There seem to be some assumptions in your question that might not be realistic
- Apparently the machines are entirely capable of repairing themselves. That sounds obvious, but it isn't. Humans get a scratch and can all but ignore it and it heals by itself. Unless you have Clarkean Magic, when a machine gets a scratch, things seriously stop working (don't believe me? Scratch one of the copper traces on your computer's motherboard. Just one. Pick a small one.) This is a fairly massive assumption.
Detection of hardware failures has been a nonstop effort in the computer industry. I remember working on JTAG Boundary Scan tech back in the day. You'd be surprised how difficult it is to distinguish between "that wasn't expected, but it was legitimate" and "whoa! Something's broken!"
Conclusion: Humans are immensely valuable, possibly even revered to the point of enslavement, because they can fix things when they go very, very wrong.
- Apparently this is a post-scarcity economy. It is, IMO, impossible to believe the idea of a post-scarcity economy. Unlimited energy, unlimited resources, unlimited opportunity, and all basically free. In a word: boring. From this perspective, the creation of the hardware necessary to house an uploaded intelligence, the cybernetic body for periodic use, the repair and recharging solutions (infinite battery? That's even more boring), are all incredibly cheap.
In reality, gallium is really rare, as are a great many other technologically useful minerals. Things will be hard (aka "expensive") to build. Frankly, the cost of shifting someone to AI will be (and should be) very non-trivial.
Conclusion: Humans are cheap to create and cheap to maintain. That's valuable in a world where if anything goes wrong, the cost of fixing it technologically could be very high. Besides, mewonders how many of the Cyborg Overlords will want to dig in the mines. (Relevant Star Trek TOS episode)
- Finally, I think it's a whomping big assumption that people would actually want to "live forever." We all think about it during our lives, and it's a truism that the people who want to live to be 100 are usually the 99-year-olds, the the truth is much more interesting. What would happen to human sanity, morality, and ethics, once uploaded to an AI that lives, theoretically, forever? It's jumping quite a long way to the isle of conclusions to assume nothing.
In reality, people get bored. Very few people want to do the same thing for decades on end. Most people, I suspect, want to retire. And then they get bored again. That leads, methinks, to either of two things: stupification or madness. And the only thing I can think of that's worse than an expensive AI that has the cognitive and social value of a tomato is the evil-AI-takes-over-the-world madness you appear to not think would happen. I'm not even sure madness would be required for that. Sociopathy and psychopathy should be thought of as dials every human has. For most people, those dials are set within what society calls "normal" ranges. For a few, they're off-center. For (thankfully) fewer still, they're turned way high. And you just gave all those people massively powerful bodies that can live forever. Remember that Facebook bully you hated so much? That person gets to live forever in a very difficult to destroy condition. I'd sure hate to see the legal system of your world.
Conclusion: More people will avoid becoming cyborgs than you might think. Unless forced to it, I suspect many if not most will remain human just because they enjoy intercourse. And beer.
- You're assuming that because computers can perform arithmetic faster than humans, that the proverbial positronic brain can "think" faster than humans. The only problem is that computers are really, really, really, really bad at doing anything associative.
Yes, neural network hardware and programming has come a long way, but unless you once again invoke Clarkean Magic, the problem you face is that computers can manipulate data very quickly, but they come to conclusions very slowly. In fact they're bad at it. Look how long it's taken to get even basic facial recognition, and it's still beaten regularly by people growing beards, wearing makeup, or puffing their cheeks when the image was captured. Humans are incredibly good at visual and auditory pattern matching and we're amazing when it comes to putting the proverbial 2 and 2 together.
Conclusion: Jumping into the AI pouch isn't all it's cracked up to be. There are pros and cons to being an AI just as there are to being human. You loose something, becoming a computer simulation. You lose a spark of intuition that lets you realize you're in danger when a computer's "judgement" based on protocols and statistical profiles disagrees. Some call that having a soul. Computers don't have them.
And Finally: From a certain point of view, this question is a bit erroneous. You did not completely explain all the rules of being an AI (its advantages and disadvantages) and, therefore, we really can't tell you what a human is worth in comparison.
The trouble is — it's your world. If you want humans to be valueless you can always come up with rules to force that to be. You literally could tick off the entries in my list and say, "yeah, my robots are better than that...," in which case, humans are worthless.
But that's a really boring story. It's just another kind of godlike character that your average reader can't relate to. One of the most brilliant moments of dialog from the Matrix trilogy was this:
Did you know that the first Matrix was designed to be a perfect human world? Where none suffered, where everyone would be happy. It was a disaster. No one would accept the program.
Neither will your readers.
We already know that computers are flawed and that it's impossible for imperfect humanity to create a perfect machine. So, the real questions you should be answering yourself are, "what are the limitations and restrictions of being an AI in my world? What is the price people pay for cybernetic immortality?"
And if you keep finding roses, you haven't found the right answers. If you have not done so, I strongly recommend reading Asimov's The Bicentennial Man.