75% Of US GDP Growth In The First Quarter Was Due To AI

On the surface, today’s final revision (aka 3rd estimate) of the US Q1 GDP print was unremarkable: Real GDP grew 2.1% annualized in the first quarter, a reversal of last month’s downward revision of 1.6%, but back to where the original print was when it was reported in April, when the BEA reported 2.0% growth. 

The print reflected a downward revision to imports, which are a subtraction in the calculation of GDP, that was partly offset by a sharp downward revision to consumer spending.

Taking a closer look at the components, net exports contribution being revised sharply higher to -0.4% from -1.3% previously drove the improvement while consumption was much weaker. Real personal consumption expenditures revised sharply lower to 0.5% (saar) from 1.4% (saar). This is unexpected as virtually everyone was convinced that bumper tax rebates from Trump’s OBBBA “stimulus” would push Q1 personal spending; in retrospect, spending in Q1 was far weaker than expected. 

That said, real spending in May climbed 0.3% (3.2% annualized), while April was revised to 0% from 0.1%. This suggests an okay pace of spending but not boomy across the two months (1.6% annualized) considering bumper tax refunds putting extra money in people’s pockets. 

Yet, as before, when we get to fixed investment, something remarkable emerges: Residential housing investment declined 1.7% and subtracted 0.3% from the bottom line GDP print. This was the 5th consecutive decline as residential investment has declined, and 7th of the past 8 quarters. To be expected at a time of rising interest rates. 

But Nonresidential fixed investment was the outlier, soaring by 8%, and responsible for 1.42% of the 2.1% bottom line print.

Let’s take a closer look at the breakdown.

The chart below shows quarterly annualized GDP growth broken down by components. It shows that Q1 GDP grew at exactly 2.100% in Q1. Also notable is that traditionally strong consumption, added just 0.37% of the bottom line number, as per the discussion above; this was offset by net trade being a far smaller detractor from GDP growth at -0.37% with, inventories (0.23%) and government (0.74%) providing a modest offset. 

The highlighted block is Fixed Investment, which contributed 1.11%. However, keep in mind that residential fixed investment subtracted 0.30% from the total number, which means that Nonresidential fixed investment was responsible for 1.42% of the 2.1% GDP print.

Focusing on the fixed investment component, we find the following: as noted above, it was all about non-residential fixed investment.

Zooming into this segment, we find that Nonresidential equipment grew by 5.8%, or contributing 0.8% to the 2.1% GDP, while Intellectual Property products grew just over 5.3%, and added 0.74% to the bottom line GDP. 

While IP is clear – it consists primarily of Software, the kind that one uses to create and develop AI tools, as well as R&D – the components behind Nonresidential equipment need a closer look again, and here we find that Information Processing equipment, i.e., data centers, grew at a stunning 14%, comprising virtually all of the 0.81% contribution to 2.1% GDP growth.

And there you have it: between Software (0.74% of the GDP growth) and Nonresidential Equipment (0.81%), AI – which was the primary driver behind growth in both – contributed just over 1.5% to GDP growth of 2.1%; in other words about 74% of all US growth in Q1 was due to AI.

Another way to visualize the remarkable impact of spending on “computers” is the chart below: it clearly shows just how reliant the US has become on spending on computer products.

And that’s why AI is now not only a market bubble, but it has become a core anchor propping up the entire US economy; it’s also why the US government will have no choice but to backstop it once the inevitable AI bubble pops.