They go back about 11 years.
Depending how you count, parent comment is accurate. Hardware doesn't just appear. 4 years of planning and R&D for the first generation chip is probably right.
The first TPU (Seastar) was designed, tested, and deployed in 15 months: https://arxiv.org/pdf/1704.04760
They started becoming available internally in mid 2015.
I was wrong, ironically because Google's AI overview says it's 15 years if you search. The article it's quoting from appears to be counting the creation of TensorFlow as an "origin".
That's awesome. :) and even that article is off. They probably were thinking of DistBelief, the predecessor to TF.