On 5 October 2026, Etisalat (the e& group, reverting to its original name for its 50th anniversary) launched a project to raise the group's international connectivity capacity from about 20 Tbps to more than 500 Tbps by 2030, a rise of more than 25 times. Sheikh Mansour bin Zayed Al Nahyan attended. According to Khaleej Times, the expanded network is meant to serve UAE AI campuses, global cloud providers and the group's own customers, using diverse, low-latency routes. No investment figure has been published.
The case for state-steered infrastructure
The strongest argument for heavy state involvement is that international capacity is a national-security asset. When a cable is cut or a landing station fails, every bank, hospital and cloud tenant behind it feels it at once. Private operators, left alone, tend to underbuild redundancy because a spare route earns nothing until the day it is needed. A regulator that sets resilience conditions, and a state-linked shareholder that can wait out slow payback, can correct that gap in a way a purely commercial market might not.
The UAE has arranged things along these lines. Sheikh Mansour was briefed in his capacity as chairman of the Emirates Investment Authority, which is the group's main shareholder. Emirates247, relaying the state news agency WAM's report, describes the project as the group's first major commitment under a new four-engine strategy covering telecom, AI and business solutions, infrastructure and fintech.
Regulation arrived first, and it is outcome-based
The more interesting detail is the regulatory backdrop. On 12 August 2026 the Telecommunications and Digital Government Regulatory Authority announced that its board had renewed the public telecom licences of e& and du for 20 years, from 9 August 2026 to 8 August 2046, replacing the 2006 licences. TDRA says the licences treat networks as critical national infrastructure and require licensees to diversify international connectivity across multiple terrestrial systems and submarine cables and to eliminate single points of failure.
The licence text is public. The du licence, No. 3/2026, issued under Federal Law by Decree No. 3 of 2003, devotes Article 12 to network resilience, redundancy and international connectivity. Its table of contents lists sections on diversified international connectivity, elimination of single points of failure, disaster-recovery and failover mechanisms, testing, and compliance with TDRA directions. I could read only the contents pages, not the full clauses, so the exact thresholds are not something this article can characterise. The e& licence is described by TDRA as renewed on the same basis, but I did not review its text.
That design matters for a pro-innovation reading. The obligations are framed around outcomes (no single point of failure, tested recovery) rather than prescribing which cable, vendor or data centre to use. An operator remains free to decide how to reach 500 Tbps. The twenty-year term also helps. Subsea and terrestrial capacity are slow-payback assets, and long licence certainty lowers the risk premium on them. TDRA itself ties the renewal to the Digital Economy Strategy, which aims to lift the sector's share of GDP from 9.7% in 2022 to 19.4% over ten years.
What the announcement does not tell us
Several things are unknown. There is no investment figure, no breakdown between subsea, terrestrial and satellite routes, and no named landing points or partners in the sources reviewed. Nor do they say how much of the 500 Tbps is lit capacity versus potential capacity, a distinction that often separates headline numbers from usable throughput. The 25-fold figure is a target for 2030, not a delivered result. Readers should treat it as a statement of intent from the operator and its principal shareholder.
There is also an openness question. The EFF's September 2026 essay on digital sovereignty notes that governments across the global majority invoke sovereignty to argue for control over data and communications infrastructure, and that the same framing can be used to fragment or restrict internet access. It also argues that local hosting alone does not give users control. Nothing in the Etisalat announcement suggests restriction. Capacity at this scale, built to serve global cloud providers, is a bet on more traffic flowing in and out of the country, not less. But the same infrastructure sits under licences that, like any telecom licence, include public-interest and national-security provisions (Article 8 in the du licence's contents), so the openness of the pipes depends on how those powers are used.
A proportionate model worth watching
The sound lesson is narrow. When a regulator specifies what must be true of a network (diverse routes, no single point of failure, tested recovery) and leaves the engineering and the capital allocation to the operator, the result can be both safer and faster-growing. Compare this with data-localisation mandates, which raise costs without improving resilience. Markets that want AI campuses need power, land and, above all, bandwidth to the rest of the world. Capacity is the scarce input, and the UAE is moving to supply it in advance of demand.
The test will come between now and 2030. Watch for three things: whether capacity is published in verifiable terms, whether routes are genuinely geographically diverse, and whether the licence's interconnection and anti-competitive-practice provisions keep wholesale access open to du, cloud providers and smaller carriers. If they do, a 25-fold expansion will serve the whole market rather than one incumbent.