Tech Stock Thoughts #7: Microsoft, AMD, GameStop, Chip Sales

Microsoft’s Earnings: Enterprise and Gaming Both Stand Out

Revenue growth accelerated across a number of businesses.

Sources: Microsoft (earnings report, earnings slides), Unhedged (CC transcript)

• Microsoft’s constant-currency revenue growth accelerated to 15% from the September quarter’s 12% in spite of some tough comps for businesses such as Office consumer, Windows and Surface.Office commercial, Azure, Dynamics, LinkedIn and of course gaming all saw accelerating growth. And though the company has historically guided conservatively, its March quarter sales guidance was fully above consensus, with the midpoint implying 16.5% growth.
• The enterprise/commercial strength might be the thing that pops out the most here — both in terms of its breadth and Microsoft’s commentary about how various pressured end-markets and revenue streams (i.e., SMBs, industries hard-hit by COVID, transactional software license sales) were recovering.
• Some enterprise numbers of note: Azure revenue rose 50% Y/Y, Dynamics revenue rose 21%, security revenue was up more than 40% on a trailing 12-month basis, LinkedIn ad revenue rose 53% (another sign of online ad strength), and commercial RPO (i.e., the commercial contract backlog) was up 24%. Also, unlike so many peers, Microsoft’s on-prem server software revenue is still growing, and doing so in spite of difficult comps.

Whether in constant currency or dollars, a number of Microsoft’s business growth rates look impressive.

• It’s not a shock that gaming revenue grew strongly, but it’s worth noting that said revenue topped $5B for the first time in Microsoft’s history, and that growth of 51% blew away guidance for high-20% growth. And it wasn’t just Xbox hardware sales (up 86%) that were responsible: Xbox content/services revenue rose 40%, more than $2B in revenue was produced by third-party titles, and Game Pass subs rose by 3M Q/Q to 18M.
• If one wants to nitpick, one can argue Microsoft’s Windows OEM and Surface revenue — up 1% and 3% Y/Y, respectively — could’ve been better even with growth rates pressured by tough comps and (in the case of Windows OEM) a mix shift away from high-ASP OEM Pro licenses, given how strong notebook demand is right now. Chromebook and Mac demand were likely headwinds here.
• But otherwise, this was a very strong report. And for a company that not long ago was written off as a tech dinosaur, it’s kind of remarkable how such a report surprises very few people these days.

AMD’s Earnings: No Seasonal Weakness to See Here

CPU share gains continue, and end-market demand is strong. It’s just a question of what multiples to pay.

Sources: AMD (earnings release, earnings slides), Unhedged (CC transcript)

• AMD reported 53% Y/Y sales growth for Q4, issued Q1 sales guidance that implies 79% growth at its midpoint and forecast ~37% 2021 growth. With the Q1 guidance midpoint just $44M below AMD’s Q4 revenue, there’s a decent chance the company sees Q/Q growth in what’s normally its seasonally weakest quarter.
• These aren’t normal times, of course, and AMD is benefiting from unprecedented gaming hardware demand as well as a giant uptick in consumer notebook demand. Also, thanks to ongoing 7nm wafer supply constraints, there was a fair amount of unmet demand going into the new year. However, a quick comparison of AMD’s Q4 numbers with Intel’s makes it clear that AMD is also benefiting from meaningful desktop and server CPU share gains.
• Moreover, with AMD indicating that its 2021 server CPU sales growth will outpace expected company-wide growth, the company seems pretty confident that Epyc will gain a fair amount of additional share this year as Milan (due to formally launch in March) ramps. Judging by AMD’s commentary, internal-workload share gains at hyperscalers — a customer base that tends to care a lot about things like core counts, TCO and power efficiency — are a major Epyc growth driver right now.

AMD’s server CPU and game console processor sales were both up sharply in Q4.

• Possibly weighing on the stock post-earnings: While full-year guidance was more than $1B above consensus, markets might’ve wanted something even stronger given AMD’s current growth rates. That said, AMD has issued conservative full-year guides in the past, and it’s worth noting that the 2021 guidance only assumes mid-single-digit 2021 PC market unit growth. Given how strong unit growth is right now in spite of supply constraints, that outlook might prove conservative, even if one assumes WFH-related demand moderates some during the back half of the year.
• GPUs were the one soft spot revenue-wise in Q4. Thanks in part to lower sales related to cloud gaming engagements, GPU sales were down Y/Y in spite of the aforementioned gaming hardware strength and the rollout of AMD’s first RDNA 2 gaming GPUs. For comparison, Nvidia’s January quarter revenue consensus implies 55% Y/Y growth. On the bright side, AMD did sound optimistic about 2021 being a growth year for GPU sales, and CEO Lisa Su suggested that server GPU sales could double this year off a relatively small base, as major HPC design wins start kicking in.

While GPU sales fell Y/Y, AMD’s desktop and notebook CPU sales were each up by double-digit percentages.

• One other thing that might be weighing on shares: Non-GAAP operating expenses rose 45% Y/Y to $789M (topping guidance of ~$750M), and are forecast to grow to ~$830M in Q1. But with Intel’s manufacturing stumbles having arguably opened up a major opportunity for AMD to grab share over the next couple of years, there’s a logic to dialing up both R&D and go-to-market spending at this juncture.
• Overall, I don’t think there’s much to complain about here, outside of maybe the GPU sales. While it’s worth paying attention to what moves Pat Gelsinger chooses to make after taking over as Intel’s CEO next month, AMD’s CPU share-gain story still looks very much intact for now, and it’s hard to object to the current demand environment. As with Nvidia, the major question is how much growth is now priced in for AMD. With the caveat that this EPS figure might end up proving low, AMD still trades for 36 times a 2022 EPS consensus of $2.51.

GameStop-Like Gamma Squeezes Are Bound to (Eventually) Produce an Institutional Response

But until then, things could get even crazier.

Sources: GameStop’s 1-month stock chart, Wallstreetbets

• Jeff Bezos’ line about social media being a nuance destruction machine feels quite relevant to the heated Twitter debates going on about the massive short/gamma-squeezes that determined Robinhooders/Redditors are driving in names such as GameStop, BlackBerry, Express and Nikola. Broadly speaking, I think all of the following are true:

a) It’s not necessarily easy — as Bloomberg’s Matt Levine spelled out on Tuesday — for the SEC to use its existing playbook for dealing with perceived market manipulation to crack down on message board posters teaming up to bid up a stock that they claim to be bullish on (and making their intentions transparent to all), even if they’re hoping that their purchases trigger a short/gamma-squeeze.

b) As many others have opined, there’s arguably a “what’s good for the goose is good for the gander” element to what’s been going on with GameStop and other names.

c) There would probably be a lot less sympathy right now for what’s going on if the goal was to tank stocks (and trigger margin calls along the way) by aggressively shorting and buying OTM put options on select names with high margin maintenance requirements.

d) Whether or not such actions violate the letter of the law, they do violate the spirit of how equity markets are supposed to function — whether for retail investors, institutions or publicly-traded companies — in terms of price discovery, capital allocation or the simple relationship between a company’s valuation and expected future cash flows. To be fair, they’re far from the only thing responsible for creating such market distortions right now, but they’re certainly now responsible for some of the more egregious ones.

e) In the absence of an institutional response, human nature and media attention make it inevitable that this phenomenon will keep expanding in scope, turning markets into a speculative madhouse for a greater number of companies and ultimately burning legions of investors (both retail and institutional) even as it makes some others rich.

• In light of all that, I’d be quite surprised if we don’t ultimately see a strong institutional response — from the SEC, options market makers and/or others — to put an end to this phenomenon. Defenses founded on arguments that others, too, have been behaving badly will probably work as well as they do in traffic court.
• But until then, wagering against anything with a large percentage of its float shorted might amount to playing with fire. Those willing to brave the waters might want to spread out their bets against overvalued names to keep one or two from blowing up their accounts, or perhaps stick with ETF trades (either shorting an ETF that has meaningful exposure to such companies, or going long an inverse ETF that does).

Signs Point to Strong Q4 Social Media Ad Spend

It looks like growth rates accelerated sharply last quarter.

Sources: Mizuho (research report), Socialbakers

• Mizuho reports that a major online ad agency it spoke with saw its clients’ Facebook-related ad spend grow 39% annually in Q4, thanks to both strong click/impression growth and improved ad pricing. The Q4 figure is up sharply from Q3’s 16% and is the highest growth rate seen by the agency over the last two years. The agency also indicated Facebook ad spend has remained strong in Q1.

• Separately, social media ad agency Socialbakers reported that the brand social media ad spend it tracked was up 56% annually at the end of 2020, and that peak holiday season spend was 50% higher than the peak reached during the 2019 holiday season. Spending growth was particularly strong in North America.

• Mizuho also talked with a Google ad agency: It reported that its Google-related ad spend rose 12.1% annually in Q4, a slight improvement from Q3’s 10.9%. Double-digit increases in spending from retail/consumer goods and financial services advertisers were partly offset by a 20% drop in travel ad spend.
• One other interesting comment in the Mizuho report: The ad agency that provided the Facebook data said that (unlike brand advertisers) it’s still difficult for direct response advertisers to build Twitter ad campaigns “because users are mostly consuming news content passively on the platform,” and thus sharing far less purchase intent data than they do on other platforms. Twitter posted stronger-than-expected Q3 ad sales, but has had its share of issues with its direct response ad business (mobile app install ads particularly) over the last two years.
• Overall, Mizuho and Socialbakers’ data provide additional signs that Q4 was a pretty strong quarter for social media ad spend, amid elevated e-commerce spend, political ad buys and an accelerated shift in ad budgets from offline to online channels. Facebook reports on Wednesday afternoon; Snap and Pinterest report on Feb. 4, and Twitter reports on Feb. 9.

Disclosure: Long Facebook. I expect a strong Q4 report, but would note that with the stock up about 13% from its mid-January low, some of the aforementioned good news is likely now priced in.

Chip Shortages (and Price Hikes) Remain a Big Deal

Demand is still well ahead of supply in many areas.

Sources: Reuters (I, II) Nikkei, Digitimes (I, II, III, IV), Texas Instruments (earnings report), Unhedged (TI’s CC transcript)

• Following a plea from Berlin to help ease the chip shortages that German automakers (like many other automakers) are dealing with, Taiwan’s Economics Ministry promised that TSMC will prioritize the production of auto chips (often made using more mature processes) to the extent that it can boost capacity.
• Separately, Nikkei reports that TSMC, UMC and other Taiwanese chip manufacturers are mulling auto chip price hikes of up to 15% in response to this supply-demand imbalance. ASE (a major chip packaging/testing firm) was said to be mulling price hikes of about 10%. The report follows confirmed price hikes from major auto chip suppliers such as NXP and Renesas.
• Digitimes, which these days can be trusted to share fresh dispatches about chip industry craziness on a near-daily basis, has reported that some Taiwanese chip developers have chosen to “decelerate the development of their new chip products” due to tight wafer supplies for advanced TSMC processes. It also claims Mediatek chose to use TSMC’s N6 (6nm) process rather than its N5 (5nm) process to make its recently-launched Dimensity 1100 and 1200 5G SoCs due to N5 wafer tightness (likely caused in large part by big orders from Apple).
• Also mentioned by Digitimes this week: Prices for resistors and monitor display panels are rising, and notebook ODMs both expect stable orders in 1H21 and are beginning to have “clearer” visibility for second-half demand. Notebook growth rates will probably cool later this year as comps get tougher and more offices and schools reopen, but with PC upgrade cycles having steadily lengthened over the last decade and legions of companies rethinking their remote work policies in recent months, 2021 is still shaping up to be a pretty strong year on the whole.
• Though it’s seeing a bit of post-earnings profit-taking, Texas Instruments’ Q4 results and Q1 guidance (both soundly above consensus) largely fit with this narrative. Revenue was up 22% annually in Q4 thanks to strong demand from automotive, industrial and personal electronics clients, and the midpoint of TI’s Q1 sales guidance implies 19% growth.
• The one way in which TI’s commentary broke with the narrative: The company insists its fabs and supply chain control have kept it from seeing major supply constraints thus far, and that as a result, it hasn’t found it necessary to raise prices.
• An environment like the current one — one where shortages, aggressive ordering and stretched lead times are common — could eventually produce an inventory correction as demand cools. But Q1 and Q2 are clearly poised to be very good quarters for the chip industry. Upcoming earnings reports of note: Xilinx (1/27), Western Digital (1/28), Skyworks (1/28), NXP (2/1), Qualcomm (2/3), Qorvo (2/3).

Disclosure: Long Western Digital. Started a position last week due to its valuation, strong demand within several key end-markets (notebooks, game consoles, cloud servers, etc.) and the potential for NAND pricing to improve thanks to healthy demand and easing capex growth. Would caution that the stock has a history of being volatile post-earnings.