That's not all.
As well as $8b for specific projects, Robertson has added $4b to the pool of uncommitted dosh for capital projects.
Expect some of that to be allocated in Budget 2020 or in the 2020 election campaign in the way of promises.
Robertson denied that delaying the big reveal until 2020 is anything to do with election year - he would say that, wouldn't he?
But when it comes, it will feature what until now has been something of a dirty word for this government - "roads".
The lion's share of the infrastructure spend, $6.8b, will be on transport projects - with a significant portion for roads and rail.
That is unlikely to see a re-emergence to National's raft of Roads of National Significance because the package has to be negotiated with the Greens and New Zealand First.
But it may signal an end to the Government's aversion to large roading projects.
The extra borrowing for infrastructure spending pushes the Government's net debt target beyond the 20 per cent of GDP as originally promised to just over 21 per cent.
But effectively the 15 per cent to 25 per cent band that was to have replaced the 20 per cent target in the 2022 - 23 financial year has been brought forward.
And as Robertson himself notes, no economic commentator is likely to quibble with going to 21 per cent especially when the very same commentators have been calling for the Government to take borrow cheaply for infrastructure spending.
Robertson will be disappointed that Treasury's forecasts have revised the $1.3b surplus for the current year to a $943m deficit mainly due to lower growth and revised ACC liabilities.
But it a single year of deficit, it may not actually eventuate, and sits in a sea of surpluses.
There is no reason for it to damage Robertson's reputation for fiscal responsibility.
In the process, he has set up the Government nicely for election year.