Faster is more expensive. USB 1.0 was aimed to replace PS2 and ADB busses for keyboard, mice, joysticks... A slow, low power 8bits microcontroller with a few tens of bytes of RAM could handle that. USB 1 also can use cheap unshielded cables.

Last month I bought a optical mouse for $1, shipping included (I live in China). The thing works just fine but the cable isn't even twisted, let alone shielded. On one pair of untwisted and shielded wires, 12Mbps is extremely fast IMO.

How many mice, keyboards, and printers have you seen with Firewire or Fast Ethernet ports?

@JörgWMittag - Printers with (at least) 100 Mb/s Ethernet these days? Seems like most of them....

It’s probably hard to believe now, but 100Mbps Ethernet seemed insanely fast at the time. It was so fast compared to the rest of most systems (including workstations) that reliably capturing a 100Mbps network stream (what you’d do with tcpdump nowadays) was rather difficult.

@JörgWMittag The point is that it would be possible to build ethernet connected keyboards.

@zomega: Sure, but the Ethernet chipset alone would have cost $50 in 1996. How many people are going to accept that a device that used to cost < $10 now costs > $50, just because of a new connector that doesn't add any new functionality?

@JörgWMittag That is a point I did not know. 1996 was before my computer life.

@zomega This principle holds in general too: Supersonic airplanes exist but most airplanes are not supersonic just because it is expensive and not needed, Super cars exist most cars are not them. Fiber optic cables exist but most connections are still copper. Top-end Intel and AMD CPUs exist but 8-bit microcontrollerss are still used and outnumber them. Don't expect everything that exists at the same time to be the most that could be achieved unless it is actually required or free to do so. One example where it is almost free: simple microcontrollers have displaced moderately discrete logic.

Transferring the contents of an entire floppy in a single second is pretty GD fast.

@DKNguyen I understand this principle but USB 1 is like a horse carriage while super cars exist.

"USB 1 is like a horse carriage" Hardly, though. For example, UK copper broadband is only just being replaced by fibre so merely 3 years ago, most UK homes have <35Mbps download and can work from home on it, surf, shop, stream video, make calls, download films etc. It's lots. 12Mbps is not 1 bps :-) A responsible business case must be made for the technology, the business and application come first every time. If USB launched as 100Mbps in 1990s, you'd never have heard of USB as it'd have quickly died out for being spectacularly expensive. In mass tech, the important numbers are £s not Mbps.

@TonyM Firewire provides a useful analogy there to support your point ;-).

@StephenKitt, spot on, Firewire's a good one - my example would've been MCA but I ran out of characters :-) MCA technically outstripped ISA, a good few years before PCI did it for good and did it better, but the masses didn't require its expensive speed benefits so it got passed by and died out. Today, 3D-capable TV is an example. Technology's there for mass 2D and 3D viewing but it's seen, by voting with their wallets, that people don't see a worthwhile return on the extra expense and inconvenience. It died out and we're back to 2D.

USB was wildly faster than everything it was meant to replace. the serial standards it was made to replace were measured in kilobytes per second. At the time, it was a huge improvement.