One Cable Cut, 177 Milliseconds of Latency: Inside the Philippines Internet Slowdown

On the evening of October 2, at around 9:30 p.m. local time, two submarine cable segments connecting the Philippines to Singapore and Hong Kong went down within the same window. Millions of internet users across the country felt it almost immediately: slow, intermittent service, especially when reaching websites and apps hosted abroad. PLDT, the country's largest carrier, confirmed a loss of roughly 300 Gbps of international capacity on the affected links. The Department of Information and Communications Technology said it was investigating the cause, which has still not been officially confirmed.

By October 4, the big carriers were declaring victory. PLDT said services had fully normalized nationwide in an advisory issued at 4:30 p.m. Globe said its mobile and broadband services were restored after the Philippine cable landing station reactivated the system and normalized fiber connections. Converge said it had rerouted its Singapore traffic through other submarine cable systems. And yet a careful reading of the public record suggests that "restored" meant three different things: a measured network path, one operator's segment, and the service a customer actually buys. They did not come back at the same time, and they are not the same thing.

This is a story about how the internet actually works, told through one broken cable.

How we know what broke

The first public signal was not a repair report. It was a measurement. PhilSTAR Tech reported that the University of the Philippines Diliman Network Helpdesk had been warned that an undersea segment between the Philippines and Singapore was offline, with higher latency and intermittent performance expected. The monitoring outfit CableStatus put numbers on it: roughly 177 milliseconds of latency on the Philippines-Singapore route, against a recent baseline of about 53 milliseconds, measured across 28 RIPE Atlas probes. That is the network equivalent of taking a temperature: the probes could not prove a physical cable fault or name the affected system, but they showed exactly where the network was hurting.

The operator-level detail came a day later. InfiniVAN told PhilSTAR Tech that its C2C Philippines-Singapore segment had been restored and monitored since 12:48 p.m. on October 3, while the C2C Hong Kong link was under maintenance, limiting redundancy. Converge said it had stabilized service quickly by rerouting traffic, mostly through the Bifrost cable system. Globe said its services were fully restored on Sunday, October 4, after traffic resumed across the affected segment following reactivation of the Philippine landing station, and that it was still coordinating with international partners to determine the exact nature and location of the fault.

Restored is a word with three different meanings on a submarine network, and they rarely arrive at the same time.

How the traffic survived anyway

Here is the part of the story that never makes the panic headlines. The internet did not go down. It bent. When the segments failed, carriers shifted traffic to other paths, and the global routing system absorbed the blow. Border Gateway Protocol, the protocol that decides which path packets take between networks, does not care which cable is pretty. It cares which route is announced and reachable. When one path disappears, routers converge on the next best one in seconds, with no human in the loop.

That is why the symptoms were slowness and intermittence rather than a blackout. Traffic from Manila bound for Singapore got squeezed through whatever capacity remained: other cables, other landing stations, longer routes. PLDT noted that most traffic had been rerouted through alternative paths even as it worked to normalize connectivity. The rerouting worked, but capacity is capacity, and a country that loses 300 Gbps of international bandwidth does not get it back from a routing table update. It gets it back from a different pipe, with different limits.

Why the cable will take weeks to actually fix

The service is back. The cable is not. That gap, between restoration and repair, confuses every cable incident, and it is worth understanding because it explains most of the physical internet. A damaged submarine cable is fixed by a specialized repair ship that must locate the fault, grapple the cable off the seabed, haul it aboard, splice in a replacement section, and lay it back down. Finding a ship, getting it to the site, and waiting for a workable weather window routinely stretches the job into weeks.

This is also why the cause matters and why it stays uncertain for so long. Most subsea faults are mundane: ship anchors dragging across the seabed in busy straits, fishing trawlers, undersea landslides. Sabotage makes headlines, but accident is the industry's default explanation, and experts have said the same about recent Red Sea cuts. The DICT has not confirmed what severed the Philippine segments, and until an inspection examines the fault, nobody else can either. The resilience of the modern internet does not come from cables that never break. It comes from there being enough of them that one break does not take a country offline.

A pattern, not an anomaly

The Philippines is not alone in learning this lesson. In September 2025, cuts to the SMW4 and IMEWE cable systems near Jeddah degraded connectivity across South Asia, the Gulf, and parts of Africa, with monitoring group NetBlocks flagging the disruption and Microsoft warning that some Azure traffic through the Middle East would see increased latency. The Red Sea corridor carries an estimated 17 percent of global internet traffic, which is why a fault in one crowded strait ripples across continents.

The pattern has not gone unnoticed. A Manifold prediction market asking whether a major internet outage caused by breaking undersea cables will occur by the end of 2026 was trading at a 90 percent implied probability of YES as of October 7, 2026, with the Philippine event as the latest data point. It is a play-money market, not a regulated exchange, and its resolution hinges on whether a two-day, single-country slowdown clears the bar for "major." But the sentiment behind the number is real: subsea faults happen with regularity, and the next one is a matter of when, not if.

That is the perspective worth keeping. The internet is not a cloud. It is a collection of glass fibers lying on the ocean floor, tended by a small fleet of ships, measured by volunteers running probes, and rerouted by protocols written before most of its users were born. The Philippines incident is a reminder that the whole system is both more fragile and more resilient than it looks, which is exactly how engineered systems are supposed to behave: they fail, they route around the failure, and then, weeks later, a ship sails out to make the ocean whole again.