Telegram Data Center Outage in Singapore Affects Seven Countries
On September 8, 2026, at around 10:20 AM Singapore time, the messaging app Telegram experienced a widespread service outage in the Asia-Pacific region. Users in Singapore, Indonesia, the Philippines, India, Malaysia, Vietnam, and Thailand reported being unable to send or receive messages, load conversations, or use the app.
According to data from outage monitoring platform Downdetector, the peak report volume in Singapore reached about 4,000, with Indonesia at 1,284, the Philippines at 933, India at 725, Malaysia at 544, Vietnam at 530, and Thailand at 253. Approximately 50% of the reports involved overall app functionality issues, 40% involved message sending and receiving failures, and 5% involved login problems.
Affected users described the interface being stuck on "Connecting" for long periods, sent messages only showing a single tick (not delivered), media files failing to download, and some conversation threads not completing sending, with channel content unable to refresh.
Telegram's global infrastructure consists of only five physical data centers: DC1 and DC3 in Miami, USA; DC2 and DC4 in Amsterdam, Netherlands; and DC5 in Singapore. User accounts are permanently bound to one of these data centers based on the country code of their phone number at registration, and cannot be changed. This outage was identified as occurring only at the DC5 node serving the Asia-Pacific region, while the other four data centers operated normally. DC5 has a history of instability, with users even creating monitoring tools to track its status. As of now, Telegram has not released any explanation regarding the cause of the outage.
Due to Telegram's architecture of permanently binding accounts to a single data center rather than employing multi-node disaster recovery, when a single node like DC5 fails, the impact is concentrated in its fixed service area rather than being dispersed among global users. This means that the Southeast Asia and South Asia regions, which have a large number of Telegram users, repeatedly face the risk of regional outages due to centralized infrastructure. This is particularly relevant as founder Pavel Durov has been promoting the integration of a TON blockchain-based wallet into Telegram, aiming to embed payment functions for approximately 1 billion monthly active users.
According to Downdetector data, the report volume had decreased from the peak of 4,000 to about 350 around 11:30 AM, with the outage gradually easing about an hour after it began, and users were able to resume normal use without needing to update the client.
Source: Public Information
ABAB AI Insight
Pavel Durov, the founder of Telegram, has long established the company's infrastructure with only five physical data centers: two in Miami, USA, two in Amsterdam, Netherlands, and one in Singapore. User accounts are permanently assigned and bound to one of these nodes based on their phone number's country code at registration. This architecture has been in operation for many years, with DC5, responsible for the Asia-Pacific region, consistently being the least stable node, to the extent that the user community has built monitoring tools to track its status. Meanwhile, Durov has been pushing for the integration of a TON blockchain-based cryptocurrency wallet into Telegram in recent years, aiming to enable approximately 1 billion monthly active users to utilize its payment functions, and he has publicly expressed a desire for the company to achieve profitability.
This outage itself did not involve any financial transactions, but the underlying capital logic is clear: Telegram has not built a distributed disaster recovery cloud infrastructure for nearly 1 billion users like other large-scale platforms; instead, it has maintained a streamlined layout of only five data centers for a long time. While this approach can lower capital expenditures for infrastructure, the cost is that any issue at one node results in a concentrated failure affecting an entire region rather than a minor disturbance dispersed globally. The Southeast Asia and South Asia regions, which bear this risk, are precisely where Durov's TON wallet and payment business need user trust to truly take off.
This "few nodes, regionally concentrated" architecture has similar precedents in other industries: some social platforms experience localized service interruptions when edge nodes in specific regions fail, and systems in the electricity and telecommunications industries that rely on a few regional substations or backbone nodes to support the load of entire areas also exhibit the same pattern of "fewer nodes, greater impact from single point failures." Telegram's current position in the communications industry is quite unique: with nearly 1 billion users, it is considered a large-scale platform, but its infrastructure investment is more akin to that of a medium-sized company, and this disparity becomes evident when such failures occur.
The structural judgment ultimately pointed to by this event is capital concentration: when a platform chooses to concentrate its physical infrastructure in a few regional nodes rather than spreading risk across numerous redundant sites, it can effectively lower fixed costs and operational complexity, but it also concentrates the risk of failure at any single point that may encounter issues. The fewer the nodes, the larger the user impact when a single node fails; this is the fundamental reason why regional chain failures repeatedly occur whenever a single facility, whether a data center, power substation, or undersea cable, is required to independently handle the traffic of an entire area without a true backup path.
ABAB News · Cognitive Law
The redundancy costs saved ultimately turn into downtime.
Fewer nodes mean a larger range of impact from failures.
The reliability of 1 billion users depends on the weakest node.