamentary reply while IMDA's investigation was continuing.
- October 20: IMDA announced its findings and the S$1 million penalty.
No duration was given for the March 19 disruption. The incidents occurred on three separate dates, with the later problems linked to the network changes made after the first failure.
Why the cooling system failed
IMDA traced the March 16 outage to a chiller system at a Singtel exchange. An exchange houses equipment and connections that route telecommunications traffic. Its network equipment needs cooling to keep operating safely.
A Singtel vendor had carried out scheduled maintenance but failed to follow the procedure requiring a manual valve controlling water flow to be reopened afterward. Singtel also lacked active monitoring of the manual valves or water flow to the chiller.
The result was a chain of failures: water flow was interrupted, cooling failed, equipment overheated and mobile network equipment shut down. Customers then lost connectivity because equipment carrying their traffic was no longer operating.
IMDA placed responsibility on both the maintenance process and Singtel's oversight. It said stronger procedures, management arrangements and monitoring systems could have prevented the disruption. Checking that maintenance had been completed correctly, and detecting an interruption in cooling water, were central to keeping the exchange running.
Why moving traffic caused further problems
Singtel redirected 5G traffic from the affected exchange to another exchange after March 16. Moving traffic between facilities is part of network failover, the process of switching work elsewhere when equipment or a location fails.
The receiving exchange faced increased traffic while its network packet optimiser was disabled and a software bug was present. Together, those conditions caused intermittent connectivity for some subscribers during the evenings of March 18 and 19, IMDA found.
A packet is a small piece of digital information sent across a network. A packet optimiser helps route those pieces along suitable paths, improving how efficiently data travels and reducing delays. With that function disabled, the second exchange had an additional problem while carrying more traffic.
Ng had previously explained that the March 18 incident followed network reconfigurations intended to stabilise and improve performance after the first outage. He said a brief surge in traffic caused the disruption, which was resolved within an hour.
IMDA's findings identified the conditions behind the later interruptions. The regulator said Singtel should have exercised greater diligence to ensure the receiving exchange could handle the additional load. Having another exchange available required checking its capacity, settings and software before relying on it to absorb the traffic.
How many customers were affected?
About 720,000 consumer and business subscribers were affected across the three incidents, according to IMDA. The regulator described the impact as significant because of the number of subscribers involved.
Singtel has around 4.5 million mobile subscribers, compared with approximately 2.2 million at StarHub, 1.6 million at M1 and 1.4 million at Simba Telecom. Using those figures, the affected Singtel subscribers represented roughly 16 per cent of its mobile subscriber base, or about one in six.
The disruption also reached services that depend on a working mobile connection. A customer unable to complete an online payment, a driver waiting for a booking or a business handling delivery orders can all encounter difficulties when mobile data becomes unavailable.
Restoring 4G first gave affected subscribers access to that service while 5G recovery continued. IMDA took this progressive restoration into account when assessing Singtel's efforts to bring customers back online.
How IMDA decided the penalty
IMDA imposed the S$1 million penalty under the Telecommunications Act. Its decision weighed the scale and impact of the outages alongside Singtel's conduct during the investigation and recovery.
The regulator credited Singtel's full cooperation with investigators, its work to restore services progressively and the goodwill rebates it voluntarily extended to affected subscribers. Those rebates were described as S$5 to S$10, with the total payout reaching about S$8 million for consumer and business subscribers.
The reported rebate total was eight times the regulatory fine. The two payments served different purposes: the rebates went to customers as a goodwill response to the disruption, while the fine was an enforcement penalty for the service failures.
IMDA also considered measures Singtel had taken to prevent another incident. Its central finding remained that the cooling failure and the later traffic handling problems could have been prevented through better monitoring, procedures and checks.
What Singtel has changed
Singtel has strengthened maintenance procedures and monitoring for its chiller systems, corrected systems at the second exchange and conducted exercises to test how effectively its network can switch traffic elsewhere during a failure, according to IMDA.
Those changes address the two main stages of the incident. Better cooling checks are intended to catch problems at the original facility. Repairs and failover exercises are intended to ensure another exchange can accept redirected traffic when it is needed.
Ng said Singtel had invested substantially in improvements to network design and was using technologies such as artificial intelligence to test failover capabilities more effectively. He also said the company would continue strengthening its systems.
The Singtel Singapore chief executive described reliability as a continuing priority:
Network reliability and resiliency have always been a top priority for us, and we are committed to continuing to strengthen our systems to provide reliable essential connectivity services to our customers.
The one hour recovery requirement
Josephine Teo, the minister responsible for digital development and information, said in her April 7 parliamentary reply that telecommunications providers are required to resolve service incidents within one hour. Providers that fail to do so may face enforcement action and financial penalties.
Teo also said providers must conduct regular audits of key telecommunications infrastructure for resilience. Such checks examine whether essential facilities and systems can withstand faults and continue providing service.
The March 16 disruption lasted far longer than the one hour requirement, although Singtel said the March 18 incident was resolved within that period. IMDA's findings addressed both the time taken to recover and the operational lapses that allowed the incidents to happen.
IMDA said telecommunications companies provide essential services and must plan, design and operate resilient networks. They must also have measures for speedy recovery with minimal inconvenience to users.
The authority stated its enforcement position:
IMDA will not hesitate to take strong action under the Telecommunications Act, including imposing financial penalties, should any lapses be identified.
Another S$1 million fine after a landline failure
Singtel received another S$1 million fine in December 2025 for a fixed voice service outage on October 8, 2024. That incident affected about 500,000 users and lasted more than four hours.
For about four hours, roughly half of calls to public service hotlines, including government agencies, banks and hospitals, could not connect. Police and the Singapore Civil Defence Force urged the public to contact them by SMS instead.
The two S$1 million penalties concerned different services and failures. The earlier incident interrupted landline calls, including access to emergency and public services. The March incidents affected mobile connectivity and services that rely on it.
Singapore has imposed other substantial telecommunications penalties. Singtel was fined S$500,000 following a December 2016 fibre broadband outage affecting around 490,000 users. A fire at its Bukit Panjang exchange in October 2013 led to a S$6 million fine in May 2014, the largest telecommunications fine in Singapore's history.
The 2013 fire disrupted phone lines, banking payments and internet services, with full restoration taking eight days. M1 and StarHub were also fined S$400,000 and S$210,000 respectively in 2020 for internet disruptions during the COVID-19 circuit breaker period, when many people depended on home connections for work and school.
Key Points
- IMDA fined Singtel S$1 million for mobile disruptions on March 16, 18 and 19, 2026, affecting about 720,000 subscribers.
- A valve left closed after maintenance and missing cooling system monitoring caused the first outage.
- Redirected 5G traffic, a disabled packet optimiser and a software bug caused the later interruptions.
- IMDA concluded that all three incidents were preventable and were not caused by a cyberattack.
- Singtel accepted the penalty, provided about S$8 million in goodwill rebates and strengthened maintenance, monitoring and failover testing.
- The penalty followed a separate S$1 million fine in December 2025 for an October 2024 landline outage.






