The problem as I see it is you are targeting a very niche market , i.e one for which an ODA + sunfire server is too small and and eigth rack of exadata and exalogic is too large. this is probably not large enough a market to be sustainable and it would eat into the standard exadata/exalogic use cases. So I don't see this happening in the near future. I bet Oracle marketing would reason that if some one is willing to buy a rack comprising of an eight exadata and eight exalogic in a single rack , they probably wouldn't mind buying two ( since there is not likely to be any price difference) and this would give customers room to expand if they wanted to consolidate.
Actually rather than a niche market, it's the case of a startup, where a relatively small volumes are currently present, and plan to spend 18 month development thean roll-out products where substantial more volumes are expected.
A 1/8 Exadata rack is affordable, even a 1/4 if this is flexible enough to allow all tiers. The nature of the business dictates for a DR site, so if you consider a 1 Exalogic + 1 Exadata for production, than 1 + 1 replicated for the DR site would be cost prohibitive now, but acceptable in three years plan.
Hence why a general purpose engineered system similar to the T4-4 for the Exastack Intel range, would be ideal,, otherwise perhaps a T4-2?
Essentially the first step entry level is still too high, unnecessarily because all the components are available literally off the shelf.