- AMZN
- NVDA
Amazon‘s (NASDAQ: AMZN) second-quarter earnings had investors on edge, fixated above all else on the company’s capital expenditure (capex) outlook. This figure came in at roughly $220 billion for the full year, a meaningful increase from the previously stated $200 billion. According to management, higher memory costs are driving the surge.
The question hanging over Amazon’s financials is whether such heavy spending can still be justified. During the earnings call, Amazon CEO Andy Jassy made some comments that offer a pointed answer.
Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a “Double Down” signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same “Total Conviction” signal is flashing for a company 1/100th the size of Nvidia. Continue »
How are capex and free cash flow related?
Heavy capital spending and free cash flow are linked by a simple accounting equation. Free cash flow equals cash generated from operations minus capital expenditures. When Amazon accelerates investment in data center infrastructure and servers, capex rises and free cash flow compresses.
In the trailing 12 months that ended with the second quarter, Amazon’s free cash flow swung to an outflow of $7.6 billion. The swing was driven by a $66.1 billion year-over-year increase in property and equipment purchases, the bulk of which was tied to artificial intelligence infrastructure.
The company’s core profitability engine, Amazon Web Services (AWS), saw revenue reach $42.2 billion in the quarter, up 37% year over year. Meanwhile, operating income from AWS jumped 64% to $16.6 billion. The contrast here is hard to overlook.
Even though Amazon’s largestuild-out of AI capacity is so large that free cash flow is turning negative. This inverse relationship is not a sign of operational weakness; rather, it is the arithmetic reality of front-loading a multi-year investment whose returns are expected to arrive only after new facilities and servers are brought online and filled with new customer workloads
Jassy explains the economics of AI data centers
During the earnings call, Jassy spent considerable time explaining why the current surge in infrastructure spending should ultimately pay off. He noted that data centers have useful lives of 30 years or more. Inside each facility, Amazon can cycle through five or six generations of servers. After the first generation, the unit economics improve because the initial capital outlay does not have to be repeated.
