What this guy said.
"Why is a high speed data bus so complex?". They evolve to be faster and more efficient each year. What's high speed today isn't in two years from now. So do you want to fast in two years, built on today's tech, so it's underutilised battery life is terrible? Or do you want to limit the efficacy of the new plug-in modules in future by having the bus under spec in two years? Or do you want to build it on something that doesn't exist yet, sooner than two years from now, so it costs a fortune to design and manufacture?
Something as simple as a "high bandwidth data bus" has capacities and costs associated with it. I don't know why it took so long for the Ara team to find this out.