As it turns out, the future is not an extrapolation of the past. Since early 2012, the cost of sequencing has remained stable. This is more remarkable than it seems at first, because before 2012, the cost of sequencing had declined significantly in every year since 2001. But since the beginning of 2012, it hasn't budged and still remains at $6,000 per genome.
What went wrong? The biggest thing that happened was nothing. There is just not enough innovation in the sequencing market. Even if the two companies that dominate the market, Illumina and Life Technologies, came up with a vastly cheaper technology, it would probably not make commercial sense for them to make it available right now.
Eventually, there'll be some new technology that will trigger a new round of innovation, but right now it is not clear from where this will come. Nanopore sequencing still seems like a possible candidate, but that has been the case for a long time.