The age of artificial intelligence is upon us. AI and machine learning have the potential to generate $ 3.9 trillion in business value by 2022, according to research firm Gartner (NYSE: IT) . The McKinsey consulting firm says that 82% of the businesses it surveyed yield positive returns from AI investments, including better customer satisfaction and productivity, better recognition of financial firms fraud and a host of other benefits. In dollars, spending on AI and cognitive systems will more than triple between 2018 and 2022, from $ 24 billion to $ 77.6 billion, for a compound annual growth rate of 37.3%, according to estimates by research firm IDC.
There are many technology companies aiming to win bits of this pie. These include cloud computing companies that own the huge data centers where information is stored and processed, chip makers that produce CPUs, GPUs and FPGAs that can process the large amounts of data involved, software companies that organize this data and produce AI programs, and IT consultants who help large companies implement these systems.
AI's "pick and shovel"
However, the heart of AI rests on the basic components of memory and storage. According to hard disk manufacturer Seagate Technology (NASDAQ: STX) 23 exabytes of data were produced in 2002 (To say on a more familiar scale, there are 23 billion gigabytes .) By 2020, 23 exabytes of data will be produced every five hours . There is a huge amount of data that needs to be stored, either on hard drives or solid-state drives using NAND flash, and then fed to processors via Dynamic Random Access Memory (DRAM).
However, the memory and storage industry is notoriously cyclical, and it is currently at the bottom of that cycle, in part due to a withdrawal of spending caused by the US-China trade war. This may have opened up an opportunity for investors to buy stock and storage, including my favorite on the spot, Micron Technology (NASDAQ: MU) .
Micron tops my AI stock list because the product portfolio spans both NAND flash – the most modern storage technology – and DRAM, where one of only three global players remains. That DRAM oligopoly can increase or decrease production flexibly as market demand fluctuates, ensuring profitability throughout the product cycle.
And as Micron CEO Sanjay Mehrotra noted during an analyst day presentation last year, artificial intelligence servers require twice as much solid state memory and six times as much DRAM as standard servers. Since we are still in the early days of implementing AI, the demand for NAND and DRAM should hover in the long run, although the current period is challenging.
In addition, Micron and Intel (NASDAQ: INTC) are the only companies capable of producing a new type of memory called 3D Xpoint, which is non-volatile, like NAND flash , but much faster. The downside is that it is more expensive, but there are many AI applications that can take advantage of its unique features.
Last year, Micron took advantage of its option to purchase the 3D Xpoint factory from IM Flash Technologies (IMFT), a joint venture Micron and Intel have developed since 2006. The two recently decided to go their own ways with 3D Xpoint -technology; Micron took advantage of its option to take over the facility, so Intel would have to develop another facility. IMFT will provide Intel with 3D Xpoint discs for up to one year.
And while Intel released its first 3D Xpoint Optane products back in April, Micron only took full control of IMFT and announced its first 3D Xpoint products last week.
Recent Exciting Announcements
24. In October, Micron dropped a number of press releases, including one that addressed the introduction of its first 3D Xpoint SSD, the X100. It will have speeds three times faster than the latest NAND Flash-based SSDs with 11 times endurance, resulting in an end-user experience that is two to four times better than flash solutions.
In addition, Micron announced several product innovations and investments related to AI. These include Authenta, a hardware key for hardware-level IoT-enabled devices, as well as Micron's first portable consumer SSD, called Crucial. And the company buys FWDNXT, a startup that creates an integrated hardware and software solution for edge AI applications. Finally, it made a venture investment in German startup Volocopter, an autonomous air taxi company, which claims the vehicles are safer than helicopters.
Everything aboard the AI train
Despite all the things Micron has gone for, the stock is still trading at just 7.6 times adjusted (non-GAAP) earnings from 2019. That's not entirely unreasonable, as revenue declines due to this year's crash in memory prices. While Micron earned $ 6.35 per share in diluted EPS for the fiscal year ended in August, analysts on average expect only to earn $ 2.54 per share for the current fiscal year.
However, it should be at the bottom of this cycle, which means that Micron trades for less than 19 times through EPS, compared to the S&P 500, which trades with 22.5 times subsequent earnings. Meanwhile, the company earned $ 11.95 per share as of fiscal 2018, which means it only trades at four times the top earnings in the last "up" cycle.
While Micron has historically lost money during down cycles, it doesn't seem like it will happen this time. The memory industry is far more consolidated, and Micron has significantly reduced its cost structure and improved its balance sheet, by $ 3.4 billion in net cash at the end of the last quarter.
Despite the recent collapse in memory prices, Micron's stock has risen by more than 50% so far in 2019. Still, it comes after a 2018 where the stock fell sharply, and the stock is still well below previous all-time highs in the middle of the $ 60's.
With a fortress balance, an unrivaled product portfolio and a valuation far below other leading technology companies, this company is at the heart of the AI revolution a great opportunity for long-term investors.