Your confusion doesn't make any sense. The point was to steal a copy of the data, not prevent the original owners from having the data any more.
> Ellis’s assignment was to break into the data center and steal around 80 servers that hosted the incriminating files
The hit was contracted by the bankers who broke the law. Why would they need a copy of the evidence?
If the goal was to know what the prosecution has on them they took the worst route. Hiring petty criminals to steal the damn servers and giving them more details than they need to know is a stupid idea. Instead of hacking the data out of the servers (which would have been even easier in 2007), you now have a trail of inconvenient witnesses who can't wait to spill the secrets for a lighter sentence. He might have been just a decoy for something bigger.
The story smells like a book promotion and a petty criminal further bolstering a high profile (but stupid) robbery.
However I take issue with your argument with "hacking would have been easier". DCs in the 2000s were (and to a large extent, still are) extremely easy to access. If the data at stake were actual "banking" data, in the 00s, it could have been still largely "air-gapped" (in the loose sense) from the internet. Hacking would have required a lot of lateral movement. And experts.
Hence, a large trail. Arguably larger than "hey bobby, here's a wad of cash, go in that building, steal these boxes in corridor A"
I also take issue with the "physical locations of DCs are still secrets" angle of the article.
Anyone working in a DC-adjacent industry and is curious enough can locate the back-entrance to any DC they have some business with. If it is a secret, it's not a well kept one (for good, operational reasons!)
Sites like DC Map give you the exact physical location, for instance.
The story is entertaining but it has clearly been romanticized.
[1]https://us.macmillan.com/books/9780374619794/cloudthief/ [2]https://www.amazon.com/gp/product/B0C9YVP3X2