There are, but you are ignoring them, you simply did not do any research or it is simply hidden.

The success of Common Lisp today is relatively small, but a careful reader could find a few interesting applications of it like the scheduling system for the Hubble Space Telescope, the design software for the Boeing 747 (and other aircrafts), the software for the Roomba, the software for DWAVE's quantum computer, the crew scheduling software for the London subway, chip design verification software at Intel (and other chip companies), ...

There are some old application platforms which survived. For example Cyc, an attempt to provide common sense AI to computers, is under continuous development since the mid 80s. The company Cycorp has 50+ employees, is very secretive and you need to guess who pays for it. Customers are among others the United States Department of Defense, DARPA, NSA, and CIA. They are using it for various applications.

Note also that prototyping software was for a long time an application area for Lisp. Have relatively small teams develop a prototype and make it a product once the idea is validated. Example: Patrick Dussud wrote the core of the first Microsoft CLR (.Net Common Language Runtime) garbage collector in Lisp. The code was then automatically translated to C (IIRC) and enhanced from there after some time. Lisp now is no longer used and the GC has a lot of new features, but the first working versions came from that Lisp code.

Not necessarily. There are other possible explanations of Lisp's relative lack of commercial success, not least of which is the fact that a widespread belief that "there must be something wrong with it because no one uses it" can become (and I think has become) a self-fulfilling prophecy.

But another important factor is that the Lisp community seems to attract people who are really good at tech but really bad at business. I think if someone (or, more likely, some pair of co-founders) could bridge that gap they could still kick some serious ass.