Teradata might be the pioneer of data warehousing on cheap x64 server clusters and the use of appliance packaging to tune machines and their software to attack specific workloads, but Oracle and IBM want to eat Teradata's lunch. And its breakfast and dinner, too. That means Teradata has to keep upgrading its hardware and …
Strengths and weaknesses
Teradata's laser-like focus on data warehousing is both their biggest strength and biggest weakness. For example, the proprietary way in which they do storage connectivity (with disks essentially being direct-attached to their parent node) makes them lightning fast, but at the same time, the company's failure to understand common IT industry standards when it comes to stuff that happens OUTSIDE their magic box is their undoing. Why is the system cabinet a couple of inches deeper than the other cabinets, thus monopolising a floor tile that should be in the centre of the aisle? Why do they insist on appending hostnames with those dodgy cop1, cop2, strings that completely stuff up coporate DNS addressing schemes? They either need to partner with, or create a subdivision of, non-warehouse-geeks who can integrate them into the wider world of IT with less pain. That way, Oracle, IBM etc will be even less likely to catch them.
Pimp my ride
Teradata's MPP architecture (on all but the single node SMP machines) is predicated on scaling out, not scaling up by building fat SMP nodes. You want to go faster? Deploy more nodes! (economics and data centre space notwithstanding).
As a result there is not the same pressure to always offer the absolute latest and greatest CPUs, especially when the processing capability has to be balanced with the IO sub-system capability. Not that more oomph is not welcome. The relatively recent move to x64 and SLES (from NCR's 32bit MP-RAS) has been an enabler on the performance front.
SSD disks may just have met their match with the 'sub-optimal' queries Teradata end users are fond of running ;-)