Citi Research analyst Mark May recently told investors that Amazon (AMZN 1.21%) could "minimize or avoid the risk of increased regulatory pressure" by splitting into two companies: one for its marketplace business and the other for Amazon Web Services (AWS).
May believes that splitting the two companies would also "better align" the two units' stock-based compensation expenses, help Bezos retain top execs and forge a succession plan, avoid conflicts of interest, improve shareholder selection, and provide "a more attractive M&A currency for potential acquisitions." May also believes that the split could help both companies achieve better valuations as "pure plays" in their respective markets, and be well timed to coincide with the construction of HQ2, Amazon's new corporate headquarters.
However, May's argument is one that we've heard from analysts ever since Amazon started disclosing its AWS numbers separately in 2015. The only significant change this time is an increased threat of regulatory pressure. Back in 2015, I argued that "it makes much more sense to keep AWS than to spin it off." That argument still holds true for four simple reasons:
1. It's Amazon's biggest profit driver
Amazon generates most of its revenue from the marketplace business, but most of its operating profits come from AWS.
During the first six months of 2018, AWS revenue rose 49% annually to $11.5 billion, and accounted for 11% of its top line. Its operating profits surged 68% to $3.04 billion, and accounted for 62% of Amazon's operating income.
If we remove AWS revenue and operating income from Amazon's total numbers, its operating margin for the first half of the year would drop from 4.7% to 2%. Therefore, AWS might look better as a stand-alone investment, but Amazon wouldn't.
2. AWS profits support Amazon Prime
Since Amazon's investors would receive shares of both companies in a split, the stronger growth of AWS' stock likely would offset Amazon's slower growth. That's what happened after eBay spun off PayPal in 2015.
However, AWS' higher margins enable Amazon to expand its marketplace and Prime ecosystems with lower-margin (or loss-leading) products like Kindles, Echo speakers, and Fire TVs, as well as costly delivery efforts like AmazonFresh.
The ongoing expansion of that ecosystem helped Amazon surpass 100 million Prime members earlier this year. If Amazon spins off AWS, it could need to rein in that spending -- which could narrow its moat against retail rivals like Walmart.
3. There are cost-cutting synergies
Amazon runs its own streaming services (Amazon Video and Amazon Music) and its cloud storage services on AWS -- which significantly reduces its hosting expenses. Amazon also owns a video game studio, Amazon Game Studios, which runs its games on AWS. And it owns a growing advertising business, which relies on AWS for analytics and content delivery.
Most other companies use third-party cloud services. Netflix hosts its content on AWS, and Spotify uses Alphabet's Google Cloud. Epic Games runs its hit game Fortnite on AWS.
Here's the problem: Amazon's video, music, and gaming units are part of its core marketplace division. If AWS was split off as a stand-alone company, Amazon would need to pay AWS cloud hosting expenses, which would significantly throttle its profit growth.
4. There's plenty of room for growth (as part of Amazon)
Lastly, AWS still has plenty of room to grow as part of Amazon. Earlier this year, Citi's May estimated that AWS' annual revenue could hit $44 billion by 2020, compared with $17.5 billion last year.
It's tempting to think that spinning off that high-growth unit would attract more investors, but AWS is better suited for supporting the growth of Amazon's entire ecosystem (marketplace, online services, advertising, and more) instead of as a stand-alone cloud platform company.
A reason to hold Amazon
I strongly believe that AWS should remain a part of Amazon. It's one of the main reasons I bought the stock in the $600s, and why I continue to hold the stock after it has tripled. If Amazon spins off AWS (which I doubt it will ever do), I'd need to reevaluate my original investment thesis.