Amazon (AMZN 1.19%) might not be the first company that comes to mind when you think of artificial intelligence (AI). But it may need to be. Amazon is one of the most important companies in the AI ecosystem, stemming from Amazon Web Services (AWS) being a top provider of AI computing hardware.

I’ve got three reasons why it should be at the top of investors’ radar, and understanding these points may help you to position your portfolio for maximum returns.

1. Amazon’s custom chip business is thriving

Since the beginning of the AI build-out, graphics processing units (GPUs) have been the primary computing chip of choice for AI companies. GPUs are great for a wide variety of workloads, but custom AI chips can be more cost-effective if the chip is configured for that specific workload. As a result, more companies are shifting toward custom AI chips, and Amazon has felt this change.

Amazon

Today’s Change

(-1.19%) $-3.19

Current Price

$264.03

Amazon designed its own custom AI chips — a process it is no stranger to doing. In 2018, nearly all data center workloads were run with Intel central processing units (CPUs). Then, Amazon created its Graviton chip that widely replaced Intel’s products. Now, 98% of workloads use Graviton chips versus Intel’s CPUs.

Amazon believes the same thing will happen in the AI training space, with GPUs being widely replaced by custom AI chips, like its Trainium series. Amazon is already seeing huge demand for its custom chips business, with its revenue growing at a triple-digit pace and nearly all available training capacity for these chips being sold out.

If Amazon can capture a large chunk of the AI computing market with its own chips, that will be a major boost to margins, causing Amazon’s stock price to soar.

2. Amazon is spending big to meet demand

This year, the big four AI hyperscalers plan to spend $650 billion on capital expenditures to meet rising AI demand. Amazon is the biggest spender in this group, with plans to spend around $200 billion this year. Amazon isn’t doing this on a whim; it already has commitments from major clients to use this computing capacity. And the faster the business grows, the more companies will have to spend to keep up with demand. That’s the way cloud computing works.

Because demand is continually increasing thanks to AI, Amazon will likely have to spend big for many years, which may seem like a negative. However, once that initial investment is made and the data center is operational, the cash flows from these investments will far outpace the short-term spending pains. This makes Amazon a great long-term investment in the AI space, as it will be able to turn these major investments into a recurring revenue model that will produce for years to come.

3. Amazon doesn’t carry a massive premium

Valuing companies that are spending big on capital expenditures isn’t easy, as there are several ways it can skew traditional earnings metrics. As a result, I think valuing a stock like Amazon on operating cash flow is a smart move.

AMZN Price to CFO Per Share (TTM) data by YCharts

Amazon’s valuation has come down significantly in the past few years, but even now it’s still not that expensive compared to some of its peers. Alphabet and Apple trade for about 28 and 32 times operating cash flow, respectively.

This erases the notion that Amazon is an expensive stock; its valuation metrics are just skewed due to huge reinvestment that should pay off big-time over the long term. As a result, I think Amazon is a top AI stock to buy now, as the shift to custom AI chips could be a major driver for its stock.