Things aren't as bad for Intel (INTC) as some of its harshest critics would make them out to be. But the number of headwinds faced by the chip giant is steadily growing.
First, the positives. Though some of their end-markets are stronger than others, Intel's bread-and-butter PC and server CPU businesses appear to be doing well right now on the whole.
While desktop demand is soft, notebook demand continues benefiting from purchases made to support remote workers and students. And server CPU sales to Internet/cloud giants (the proverbial hyperscalers), which were already on the upswing before COVID-19 lockdowns arrived, have gotten a fresh boost from the traffic spikes that have followed the lockdowns.
Intel is also seeing -- with the help of favorable technology and capex trends -- strong demand from telecom service providers, both for server CPUs and its Atom P5900 5G base station processor. And while turning a profit on the business remains a challenge, Intel's non-volatile memory unit is seeing healthy growth as well, with the help of higher NAND flash prices and growing demand for Intel's Optane next-gen memory.
Finally, though AMD (AMD) has been gaining PC and server CPU share (more on that shortly), Intel has done a better job of holding its own in some end-markets, such as desktops and servers purchased by traditional, Global 2000-type enterprises, than what many had feared.
But with all of that said, some of the end-markets where Intel has seen strong early-year demand are likely to soften during the back half of the year.
Intel and AMD both cautioned on their Q1 earnings calls that their PC-related sales might soften during the second half of 2020 thanks to macro pressures. And recently, there have been supply chain reports pointing to second-half headwinds.
Moreover, in Intel's case, PC CPU sales will also be pressured a bit by Apple's (AAPL) plans to start selling its first Macs to be powered by its own processors before year's end. The initial impact probably won't be massive -- Apple says it will take two years to fully transition to its own Mac processors, and still has some Intel-powered Macs in its pipeline -- but with Apple claiming about a 7% PC unit share and a 15%-plus revenue share, there will be a meaningful impact over time.
Intel also cautioned on its Q1 call that it thinks its server CPU division (the Data Center Group, or DCG) will see demand from traditional enterprise and government end-markets soften during the second half of the year. Given that DCG's sales to these end-markets were up 34% annually in Q1, and that the likes of Dell (DELL) , Hewlett-Packard Enterprise (HPE) and NetApp (NTAP) have been reporting declining server and/or storage sales, it looks like Intel got a Q1 boost from inventory-building by enterprise hardware OEMs.
Separately, Intel warned that its Internet of Things Group (IOTG) and Mobileye ADAS vision processor unit would see lower demand in Q2. IOTG's sales are being hurt by COVID-19's impact on demand from end-markets such as retail stores and factories; Mobileye is naturally being stung by lower auto sales and production.
Then there's the issue of AMD, which has been steadily gaining ground in the consumer PC, cloud server and high-performance computing (HPC) server markets.
In the notebook CPU market, Intel (in spite of Apple's plans and tougher competition from AMD) is still in reasonably good competitive shape, and will benefit in the coming months from the rollout of its Tiger Lake notebook processor line, the second to use its 10-nanometer (10nm) manufacturing process node. Among other things, Tiger Lake will deliver major integrated GPU performance gains.
But in the consumer desktop, high-end desktop (HEDT) and cloud and HPC server markets, where specs and benchmarks can often have a big impact on buying decisions, Intel is in a tougher spot, since it's still relying on desktop and server CPUs made using a 14nm manufacturing process node that (though improved over the years) is showing its age.
Last year, AMD launched its first desktop and server CPUs to be made using Taiwan Semiconductor's (TSM) 7nm manufacturing process node, which is seen as competitive with Intel's 10nm node. And before 2020 is over, AMD is expected to roll out new 7nm desktop and server CPUs that rely on a next-gen CPU core microarchitecture known as Zen 3.
Intel has said it will ship its first 10nm server CPUs before year's end, but hasn't given much clarity about what 2020 volumes will look like. And it still hasn't said when it will start shipping 10nm desktop and HEDT CPUs.
Also, though this may be more of a long-term challenge given corporate inertia and lingering software compatibility issues, ARM-architecture server CPUs from the likes of Amazon Web Services (AWS), Marvell Technology (MRVL) and private Ampere Computing are likely to gain some share over time. AWS' recently-launched Graviton2 ARM CPU, which is available via aggressively-priced AWS cloud computing instances, has fared well against Intel and AMD server CPUs in some (though not all) popular benchmarks.
Intel's stock is admittedly cheaper than that of many chip peers. But with Intel's free cash flow (FCF) now meaningfully below its earnings thanks to elevated capital spending, it's not quite as cheap as it might look at first glance. Intel's $280 billion enterprise value is equal to 17 times reported 2019 FCF of $16.9 billion, and 18 times both its 2020 and 2021 FCF consensus estimates, which each stand at $15.8 billion.
And between the macro and company-specific challenges Intel is facing right now, as well as its need to keep investing heavily in capex, there is arguably some downside risk to those FCF estimates.
