Tech Thoughts Newsletter – 3 February 2023.

Conclusion

The market rally and “risk-on” trade has continued this week, driving up tech. The fed remains a focus for the market, which seems to be behaving as though inflation worries and rate hikes are over – we’re less sure. Tech fundamentals are deteriorating in places, and the timing and strength of the recovery remains uncertain. We continue to focus on fundamentals and intrinsic value of the companies within the portfolio. 

Portfolio

We had reduced our holding in Microsoft ahead of numbers as we wrote last week – and we started to build the position again by the end of last week and the beginning of this week.

Onto results: 

Firstly the big tech reporters Apple (owned – small position), Amazon (owned – small position), Alphabet (owned) and Meta (not owned) – all had slowing top lines, all with a focus on cost cutting and efficiency to drive earnings upside. 

• Apple missed already fairly reduced expectations – it had warned on iPhone shipments given the COVID interruptions in China back in November (though important to note that it did not warn on demand), and so the market was expecting a weaker iPhone revenue number. We also expected very weak Mac results, given the product releases this quarter, and a services slowdown, given the weight of gaming. 
• They reported a decline in revenue (-5.5% yr/yr) which missed expectations. The iPhone revenue number was particularly weak, again driven by the known, but clearly still underestimated, supply issues. 
• Services growth slowed (to 6% yr/yr). We think Services (mostly gaming and digital advertising within apps) can be a good early indicator of the macro – in particular the feedback loop for games developers in terms of in-game purchases and their ROI on advertising within the app store is amongst the fastest in advertising. The implication is that the iPhone installed base (which so far has seemed pretty immunne) is being hurt by the macro. 
• Looking at the forward guidance and trying to unpick demand going forwards, iPhone sales are guided to accelerate – that’s clearly helped by the unwind of last quarter’s supply disruptions, and a potential for inventory rebuild. We’re less convinced that you’ll see a significant acceleration in demand – that needs to be driven by an iPhone upgrade cycle that looks to be lengthening. 
• The forward guidance for Mac and iPad are for a double digit decline in the March quarter – again driven by macro headwinds. 
• Meta was better than expected, the biggest positive surprise being the bigger than expected capex (down ~$4bn vs expectations) and opex reduction (down ~$5bn). The call started with Mark Zuckerberg pitching 2023 as “the year of efficiency” (which as we’ve commented before seems to be the driver of much of tech right now). They’ve also increased their buyback to $40bn (and bought back $7bn of stock in Q4). 
• Q4 revs of $32.2bn were ~2% ahead of consensus. Q1 guide of $26-28.5bn in line with consensus and will mean a 2% yr/yr decrease in revenues. 
• With top line going backwards (we have $113bn revs vs $117bn for 2022, from $118bn in 2021), Meta is really about (1) trying to limit/ the share loss in digital advertising and the decline in top line; and (2) cutting opex to get compounded earnings growth. 
• The call sounded positive on Reels monetisation and improving ad performance (leveraging some of their own internal data/AI to try to get back some of the targeting that was lost to ATT). 
• Alphabet just missed revenue expectations ($63.1bn reported vs cons $63.2bn) but the real alarm bells were in YouTube, which declined 8% (vs search -2%). For us YouTube is clearly much more sensitive to new competition and increased industry inventory, and to ATT. Google Search is more resilient – still hurt somewhat by the macro and the consumer, but as we touch on below, we think more cyclical than structural. 
• Google Cloud grew strongly +32%, though there were the same comments on customer spend optimisation that we’ve heard from all the players. 
• On their costs, CFO Ruth Porat commented that they would reengineer the cost base and slow the pace of hiring. Again, no big surprises. 
• Clearly AI (with ChatGPT and the potential structural threat on everyone’s minds) was a big topic on the call, with CEO Sundar Pichai announcing that LaMDA and other language models will be available. 
• Amazon reported arguably the most robust results of the lot with revenues of $149bn coming in ahead of cons $146bn
• The retail environment in Q4 we know was slightly better. AWS slowed but as per expectations (and has continued to decelerate this quarter) – again similar to Microsoft/Google there is an ongoing process of spend optimisation – their customers across the board are trying to get spend down as the economic picture deteriorates.
• A lot of the call came down to Amazon’s ability to cut costs. CEO Andy Jassey spoke of a different era and different priorities – his now is to drive cost efficiencies. The $0-4bn EBIT guide is within the expected range – getting to the top end of that will be about costs and revenue. 

Our conclusion – while top lines are slowing across the board, we think we should differentiate between the cyclical and structural. Meta’s top line growth decline is for us a structural issue of competition and deteriorating barriers to entry, while Apple, Google Search, and Amazon (AWS) are more cyclical – around advertising growth and consumer sentiment. 

For Apple, its position in the industry is still the strongest – indeed arguably it’s strengthening (if you consider the broader smartphone market). For Google, we think search is the most resilient and has the biggest barriers to entry within the digital advertising space. AWS has a captive customer base as the result of its retail business – this provides it with meaningful barriers to entry in cloud services – and ultimately Amazon is moving from a low returning retail business to a higher returns cloud business. 

Meta CEO Mark Zuckerberg commented that 2023 will be the year of efficiency – we think this is true for tech more broadly and tech calls across the board have been focused on opex and capex discipline. We saw a myriad more workforce cuts this week (Workday, Salesforce, Okta, Match, Pinterest). 

There is clearly room for most of tech to cut opex further, but more important for us is their ability to improve top line growth. Without that it’s very hard to turn this into a compounding returns thesis. That means we continue not to hold Meta – where we view the runway for opex reduction as limited – given the spend on AI is in large part about mitigating the structural ATT issue and maintaining some competitive advantage. 

Semis markets still bifurcating…Autos still the winners: 

• NXP (owned) reported a slight beat. Overall revenue declined 9% qtr/qtr with auto revenue flat. Auto semis remains a supply constrained area and an area where there is still positive pricing – we see better qtr/qtr growth rates for all the auto exposed semis stocks. 
• For the broader business they’ve reduced utilisation rates in areas where they’re seeing market softness – again it’s key for us that with the lower utilisation levels you’re seeing across all the players ultimately feeds through to healthier looking inventory levels – that isn’t happening yet. 
• Tougher times call for tougher internal management and NXP continue to manage channel inventory to ~one month below their long-term target (equivalent of ~$500m revenue). That means keeping more inventory on their balance sheet and ultimately means $500m of revenues they could (and previously would) have recorded in Q4 as they kept shipping. Instead they’ll now keep on their books in order to minimise channel inventory buildup/obsolescence if demand falls, and to potentially reship it to areas where there’s sell through.
• NXP flagged China macro which remains an area of uncertainty and potential downside for all players 
• Infineon (owned) delivered a good beat and raise, – Q1 (Dec year end) revenue $3.95bn vs cons $3.99bn. Q2 guide $3.9bn vs cons $3.8bn but with a change in FX assumption so a higher implicit upgrade. The real upside came on margin – 28% for Q1 vs 25% cons. 
• For the full year the revenue outlook (which was upgraded in November) is kept $15.5bn, but the FX assumption is going from $1.05 from $1.0, so an implicit 10% upgrade for the second half of the year. And margin going to 25% from 24% (despite the currency headwind so implicitly more). 
• Again looking at the relative strength – yr/yr growth of auto in the quarter was +35% yr/yr. Similar to NXP, there were comments that auto semis supply remains tight (they are fully booked for FY23) and that pricing dynamics remain strong. 
• In the comments, the shift to EV which comes with semi content increases, and renewables is still driving demand, while smartphones/PC/Data centre are all weaker – in line with what we’re hearing from others. 
• The EV price war continues: We’ve commented before on the pricing dynamics within EV. Tesla seems to have found that in order to grow sales it needs to have a more compelling value proposition for customers – which led to its ~20% price cuts earlier this year. This week, Ford too announced that it would cut prices of its Mustang Mach-E by ~$4,500. We continue to think that pricing for EV will remain under pressure – also from new Chinese competition, and related to government mandates requiring OEMs to increase their EV sales. That’s good news for Infineon and NXP – stimulating the consumer shift to EV. It’s clearly very bad news for the auto OEMs and their margin profile. 
• Relatedly, Ford results showed that a number of car makers are still having problems with logistics and components. The company stated that they left $2bn on the table not being able to supply customers. Ford is trying to ramp up their EV production capacity but is struggling and is running at a monthly capacity of 12,000 by 4Q while their target is to run at 50,000 per month by the end of 2023. Good news for their semis suppliers. 

Cloud and datacentre – choppy around product upgrades, question marks over a pause in underlying demand

• AMD (owned) results were happily a world apart from Intel last week. Good beat, and better guide. Q4 $5.6bn revs vs. cons $5.5bn, Q1 guide $5.3bn (vs a soft consensus of close to $5bn). 5% qtr/qtr decline is very solid in the context of the still ongoing PC inventory correction. Gross margin 50%. Segment-wise datacentre grew 42% yr/yr; client -51% yr/yr, hurt by PC market and inventory correction clearly; gaming -7%.
• While they didn’t provide specific guidance, they did say datacentre and embedded will grow due to share gains, while Client and gaming will decline. 
• Non-GAAP gross margin to be flattish in H1 and increase in H2. The gross margin guide is positive in that it indicates that they are able to keep pricing given their superior product (vs Intel) even with Intel quite likely putting pressure on pricing – that sounds like AMD are reducing some older product pricing which speaks to the H2 increase on H1 but still a pretty minimal impact overall (it does look like they can still increase overall 2023 margin vs 2022). 
• There is increasing evidence that they are gaining share from and executing significantly better than Intel. We know their latest Genoa performance and energy efficiency gains have continued to outperform Intel’s latest efforts – and the reality is that what’s playing out in terms of share gains is very hard to reverse – it goes back to Intel falling significant behind TSMC’s process technology leadership (which AMD benefits from). It’s just not something that can reverse that quickly, however much Gelsinger is willing to spend. 
• The negatives: They are saying there are some elevated inventory levels at some of the cloud customers which is leading to expectations of a lower H1 followed by a recovery as that is worked through – that will be different for each customer. 
• They believe they’ve made progress on PC inventory in Q4 – and are still shipping below consumption/sell through. 
• They’re guiding for a 10% TAM decline this year (which goes back to Intel’s very unrealistic expectations from their PC day earlier this month)
• Super Micro already positively pre released their 2Q numbers so the focus in the report was completely on forward looking guidance.  The company is now guiding for sales of $6.5bn to $7.5bn for 2023 inline with the consensus forecast of $6.9bn. Clearly winning share in the server market.  
• Juniper quarterly sales fell slightly below forecasts growing 11.5% yr/yr while EPS was in line with expectations. The company states that the timing of supply and some logistical challenges affected sales negatively. Enterprise and cloud sales were strong while service provider sales were weaker. 
• The order backlog is still strong helping the company grow sales this year, though bookings are weaker down 20% yr/yr. The company states that the moderation in orders is due to a normalisation of order pattern after a strong built up orders during the component shortage period that the company suffered. We are not surprised about the carrier weakness (re wolfspeed, telco operators, Ericsson warning, also see a weaker cloud ordering in general for the moment.

Our conclusion: cloud and datacentre spending remains an area of uncertainty – and there are multiple factors at play, around hyperscalers cutting back spend; new product releases, inventory builds. Looking at the hyperscaler capex comments, while there is a focus on spend reduction overall, it’s not obvious that this capex reduction is happening at the datacentre investment level. Google CFO Ruth Porat commented that spend on tech infrastructure (servers/cloud/AI) would be higher. 

There are clearly share shifts happening (AMD and Super Micro both gaining at different parts of the value chain – arguably Super Micro benefitting from an early bet on AMD’s Genoa release). 

Elsewhere in semis, consumer still weak

• All of Infineon, AMD, NXP mentioned above called out consumer as the weakest part of their business, and all with elevated inventory levels. For all of them the visibility is lower given much of supply goes through China distributors. 
• Microchip had already reaffirmed its guidance earlier this year and reported solid results and better guidance. Their exposure towards industrial, automotive, aerospace and defense, data center and communications infrastructure end markets which make up 86% of net sales, remain solid, while consumer (mostly home appliances) was weak – more of the same. The company stated that it expects supply constraints to remain for much of 2023.
• Qualcomm revenue for the quarter was weak while guidance for the semis product business (QCT) was a bit more of a relief (-2.5% qtr/qtr) – and better than the headline revenues which includes the backward looking licenses business. We are not surprised in the weak quarter given the production problems for Apple and the weak performance Samsung had in the quarter.They comment that inventory levels are expected to remain elevated at least for the first half. We think they should benefit from more normalised production at Apple and the release of the Samsung S23 (single supplier, with no Samsung internal solution) – thoughthe environment clearly remains tough. 

Our conclusion: We continue to avoid direct exposure to consumer – the reality is that consumer exposed semis will still struggle to perform if consumer end demand remains weak. Consumer is an area particularly sensitive to supplier discipline, because so much consumer semi content is commoditised or fungible. There are times to be brave in consumer semis but we still don’t think it’s time yet, and we remain cautious.  

Qualcomm is good news for TSMC – its primary foundry supplier. 

Semi capex – good news this week for the semi capex space 

• We’ve seen both Micron and Hynix cut their capex this year given the weakness in the memory market, and everyone was waiting for Samsung to do the same. But this week Samsung reported and guided that it would keep spending at the same level (its capex came in at $39bn last year). That’s good news for the semicap space – most expected Samsung to cut. It also leads to potential upside to the guided spend for Hynix and Micron to keep share. Terrible for the DRAM market (we don’t hold any memory names)
• Texas Instruments raised its long term capex guide – it now expects to spend $5bn/year 2023-2026 (vs $3.5bn previously) and targets capital spending of 10-15% long term (from 10%). 

Our conclusion – semicap remains an area we’re happily exposed to – we own ASML, LAM Research, AMAT, KLA. Texas speaks to the supply shortages at trailing edge nodes and the tailwinds around 300mm investment from IDMs (Infineon too). 

Digital advertising – share losses for the “old” incumbents continues

• As we spoke about above, YouTube and Meta are both clearly being impacted by share losses in the industry. 
• Elsewhere Snap (not owned) significantly slowing (Q4 came in flat, Q1 is guided to decline 2-10% and is currently tracking down 7%) speaks to the ongoing pressure in the broader digital advertising space. Impacted by ATT/Increasing inventory (competition – the most obvious being TikTok) and the resulting impact on ASPs, weaker macro impacting overall spend. Updated chart from our digital advertising report below (Meta/Alphabet reporting this week too) – significant headwinds in the whole space:
• Publicis guided 2023 revenue to grow 3-5% organically above market expectations of flat growth.  The company states that they have not seen any major client behaviour changes but that they have seen some small cuts by some accounts. 

Our conclusion: The general advertising market is holding up well, and while overall digital growth is slowing (given the higher levels of maturity), the real issue for players is share losses to new competition. Worth noting as a slight aside than Amazon’s advertising business grew 23% ex FX. 

Gaming remaining a hit driven market 

• Sony (owned) reported a 3Q inline with expectation but with an interesting business mix. The gaming division surprised positively while the Pictures/movie division struggled somewhat. The company shipped a massive 7.1m PS5 in the quarter up from 3.9m the year before and raised its full year 23 target to 19m. “Gods of Ragnarök” success continues – selling 11m in the first 10 weeks which also contributed to the healthy development in the division.
• The company tweaked their sales forecast from 11.6 trillion yen  to 11.5 trillion yen while the operating income was slightly raised from 1.16 trillion yen to 1.18 trillion yen. In total very minor changes.

Our conclusion: The results do not change our view of the company although we think the good sales of PlayStation is a positive sign for the gaming unit/software sales going forward. 

For weekly insights on the latest market updates, please subscribe to our Tech Thoughts podcast.
For more information about Tech Thoughts, please click here.

We provide investors with access to category leading technology companies, globally. Our assets under management have a total value of more than €1bn, and our limited partners include institutions, family offices and entrepreneurs. Learn more about our funds here.

Enquiries
For enquiries, please contact:
Inge Heydorn, Partner, at inge.heydorn@gpbullhound.com
Jenny Hardy, Portfolio Manager, at jenny.hardy@gpbullhound.com
Nejla-Selma Salkovic, Analyst, at nejla-selma.salkovic@gpbullhound.com

About GP Bullhound
GP Bullhound is a leading technology advisory and investment firm, providing transaction advice and capital to the world’s best entrepreneurs and founders. Founded in 1999 in London and Menlo Park, the firm today has 13 offices spanning Europe, the US and Asia. For more information, please visit www.gpbullhound.com.