Where traditional subsea cables carry capacity in terabits per second, Meta’s newly announced Petal system will deliver 1 petabit per second. Petal will run roughly 7,000km between the US and France and is expected to enter service in 2029. Meta says it will double the capacity of today’s most advanced transoceanic cables, the largest generational jump any system of its kind has made, without a proportional increase in power or physical infrastructure.
Meta is building the cable with NEC Corporation, which is the turnkey supplier responsible for manufacturing and installing the system, and Sumitomo Electric Industries, which is providing the fibre. Orange is working with Meta on landing the cable on France’s Atlantic coast, including the terrestrial link into the European network. Sumitomo’s announcement confirms that Meta will fund and operate the system.
Doubling capacity without doubling the cable The headline number matters less than the method. For the past decade, the industry has added capacity mainly by packing more fibre pairs into each cable, a technique known as spatial division multiplexing. Meta’s own transatlantic systems show the progression: Marea had eight fibre pairs, Amitié 16 and Anjana 24, making Anjana the first 0.5 Pbps transatlantic system.
Meta weighed three ways to double Anjana’s capacity: going to 48 fibre pairs; adding L-band transmission on 24 pairs; adopting 2-core fibre. It chose the third, putting two light-carrying cores in each fibre. That gives 24 fibre pairs the equivalent of 48.
That makes Petal the first subsea system to deploy multi-core fibre at scale, moving the technology from the lab into commercial service. Sumitomo says the 2-core fibre achieves ultra-low loss and near-immeasurable crosstalk between cores, giving performance close to single-core fibre. The repeaters are where the engineering gets practical.
A 7,000km cable typically needs around 100 of them. Petal’s repeaters use a fan-in, fan-out interface to split each 2-core fibre into two single-core fibres for amplification, then recombine them. That keeps the proven reliability of single-core amplifiers.
Crucially, Petal stays within existing 18kV power-feeding limits. Going above that would force a requalification of the wider subsea supply chain. In other words, Meta has doubled capacity without asking the industry to rebuild its equipment base.
Why it matters beyond Meta Cost per bit. Meta says carrying a petabit through one cable significantly reduces materials, resources and carbon footprint compared with building two 0.5 Pbps systems. At a time when cable-laying vessels and manufacturing slots are scarce, as Capacity reported in its analysis of why hyperscalers are laying their own subsea cables, getting twice the capacity from one build is a significant efficiency.
A new design baseline. Meta frames Petal as moving multi-core fibre from experimental to mainstream, and says it wants the wider ecosystem to invest alongside it. If NEC and Sumitomo can industrialise the 2-core supply chain, the approach becomes available to other builders, not just Meta.
The transatlantic race. Petal lands in a crowded field: EXA Infrastructure’s Exa Meridian, announced this month, will run 24 fibre pairs with more than 500 Tbps of capacity between New Jersey and Brean in the UK, targeting service in Q4 2029, as Capacity reported. AWS’s wholly-owned Fastnet is designed for more than 320 Tbps across 16 fibre pairs.
Petal would roughly double either on a single system. That sharpens the difference between hyperscalers building for their own traffic and carrier-neutral operators selling fibre pairs to the same customers. France as a landing hub.
Orange’s Jean-Louis Le Roux called reaching a petabit “25 years after the terabit milestone” a significant breakthrough. He said the project extends connectivity to other European countries and underlines Orange’s focus on securing the terrestrial segment of those links. The caveats Petal is three years from service, and multi-core fibre has never been deployed at this distance commercially, so execution risk is real.
Meta has not yet disclosed the US landing point. Meta’s public framing is also worth noting. Its announcement emphasises growing demand for global connectivity and resilience rather than AI specifically.
That contrasts with the explicitly AI-driven pitch behind several other recent builds, a distinction Capacity explored in its analysis of how AI traffic is reshaping subsea cable networks. Petal’s arrival also raises the ownership question now running through European policy debates: how much of Europe’s transatlantic capacity should sit on systems owned and operated by a single US company. The question will be on the agenda at Capacity Europe in London next month, where Meta’s Ugne Kleinauskaite joins the “Chokepoints & Shockwaves” subsea panel and Orange’s Le Roux sits on “The New Rules of Network Design”.
RELATED STORIES How is AI traffic reshaping subsea cable networks? Why hyperscalers are laying their own subsea cables Exa Infrastructure unveils Exa Meridian, the first single-operator transatlantic cable in a decade Six European submarine cables to watch in 2026 Capacity Europe 2026 13 October 2026 The 25th anniversary edition of Capacity Europe 2025 will bring together 3,500+ decision-makers from the global connectivity and digital infrastructure community. Register now See event details Saf Malik Senior Content and Insights Manager