Investors looking for bargains among chip developers are more likely to find them among one type of chip developer than another.
Specifically, chip suppliers which partly or wholly rely on their own chip manufacturing plants (fabs) are on average trading at lower earnings multiples than "fabless" suppliers that outsource chip manufacturing to third parties.
There are admittedly some exceptions to this rule. For example, Texas Instruments (TXN) and Analog Devices (ADI) , each of which rely heavily (though not entirely) on their own fabs, are both valued at over 20 times their consensus analyst EPS estimates for their 2020 fiscal years.
But on the whole, it's fair to say that discount valuations are a lot more common among fab-owning chip suppliers than fabless ones. Here's a list of some notable fab-owning firms trading at low multiples.
- Intel (INTC) trades for 11 times its expected calendar 2020 EPS.
- NXP Semiconductors (NXPI) trades for 12 times its expected calendar 2020 EPS.
- ON Semiconductor (ON) trades for 11 times its expected calendar 2020 EPS.
- Microchip Technology (MCHP) trades for 13 times its expected fiscal 2021 (ends in March 2021) EPS.
- Skyworks (SWKS) trades for 13 times its expected fiscal 2020 (ends in Sep. 2020) EPS.
- Qorvo (QRVO) trades for 12 times its expected fiscal 2021 (ends in March 2021) EPS.
- With the caveat that estimating what a memory maker will earn in a year or two is very much a guessing game, Micron (MU) trades for 10 times its expected fiscal 2021 (ends in Aug. 2021) EPS. It also trades for just 4 times what it reported for fiscal 2018.
Meanwhile, quite a few fabless firms trade at markedly higher valuations. Some examples:
- Xilinx (XLNX) trades for 25 times its expected fiscal 2021 (ends in March 2021) EPS.
- Silicon Labs (SLAB) trades for 31 times its expected calendar 2020 EPS.
- AMD (AMD) trades for 30 times its expected calendar 2020 EPS.
- Nvidia (NVDA) trades for 25 times its expected fiscal 2021 (ends in March 2021) EPS.
- Semtech (SMTC) trades for 24 times its expected fiscal 2021 (ends in Jan. 2021) EPS.
- Following a recent post-earnings surge, Ambarella (AMBA) trades for 98 times its expected fiscal 2021 (ends in Jan. 2021) EPS, and 60 times its expected fiscal 2022 EPS.
Differences in growth expectations can partly explain the valuation gap that has emerged for the two sets of companies. Xilinx, for example, is expected to see double-digit sales growth with the help of 5G network rollouts and growing FPGA usage within data centers. AMD's sales are expected to get a healthy boost from CPU share gains and a new game console cycle, and Nvidia's are expected to get a lift from growing demand within data center and automotive end-markets.
On the flip side, Intel is three months removed from forecasting it expects just low-single digit annual revenue and EPS growth through 2021. And for now, weak smartphone demand is a clear headwind for companies such as Skyworks and Qorvo.
But many of the fab-owning companies still have growth drivers that -- together with operating leverage and possibly margin expansion -- could fuel double-digit EPS growth in the coming years, provided macro conditions aren't terrible. For example, suppliers of analog/mixed-signal chips and microcontrollers (MCU), such as NXP, Microchip and ON Semi, are set to benefit from IoT device growth and rising chip content within cars, with some (such as NXP and ON) also well-exposed to 5G infrastructure ramps. And even if smartphone sales don't grow much in the coming years, Skyworks and Qorvo could see sales growth thanks to the RF chip demands placed by 5G phones, together with their exposure to automotive, IoT and 5G infrastructure markets.
Given that it's far from impossible to see many of the fab-owning firms also posting double-digit EPS growth, a factor other than growth expectations arguably has much to do with the valuation gap that has emerged. Namely, investors worried about a major cyclical downturn are assigning a valuation discount to fab-owning firms, owing to worries that the fixed costs attached to their manufacturing footprints will result in greater earnings pressure than fabless firms will see if macro conditions get markedly worse.
To some extent, that's a fair concern. But while one can justify placing a bit of a valuation discount on the shares of fab-owning chip suppliers, placing forward EPS multiples in the 10-to-13 range on companies that have meaningful growth opportunities feels questionable -- particularly given that EPS estimates for these companies have already been cut (often sharply) amid the current downturn, and that many of them spend much less than 10% of their annual revenue on capital expenses for their manufacturing footprints.
As a result, while one can make bull cases for some fabless chip suppliers as well on account of their growth potential, more value-oriented investors are likely to find more compelling opportunities among some of the fab-owning firms.