"DoD saying no conflict of interest in a contract obviously intended as a cloud contract seems specious... I mean really, did DoD expect that AWS would not be in the running while Ubhi was scratching out pieces of the early contract terms?"
But the DoD did identify conflicts of interest (the people involved raised the conflict) and they were separated from the decision process.
Oracles counter-argument is that key decisions (using one provider) were already made at the point the people stood down from the process.
On the larger point around cloud computing, the DoD currently has 50+ data centre operators for their systems and will pay almost US$17.4B in 2019. JEDI is estimated at being worth around US$10B for a 10-year contract to put that into perspective with overall DoD IT spend. Current DoD cloud spending was estimated at US1.15B/year in 2018, so the use of cloud is not new, and the consolidation of services into one provider (with multiple facilities) is expected to save significant amounts.
AWS, Azure and Google have existing, large facilities dedicated to government services in the right locations for the DoD to use them for this contract. IBM and Oracle don't have suitable facilities at present - they expect the government to pay to build those facilities delaying implementation and increasing costs putting them at around a $100M-$200M cost disadvantage and 1-2 year disadvantage for bringing services (and savings) into service. IBM's cloud business will survive missing out on this contract, I suspect Oracles won't given it's chequered past.