Coinbase is not the next Google. They provide no actual value, beyond making it easy to join a bubble. Google's value proposition in the early days was enormous, very obvious, and continues to this day.
Most Americans have not accepted that premise (and I don't exclude myself...I'm on the fence as for whether crypto is a bubble or the future of world commerce). However, I'll point out that in 1996 when Larry/Sergey/Scott Hassan were first working on Google, it was far from obvious that the web would be a big thing. Streaming audio (via RealAudio/WMP/etc.) was a year in the future; most Americans had 14.4K connections that made it impossible (my first MP3 took 2 hours to download a 5 minute song). E-commerce was clunky and sluggish, and the big "order everything online!" push was 2 years in the future. There were only 36M users online, roughly the same order as the number of Bitcoin users in the U.S. (Bitcoin has more international adoption though). Interactive websites were almost unknown, because the technology really didn't support it. Most people used the web for porn, or to lookup song lyrics and share their favorite bands, or to share scientific information.
It wasn't until the early 2000s, after the dot-com bubble burst and yet people continued using the Internet (and found a lot of new uses for it, like dating/socializing/paying for things/watching videos/sharing photos/finding lodging), that it became apparent that it was really here to stay. Not coincidentally, those were Google's biggest growth years.
No such thing is happening with bitcoin. You can't scale something to internet levels if its banned in half the worlds research labs.
If you accept it sure. However, the utility between the internet and cryptocurrencies are vastly different.
The internet provides light speed communication and seeing as we are a species built upon communication, communicating from days to fraction of a second which is also orders of magnitude cheaper (granted, not at the start but it got there and cryptocurrencies have not solved the scaling problem (cost) yet) is hugely beneficial. The benefits are completely obvious.
Cryptocurrencies provide you the ability to incentivize people you cannot trust to spend money to secure a network in the hopes that the cost of securing the network is greater than the cost of someone oversupplying your network and stealing your money.
Also, the benefits of control not being enforceable by a central authority is a bad thing in most cases and the few that are useful are outweighed by in pretty much every use case the billions of dollars it costs to run those computers to secure it.
It has been almost 10 years since the bitcoin network started, where is its killer use case?
Lastly, the billions of dollars spent to pay for the decentralized network to secure it are mostly payed for by creating new coins. At the end of the mining timeline, they will have to rely purely on transaction fees to pay for the costs to secure the network and the billions of dollars used to do so currently at a fraction of the scale Visa does will likely never outweigh the benefits of decentralization.
TLDR: Running decentralized nodes to recompute the same transactions to protect against untrusted nodes will always be more expensive than running each transaction on a single node on a trusted server.
When you say "users", what do these users "use" Bitcoin for besides speculating on its price?
Do you know of any real usage?
Ethereum is currently used for startup fundraising via ICO, for tokens in collectible games (Cryptokitties), to distribute ad revenue from browsing the web (Brave), to incentivize participation on curated social news sites (Steemit), for decentralized prediction markets (Augur), to get paid for donating computing power (Golem), and of course for speculation. Possibly others I'm not aware of too.
Coinbase is a bet on the idea that there's enough crypto-enthusiasm that they can capture a small part of a lot money while it lasts.