Ah, so this is a passion-project for you, and also an open-source project that you're not planning to get rich from. That changes the metrics a bit.
In that case, the two main hurdles I see are:
1) Why would bicycle manufacturers do something different when what they have sells well enough and doesn't require taking any new risks? Big businesses hate taking risks, because everyone wants to keep their steady jobs. (though having just ridden my Christini with its derailleur drivetrain after a few weeks of riding my new IGH bike, I can say the stuttering in the drivetrain caused by shifting sprockets was unexpectedly jarring, and I agree if more people got to ride gearbox bikes of any kind they'd probably prefer the "feel".) Admittedly there's nothing you can do about institutional inertia, short of paying millions of people to try your product, or launching a fabulously-persuasive advertising campaign.
2) Humans are all torque and no horsepower, which is a big problem for any drivetrain design. If you've ever had the opportunity to take apart a car transmission (which I'm guessing you have), you know the gears are big, chunky, and heavy. That is entirely because the gears need to withstand the torque the engine generates -- it's actually a disadvantage to have heavy gears spinning at high speed, but it's much more of a disadvantage to have broken gears not spinning at all. Diesel transmissions are even beefier, for the same reason. A reasonably strong human struggles to generate 0.5 horsepower, but even my desk-job physique can generate over 120lb-ft of torque (as much as a small car engine) just by
standing on the forward pedal on my bike -- my torque output goes up even more when I brace against the handlebars and push hard on the pedals.
That is a huge problem for designing bicycle drivetrains, because they have to be strong enough to withstand as much torque as a small car engine can produce, but also be light enough that they don't make pedaling uphill significantly harder than it needs to be. The saving grace is that bicycles don't need to work as long between repairs, so their drivetrains can be made out of normal metals instead of gold-plated unobtainium. Derailleur drivetrains further improve their torque-carrying capacity by using large gears that can be made out of even lighter metals, whereas IGH gearboxes don't have that option, so they have to use hardened carbon-steel instead. IGH gearboxes
do benefit from the torque-reducing effect of having a large drive sprocket connected to a small driven sprocket, so the full torque of the rider is reduced by ~66% in exchange for 3x faster rotational speed at the hub. That means IGH gears can at least be made small to minimize the weight penalty of using an IGH, without also needing to be made from gold-plated unobtainium to handle the rider's torque. But a mid-frame gearbox doesn't get the benefit of being on the output side of a torque-reducing/speed-amplifying chain-drive. A mid-frame gearbox is the worst of both worlds, in that respect; it has to handle as much torque as a small car engine can produce, while also being light enough
and durable enough that the rider won't regret buying one.
I'm sure you've heard the old engineering joke "Strong, light, fast: pick two." In this case we're talking about "strong, light, durable" instead, and mid-frame gearboxes demand that you pick all three. In that case, the unavoidable tradeoff is that it will be horrendously expensive. Lightweight metals that can handle high loads for lots of miles without deforming or cracking are horrendously expensive, not just in materials costs, but also in machining costs.
Also this one might surprise you, but fancy alloys are also nearly impossible to lubricate effectively. I don't entirely understand the chemical reasons for that, but the 50-cent explanation that I've scraped together is: the atoms in high-strength alloys are so tightly-bound to each other that their outer electron shells don't have enough available electrostatic charge to attract oil molecules very well. (hopefully there aren't any physicists or chemists reading this, because they'll probably have a nosebleed from the crudeness of that explanation.) 6Al4V aerospace-grade titanium is the metal that comes to mind for the gearbox you want to build, and I know from past experience in other hobbies that cutting titanium wears-out tools super fast, and lubricating titanium is an exercise in frustration. I spent a couple years figuring out how to lubricate titanium screw-threads for a different hobby, and the best I could do was to reduce the awful scraping sensation to a minimum. I actually made some money selling small batches of a special grease that I mixed in my workshop at home for that purpose, because the minimal level of lubrication was still better than what anyone else had managed to achieve, including the big lube-manufacturing companies. But that wouldn't be good enough for the heavily-loaded surfaces of the gear teeth in your gearbox; you'd probably need to use fancy surface coatings like AlTiN or DLC that would easily double the cost of your gearbox design, and those have an irritating tendency to flake off over time. Not such a great thing for that lifetime warranty you're proposing.
Buuuut, you're not the first guy to chase the holy grail in their chosen hobby, so keep at it. I like being proven wrong when I say something won't work.
It reminds me of when I was a kid and I still believed in magic.
