Doubt if an 8008 uC would have been a commercial success. It ran at 125 KHz with instructions taking around 20 clocks. Instruction set was very primitive compared to the 8080 and the address bus was only 14 bits wide. It was best suited for calculators and controllers.

I built an 8008 machine in 1974 with a whopping 256 bytes of memory and toggle switches for program entry in binary machine language. You could watch it execute subroutines in the address bus LEDs!

That's around five times slower than a PDP-8/S, which was possibly the slowest mini around at the time - already obsolete by 1970.

So it's doubtful it would have been successful as a computing product, although it might have had some applications at the more undemanding end of process control.

It would have been more interesting to use the skills, techniques, and marketing contacts collected during the 8008 project to jump-start an 8080 project as soon as the new CPU appeared.

But the S100 era happened because the Altair had a standard bus that made the system trivially expandable.

Corporate/industrial/academic systems had expensive and relatively complicated proprietary bus and backplane systems, which limited the market and kept prices high.

So... an alternative 8080 system wouldn't have been a game changer unless it was sold with the same open hardware model. At best it would have been an expensive micromini for a niche academic and industrial market.

The 8 level deep stack of 8008 precluded most complex software though. It was also 4 to 6 times slower than an 8080 running at 2MHz, had a very small package that required extensive demultiplexing, slowing it further.

There were multiple kit computer based on 8008 in the era, ex. Mark 8, the French Micral, but they were severely hampered compared to something like the Altair 8800 and it's unlikely they could have had a similar story despite launching earlier.