The Cost of Keeping Moving

A car company is building a fleet of ships. The monopoly that makes the machines that make chips says its customers are lining up to place orders for 2030. And a country at war just found out that the satellite link its drones depend on can be changed overnight, without warning. I think these are the same purchase.

Start with the ships. BYD, the world's biggest electric vehicle maker, is going from eight car carriers to eighteen by 2029 — each new vessel good for about 9,200 cars, built by China Merchants Industry. The boring filing is "vertical integration, Chinese edition: automaker buys boats." I don't think that's what this is. I think BYD is buying the right never to be told no.

Look at what renting costs. A car carrier holding about 6,500 cars runs $95,000 a day, up from $50,000 in February, according to Clarksons Research. Most Asia–Europe carriers now crawl around the Cape of Good Hope — roughly a quarter more distance, up to fourteen more days per voyage. Moving a car from China to the Middle East went from $65 to $220 per cubic metre of cargo space. And underneath the price spike sits a quiet indignity: Chinese owners hold only about 10% of global car-carrier capacity, ships on order included. The world's biggest exporter rents its exit from somebody else's fleet.

So BYD is drilling a private tap. It announced $687 million for carrier construction back in 2022, after founder Wang Chuanfu complained of a shortage of export shipping capacity; eight ships entered service between January 2024 and September 2025; this year it told investors that shipping capacity had literally held back its exports. In September it sold 180,700 cars outside China — more than double a year earlier, and nearly 40% of total monthly sales — and it is aiming at more than 2.5 million overseas sales next year.

The number everyone quotes is the estimate that owning could cut BYD's cost to ship each car by up to 40% versus chartering. I read that number differently than most people do. Eighteen ships are spectacularly inflexible capital — steel that has to sail whether or not anyone needs it.

That 40% isn't an efficiency gain. It's the premium on an option: the option to keep shipping when the shared market breaks.

Here's the part I can't leave alone. The obvious story is war insurance, and the obvious story isn't wrong — it's just not the interesting one. Stephen Gordon of Clarksons Research puts it in order:

The very strong growth in Chinese exports, fueling strong Asia-Europe fronthaul demand growth, is the biggest single factor behind the increase in charter rates. Additionally, conflict in the Middle East has caused regional disruption and inefficiencies, which have helped to further tighten markets after a two-year period of gradual normalization.

Read that twice. The biggest single force squeezing shipping is the success of Chinese exports themselves.

Which cracks the story open. At home, Chinese sales of EVs and plug-in hybrids fell 12% over the first eight months of the year after buyer tax breaks were cut — and BYD's own domestic sales fell 33%. Meanwhile its exports keep outrunning overseas registrations, which, as Benchmark Mineral Intelligence's George Whitcombe points to, means a growing stock of unsold cars sitting abroad. The ships aren't only insurance against a closed strait. They're the exit valve for a home market that is shrinking under the company's feet. The fleet is how you keep a factory full when your own country stops buying.

It's a weapon against rivals, too. De Luca's point is blunt: carmakers that charter ride the charter market's swings, and owners don't. Whitcombe adds that Chinese automakers left without their own fleets face rising shipping costs in a weaker position to grow overseas. Geely runs one carrier on a long-term charter from Zodiac Maritime. Chery has a shipyard venture at Wuhu. Lei Xing, founder of AutoXing, is certain where this ends: "Geely... will be running its own ships eventually," he said. "You can bet on that."

There is a dissent worth taking seriously. Felipe Munoz of Car Industry Analysis calls the giant carriers "mainly a marketing exercise" — "a political statement against the established carmakers," a floating billboard for speed and resources. Sure, some of it is poster. But marketing doesn't explain Geely and Chery following, and it doesn't explain the configuration. BYD ships only EVs and plug-in hybrids, which need specially configured carriers, and most of the fleet burns liquefied natural gas because ports are tightening green-fuel rules. That is pre-complying with regulations that don't fully exist yet. Vanity doesn't order a fleet that fits the rulebook of 2030.

The ladder is the point. Tesla exports from Shanghai with no hulls of its own — other companies' carriers. Geely charters long-term. BYD owns. And one rung above all of them sits ASML, whose chief executive Christophe Fouquet said on Wednesday that the company cannot fully meet demand for its chip-making tools in 2027 or 2028 — despite planning 30% capacity increases in each of those years — and that customers now show "strong interest to already place orders for 2029 and sometimes even 2030." "A lot of our discussions with our customers now are shifting in fact beyond 2027," he said. When buyers start queuing years in advance for the right to buy something, they have stopped shopping and started securing.

Three industries, one week, one instinct: stop renting the critical layer.

Which brings me to the story that shows what renting looks like from the far end.

Overnight on October 13–14, Ukraine's drone units started losing their connection. The 7th Rapid Reaction Corps said it noticed "Starlink connectivity problems caused by data transfer speed limits" — "the drones kept flying, but the interruptions made them harder to control." A unit in the Zaporizhzhia region cancelled planned operations. According to President Zelenskyy's defense tech adviser, Serhii Beskrestnov, SpaceX had quietly imposed speed limits on terminals in motion, on the military-grade service, above roughly 100 mph.

Our defence forces received no warning of such a restriction, and this led to the failure of a number of missions.

That's Beskrestnov on Telegram. And then the detail that ought to make anyone who depends on rented infrastructure sit bolt upright: Ukraine's armed forces learned about the limitation from customer support, after filing a complaint. They found out what their own war runs on by opening a help ticket.

The motive is genuinely unknown, and I'm not going to pretend otherwise. Beskrestnov says the timing "cannot be a coincidence" and is "connected to the current political situation," and Ukrainian officials note that Elon Musk is a close Trump ally as relations spiral — but nobody has offered evidence, and SpaceX and the White House haven't commented. One roundup of Wednesday's reporting makes the precise point: SpaceX has restricted service in Ukraine before, with Kyiv's agreement. What's new is the tier it hit — the military one — and the silence.

Now put that next to the ships. This is BYD's eighteen hulls seen from the other end of the telescope: a country at war, running on a private company's satellites, discovering that the tap can be adjusted unilaterally, overnight, for reasons it is not entitled to know. BYD looked at that same world and decided the only acceptable answer to "who can change the terms?" is us.

The counterargument to all of this is the boring, correct-looking one: when the wars quieten, freight rates will fall, and eighteen hulls will look like a war premium paid at the top of the market. Maybe. Options get mistimed all the time; oversized bets are a real category of error. But notice that ASML's customers aren't scared of war — they're scared of shortage — and Ukraine's problem isn't pricing at all, it's sovereignty over a setting. Three different fears, one purchase.

So here's the habit I'm taking into the deal announcements of the next decade. Ask what each purchase insures. And ask who can change the terms without telling you. Private taps are expensive, inflexible, and faintly ridiculous right up until the shared world stumbles. A startling number of serious companies have decided that is a bet worth making anyway. That tells you something no earnings call will say out loud: this is what they expect the next decade to feel like.