So the big story in tech today is about Elon Musk combining SpaceX with xAI in a $1.25 trillion deal. Except everyone else has written about that, and everyone knows what it is really about: money for new rope, or else data centres in space. Pick your poison. No, the best story today (IMHO) is about the subsea cable market, and the challenges it faces to deliver enough capacity for the kinds of AI shenanigans Musk et al have up their sleeves. As an aside, I’m recently back from PTC in Hawaii, where the locals are apoplectic about Zuckerberg’s mile-deep end-of-times bunker on Maui (plus other concessions to tourism and military). 

The subsea story is from PTC, too, where a panel of hyperscalers (Google, Meta; there we are again) and wholesalers (Exa Infrastructure, Southern Cross) laid bare a stark reality: the planet’s under-water data highways are at breaking point already. Legacy systems from the early 2000s are end‑of‑life, just as hyperscalers demand more bandwidth. Traditional wholesalers are finding it hard to justify new builds as AI players, and hyperscalers, with deep-enough pockets to build in the oceans (and space), are reluctant to make capacity available on the open market. And then there’s long lead times, supply‑chain bottlenecks, and disjointedness with data centre builds. 

Meta’s gigantic global Project Waterworth initiative and Google’s trans‑Pacific Tabua system show huge ambition, but they have faced regulatory, logistical, and geopolitical hurdles. Red-tape around the industry’s new ‘critical infrastructure’ designation needs untangling too. It is a wonder anything gets done in the water. But at least it’s easier than terrestrial fibre, the panel says; at least it’s not a wild west like the data centre market. And there is plenty of growth; Omdia was at PTC, too, and says the subsea ‘optoelectronics’ market surged 32.5% (to $337 million) in the third quarter of last year as the industry’s focus shifts toward mesh architectures for greater resilience.