As one technology giant after another significantly increases its investment in cloud data centers and the hardware it holds, it pays to ask how much of this spending growth is not just a product of companies for the long term but also because they have to spend a lot more on investment because of recent trends.
It's also worth asking how much of this spending will give the technology giants a long-term competitive advantage over smaller competitors who can not afford it.
Alphabet / Google (GOOGL) provided the latest evidence that cloud capex is experiencing a boom when it announced on Monday that its Q1 investment was $ 7.3 billion. Although about half of this sum covers asset purchases and not cloud infrastructure investment, data center spend increased sharply compared to the first quarter of 201
Facebook (FB) has now agreed that its investment in 2018 will more than double to $ 14-15 billion. And Amazon.com (AMZN) comes from a year in 2017 when asset purchases related to capital leases (mostly in terms of data centers) increased 69% to $ 9.6 billion. Amazon also increased its direct investment, which includes inventory investment, by 53% to $ 12 billion.
Earlier this year, the Bank of America estimated that between them Google, Facebook, Amazon, Microsoft (MSFT), Alibaba (BABA), Baidu (BIDU) and Tencent would increase their investment in 2018 by 30% to 74 , $ 1 billion increase. It should be noted that Google spent much more on investment in the first quarter than its expected $ 3.4 billion.
When Chief Financial Officer Ruth Porat was asked about Google's payout on his Q1 cloud investments, he downplayed the notion that spending was a one-time event. "[O] n machines, the largest contribution [to spending growth] is the demand that we see," she emphasized. "[I] In particular, it's the growing use of machine learning in alphabet, plus the requirements for cloud and search and YouTube, and secondarily the increased cost of newer technologies – CPUs, memory, networking."
Porat's comments on machinery investment deserve special attention. There is clearly an arms race among technology giants to hoard AI research talents that are limited in availability. And as Nvidia's (NVDA) recent results reports make clear, investments in hardware for the processing-intensive task of training deep learning algorithms have also developed.
Nvidia's flagship Tesla V100 server GPUs, widely used for training purposes, go for nearly $ 10,000 apiece. And the systems they contain often also have a sufficient amount of DRAM and flash memory. Although some of this expenditure is intended for long-term AI work, it also seems to go in the direction of existing services (language assistants, search, photo apps and services, content recommendations, etc.), with technology giants sensing the need   In addition, the expenditure on hardware used for inference – the task of running trained AI algorithms against real data and content – appears to be growing rapidly. Although inference is much less computationally intensive than training, processing requirements are typically significant when tens or hundreds of millions of consumers rely on an AI-powered service. Nvidia has argued that the inference market is as big an opportunity for him in the long run as the training market.
Video growth is another reason why cloud giants see accelerated capex growth. In early 2017, Google's YouTube announced that its video viewing has increased 10-fold to more than 1 billion hours per day since 2012 (probably even higher). Facebook has also seen tremendous video growth with the introduction of live streaming services on both Facebook and Instagram that address infrastructure needs.
As Porat indicated in her remarks, higher storage prices also provide upward pressure. DRAM prices have more than doubled since the fall of 2016, and while flash memory prices have recently started to decline, they have risen significantly over the past year. This is positive for a company like Micron (MU), but not so much for its cloud customers.
And in terms of capital intensity – the amount a company spends on investments relative to its revenue – it's worth it Much of the user growth that companies like Google and Facebook have seen for popular services comes from emerging markets that generate much less revenue per user than the US and Western Europe. While Facebook had an average revenue per user (ARPU) of $ 26.76 in North America and $ 8.86 in Western Europe in Q4, its ARPUs for the Asia-Pacific and rest of the world were only 2, 54 or 1.86 US dollars. The latter two accounted for more than 90% of their monthly active user (MAU) sales in the fourth quarter.
Nonetheless, the fact that cloud giants can afford to make such huge capital investments does not diminish smaller rivals. Just ask for Snap (SNAP), suggesting that it will not expand in emerging markets that are hard to monetize, as infrastructure costs are required to support all video uploading and sharing to additional users. Facebook's Instagram and WhatsApp, which now have over 300 million users per day for their Snapchat Storys clones, have no such dependencies.
Similarly, the huge sums that cloud giants spend on AI hardware and talent, along with the vast amounts of user data they have to power their training systems, make it harder for smaller competitors to compete. And the massive data center expansions that are being carried out by the public cloud service giants Amazon, Microsoft and Google are making it difficult for smaller competitors such as Oracle (ORCL) and CenturyLink (CTL) – competitors that are also generally disadvantaged Number of features they provide – to keep up.
Either way, the hardware and chip vendors who act as arms dealers for the cloud giants are the biggest winners in the short run. For companies like Intel (INTC) and Broadcom (AVGO), whose PC and / or smartphone chip sales have come under pressure, the cloud has become a safe haven.
Jim Cramer and the AAP team hold positions in Alphabet, Facebook, Amazon, Microsoft, Nvidia, and Broadcom for their Action Alerts PLUS Charitable Trust Portfolio . Want to be warned before Cramer buys or sells GOOGL, FB, AMZN, MSFT, NVDA and AVGO? Learn more now .