Summary

  • TDRA’s renewed licences for e& and du run from 9 August 2026 through 8 August 2046 and, according to the regulator’s announcement, require geographically diverse international links, fewer single points of failure and regularly tested recovery.
  • Two cable systems count as resilience only when their end-to-end paths avoid the same disabling dependencies and can carry service after a failure. Public evidence reviewed here does not show either operator’s route topology or test results.

A licence date is not a recovery result

On 12 August 2026, the Telecommunications and Digital Government Regulatory Authority (TDRA) announced a 20-year renewal of the public telecommunications licences held by Emirates Telecommunications Group Company PJSC (e&) and Emirates Integrated Telecommunications Company PJSC (du). The announced term began on 9 August and ends on 8 August 2046, replacing licences issued in 2006. TDRA describes the renewal as a stable framework for infrastructure investment. It also lists a different kind of obligation: geographic diversity for international connectivity through multiple terrestrial systems and submarine cables, the elimination of single points of failure, disaster-recovery mechanisms that are tested regularly, and qualified national talent able to operate and restore networks in exceptional circumstances. Those are the regulator’s stated terms in its renewal announcement.

That combination changes the question a reader should ask. A long authorisation can make capital planning easier, but duration does not make a network recoverable. The operational promise sits in the topology, the capacity available during a cut, and the people and procedures that move traffic and restore service. The licence term is the horizon; resilience is what the system can do on an ordinary day when one of its assumptions fails.

Count failure domains, not cable names

A route drawing can show two submarine cables and still conceal one point of failure. The systems may approach the coast on different seabed paths, then meet in the same landing facility, rely on a common power source, enter the same inland duct, or converge on one international gateway. A fault at that shared point can remove both lines from the service path. This is a general engineering possibility, not a finding about the infrastructure of e& or du. The public sources reviewed for this article do not disclose their route maps or common dependencies.

TDRA’s own International Telecommunications Cable Regulations help explain why the end-to-end view matters. They define a cable landing station as a facility that houses power-feed or termination equipment and interconnects submarine cables with land-based networks. They also define a terrestrial international cable gateway that connects international cables to UAE networks. The cable is therefore one segment in a chain. A service can still depend on the station, its power and access, the inland path, the gateway, and the operator’s ability to switch traffic.

The International Telecommunication Union’s 2026 recommendations treat choke points, route congestion and clustered landings as subjects for resilience stress tests. That guidance gives substance to the word “geographically” in TDRA’s announcement. Diversity must extend beyond naming more systems: it has to separate the conditions that can disable them. A second supplier or cable label is not evidence of a second path if the outage boundary is still shared.

Test the service that survives

The announced obligations point to four connected tests. First, can an operator trace each international service from cable to customer, including the land segment and facilities where routes meet? Second, which single event can impair more than one route? Third, when one path is lost, does the remaining path have enough usable capacity, power and reachable staff to carry the service that the plan says it will protect? Fourth, do exercises demonstrate that traffic moves and recovery proceeds, or do they only show that a plan and contact list exist?

These are questions for an assurance programme, not thresholds published in the renewal release. That release says TDRA will monitor compliance, but it does not specify a public route-independence score, a minimum failover capacity, a recovery-time target, or the evidence that operators must publish. The difference matters. A count of routes is easy to report; a tested failure boundary is more useful to the customer who needs service after the count has been reduced by an outage.

TDRA already describes a broader business-resilience and continuity programme. It says the sector’s emergency, crisis and disaster plan began in 2011, and lists regular readiness audits, SADA ALBARQ simulation exercises, Shattat operational drills begun in 2020, and an annual continuity forum. Those activities establish that the regulator has a testing and coordination framework. They do not tell the public whether a particular operator passed a particular test under the new licences. The renewal creates an opportunity to connect the licence duties to observable outcomes without pretending that the public programme page is a performance certificate.

Make the long horizon legible

Redundancy costs money before it produces value. An operator must preserve alternative capacity, keep it powered and maintained, train people to use it, and accept that a low-frequency failure may test the investment only years later. If the evidence standard is merely “two cables,” the incentive is to count assets. If the standard asks which services remain available after a shared-risk failure, investment can be compared with a public outcome.

That outcome should not require publication of sensitive cable coordinates. A regulator can disclose whether paths were reviewed for shared dependencies, what capacity survived a tested loss, how long recovery took against the operator’s own target, and whether corrective actions closed. Independent auditors can inspect detailed topology while the public sees a bounded result. The ITU recommendations similarly connect stress testing with choke-point review, spares, repair agreements and cable-station security.

The licence announcement also names interconnection and access, national roaming, quality of service and subscriber protection. These conditions matter because continuity is experienced by people and businesses, not by a cable diagram. A technically diverse international path is only one part of service continuity; domestic switching, wholesale access and customer communication can decide how much of that capacity reaches a user. They complement physical diversity. They cannot make two paths independent if both fail at the same point.

For now, the evidence supports a narrow conclusion. The UAE has made route diversity, single-point-of-failure reduction, tested disaster recovery and skilled recovery staff explicit parts of a 20-year licence framework. It does not establish that the required routes have already been built or that either operator has met a test. The durable measure will be whether a loss stays local: a cable interruption should not become a landing-station outage, a gateway outage and a customer outage at once.

Sources