When was that ever true? Businesses have always improved something that exists. They don't really exist to create "new inventions".
Sometimes they start that way, but those are the exception, not the rule. For example, Xerox (xerography) and Google.
Apple did a lot of inventing, but they also clearly stood on the shoulders of Xerox PARC, which did a lot more inventing.
Every other company I can think of is not based on "inventions", but more "bringing to market". Netscape was a great tech company, but they didn't invent the web, etc.
Sure, lots of the big tech companies used tech invented elsewhere, but still, they were bringing new technologies to market that hadn't been seen before. Internet access through browsers was big and new. Google was a huge step over the competition. Apple iPods and iPhones big and new. Lower down: databases, network hardware, and so on.
Now it seems like a bunch of micro-optimizations, ways to avoid legislation, or lifestyle products. Not new forms of technology entering the market that really do change everything. The 'change everything' technologies, like Musks' ventures, seem to be a bunch of moonshots (or marshots).
Like you, I assume, I'd love to see more remarkable innovation, but the reality of innovation is that it's mostly a slow grind where most of the value is in optimizing the current paradigm until it breaks and we have to figure out a new 10x paradigm.
There are also dynamic effects where incremental innovation makes future radical/disruptive innovation easier.
Since most technologies are combinations of previous technologies, it's rare to find completely new technologies/inventions, and that's ok.
The fact that to reach their potential, innovation must be diffused through society also bounds the pace at which it can go. As that infrastructure develops, standards emerge that allow for innovation to be distributed faster. But then, in turn, as the underlying technologies improve incrementally, they also ossify as an infrastructure, which might delay bigger innovations.
The question of where the disruption/incrementalism equilibrium lies and what its ideal position can be is very interesting and, I find, understudied.
Everyone is trying driving cars and fully doing so is impossible. Supersonic flight already been done with Concorde. Custom microorganisms are potentially interesting but also seem to lack killer app, except killer bugs.
You're trying to have it both ways - self driving is not an invention because they have not done it yet, supersonic is not because it's done already. By that measure nothing would seem to count. Also cheap supersonic has not been done - Concorde was expensive.
I think self driving is a poorly thought out goal, and driven more by the cult of AI than by sound engineering. I don't believe it is actually possible, except in very restricted areas. It's more one of those things that someone took from a sci fi show than a real engineering problem. I'm not a fan of those kinds of technologies. Making Concorde cheaper is just optimizing something.
Sometimes they start that way, but those are the exception, not the rule. For example, Xerox (xerography) and Google.
Apple did a lot of inventing, but they also clearly stood on the shoulders of Xerox PARC, which did a lot more inventing.
Every other company I can think of is not based on "inventions", but more "bringing to market". Netscape was a great tech company, but they didn't invent the web, etc.