They built something and turns out that no one wants to use it. Happens all the time.
Happens all the time, especially if no one sees a point in using it. In other words, it didn't solve anyone's serious valuable problem. Which is something that those people that have been saying "couldn't you just do the same thing with a database" already knew.
Crypto went thru a lot of hype-memes, like NFTs, staking etc etc. But I still remember the older "nonono this is about blockchain as an underlying technology" meme.
Yup, that's pretty much the question that has bedeviled all blockchain "innovation" since day one. Cryptocurrency still works (for some value of works) because it's self-contained, but as soon as you need to track anything in the real world you run into the oracle problem: how can you trust that what the blockchain says about the world is correct? (Spoiler, you can't.)
Yep. The hype cycle at work; if you bamboozle enough people, even skeptics have to commit resources to the latest stupidity. Otherwise you're left without a "blockchain strategy" and the reason-by-checkbox types will think that's a bad thing. Unfortunately, as Gartner has well demonstrated, that's a substantial fraction of the market, so you can't ignore it.
The problem is "producing sticky, heavyweight enterprise IBM products to invoice customers with".
This retirement of the product might be more aligned with the expected time between major software replacement initiatives at typical customers than with the hype cycle of blockchains: IBM needs to sell some successor to TradeLens. Or, more simply, Maersk actually needs better software than TradeLens for themselves.
So... they successfully built the solution, but then failed to build a suitable problem for it?