My bad; QFT actually postulates locality. I was thinking about the casual set theory which strives to solve some of the QFT's difficulties, and where locality is an emergent / statistical phenomenon rather than a postulated condition.
If you couple your system to a heat path that is at rest wrt a specific Lorentz frame, you of course lose Lorentz incariance. On the other hand the lagrangian of the standard model itself is to my knowledge fully Lorentz invariant.
I don't know what they talk about there, but it sounds like some kind of thermodynamic approximation is involved there. Does thermodynamics survive Lorentz transformation?
How so? QFT is Lorentz invariant. Even has such a thing as the norm flux.