This is a cost, not a value.
But, maybe, it could be interpreted as a "direct use value".
If instead of asking "what is the direct use value of bitcoin" we ask "what is the direct use value of a Proof of Work" we could be on something.
The direct use value of a proof of work is the knowledge the work needed limit the output of the proves offered.
The idea is that Bitcoin is a proof-of-work token using the same protocol as Hashcash ( https://en.wikipedia.org/wiki/Hashcash ) which itself is/was a scheme for limiting spam, wherein inclusion of the Hashcash token would cost the user something (CPU time). Individual emails would be cheap to send, but mass emails would be expensive. Under this system, email receivers could dismiss email that included no POW tokens, and senders would thereby be required to include them.
In other words, you would not be able to get an email through without the token... you would need them in order to send email. (Also, most users would not find it convenient to generate these tokens on their own. They would prefer to buy the tokens from a producer.) This is the pre-existing commodity value of the tokens, that they are a necessity, or have a real-world utility, as tokens that legitimize email.