More cloud capex, and consumer strength?

23 July 2023 | Issue #3 - Mentions $GOOG, $MSFT, $META, $V, $AMZN, $PDD, $BABA, $SE

Welcome to the third edition of Tech takes from the cheap seats. This will be my public journal, where I aim to write weekly on tech and consumer news and trends that I thought were interesting.

Through an investing lens, I break down and analyze these topics, providing valuable insights in a concise format.

Let’s dig in.

Weirdest recession ever

Throughout this week, several influential companies unveiled their Q2 results, alongside the eagerly anticipated US GDP report, offering us a glimpse into the economy's underlying health. Among the key players sharing interesting insights were Alphabet, Meta, Microsoft, Visa, and other noteworthy names.

The US GDP rose at a Seasonally Adjusted Annualised Rate of 2.4% in 2Q23, accelerating from a 2% increase in 1Q23 and consumer spending came in stronger than economists expected at 1.6% vs 1.2%. Commentary from travel, leisure and credit card companies broadly noted the resiliency in consumer spending across travel but online ad companies also mentioned strength in the retail vertical.

From Platform Aeronaught

My quick takeaways for each segment of the travel+leisure economy:

Cruise: Incredibly strong demand both Y/Y and vs 2019. Consumers are flocking back to cruises and we haven’t seen much deceleration in demand trends (yet)

Airlines: Demand overall remains strong (especially international) but seeing some pockets of domestic weakness (from ALK and HA) and business/corporate continues to slowly recover.

Food Delivery: Northern Europe (incl the UK) is seeing strength and normalization but Southern Europe is lagging.

Hotels: Pricing and consumer demand is still strong, early bookings (especially group/corporate) for 2024 are well ahead of pre-covid levels. International still recovering but China back to 100% of 2019 RevPAR levels

Car Rental: Lots of Americans (and Europeans) traveling this summer leading to continued peak RPD pricing. We’ll see what things look like into September but late summer continues to be strong.

From Google’s earnings

In Google advertising, Search and other revenues grew 5% year-on-year led by solid growth of the retail vertical.

… Let's pivot to Retail where we had a good quarter. Profitability remains a top theme for retailers, so solutions like PMax that drive bottom line value continue to do well. We also continue to see success in helping businesses unlock efficient growth and deliver on their omni-channel goals. Take Ace Hardware who tapped into AI powered search and omni bidding to capture increased seasonal demand leading up to Memorial Day. This drove increases across online sales, store visits and in-store sales resulting in 87% year-over-year growth in omni revenue from Google Ads and led to one of the largest revenue weeks ever for Ace store owners.

and Meta’s earnings

Within ad revenue, the online commerce vertical was the largest contributor to year-over-year growth, followed by entertainment and media, and CPG. Online commerce benefited from strong spend among advertisers in China reaching customers in other markets.

Economists and strategists alike have been calling for a recession in 2023 since last year due to monetary policy choices in order to tame inflation - so what gives? From the WSJ: How the U.S. Economy Is Sticking the Soft Landing

To see what an economic soft landing looks like, search no further than business hiring.

Parts of the economy are cooling, just as the Federal Reserve would like to see to combat inflation. Freight railroads, for instance, are seeing shipping volumes decline. Construction firms are cutting back on equipment purchases. A vending-machine company’s customers are negotiating prices downward.

Yet the key to a measured, inflation-busting slowdown that doesn’t sink the economy lies in whether companies hold on to workers or lay them off. And the answer, in an economy that otherwise can send repeated mixed signals, is clear: They are making a priority of keeping workers. Apple, for one, is avoiding layoffs despite economic uncertainty.

All those retained workers, in turn, are spending their paychecks, albeit more slowly. So the economy appears to be steadily cooling, while averting the long-anticipated recession.

Fresh economic data this week reinforced optimism that inflation can fall without the U.S. suffering a recession. Economic output accelerated in recent months on the back of solid consumer spending. Inflation cooled to 3% in June, according to the Fed’s preferred gauge. And wage growth, while still elevated, slowed, the Labor Department said Friday.

I also thought this article was interesting, and I am by no means suggesting that Swifties are solely responsible for saving the US economy, but it’s hard to ignore the enormous impact that Taylor Swift’s Eras Tour is making on every city she visits.

Transformers… humans in disguise?

The FT came out with a fascinating article last weekend which went into the history behind ‘transformers’ - the technology architecture behind language processing applications such as Large Language Models (LLMs).

“What came out of ‘Attention is All You Need’ is the basis for effectively every generative AI company using a large language model. I mean it’s in everything. That’s the most insane thing about it,” says Jill Chase, a partner at CapitalG, Alphabet’s growth fund, where she focuses on AI investments. “All these products exist because of the transformer.”

The entire article is worth a read but it was basically hatched out of a team of AI researchers at Google who were looking for new methods for AI to process languages faster and more accurately. It’s evolved to become revolutionary for the AI industry, and although it was conceived within Google, the company has arguably

dropped the ball on commercialising it. The team from the monumental research paper have all left Google to found their own start-ups (Cohere, Character.AI, Inceptive, Essential.AI and Near, with two of them being unicorns) to create products using the transformer concept. Below are some interesting quotes highlighting the innovator’s dilemma large incumbents such as Google have to grapple with.

Each of the co-authors who spoke to the FT said they wanted to discover what the toolbox they had created was capable of. “The years after the transformer were some of the most fertile years in research. It became apparent . . . the models would get smarter with more feedback,” Vaswani says. “It was too compelling not to pursue this.”

But they also found that Google was not structured in a way that allowed for risk-taking entrepreneurialism, or launching new products quickly. It would require building a “new kind of software . . . computers you can talk to,” Vaswani adds. “It seemed easier to bring that vision to light outside of Google.” He would eventually leave in 2021.

“The reason why I left Google was that I actually didn’t see enough adoption in the products that I was using. They weren’t changing. They weren’t modernising. They weren’t adopting this tech. I just wasn’t seeing this large language model tech actually reach the places that it needed to reach,” he says. In 2019, he quit Google to found Cohere, a generative AI start-up that is valued at more than $2bn, with investment from Nvidia, Oracle and Salesforce, among others. Gomez is interested in applying large language models to business problems from banking and retail to customer service. “For us, it’s about lowering the barrier to access,” he says. “Every developer should be able to build with this stuff.”

“Google was an amazing place, but they wanted to optimise for the existing products . . . so things were moving very slowly,” says Parmar.“ I wanted to take this very capable technology, and build new novel products out of it. And that was a big motivation to leave.”

Related: Large language models, explained with a minimum of math and jargon h/t @fabknowledge

Can’t get enough GPUs

Three of the biggest hyperscalers - Microsoft, Google and Meta - reported their quarterly results this week, and while there was plenty of information to digest from it, I’ll write to some highlights that I thought were worth calling out.

I wrote last week that there seemed to be a disconnect between hyperscaler capex spending intentions/forecasts and Nvidia GPU demand expectations. This week’s results helped us paint more of the picture.

From Microsoft’s earnings call

At a total company level, revenue growth from our commercial business will continue to be driven by the Microsoft Cloud and will again outpace the growth from our consumer business. Even with strong demand and a leadership position, growth from our AI services will be gradual as Azure AI scales and our Copilots reach general availability dates. So for FY 2024, the impact will be weighted towards H2.

To support our Microsoft Cloud growth and demand for our AI platform, we will accelerate investment in our cloud infrastructure. We expect capital expenditures to increase sequentially each quarter through the year as we scale to meet demand signals.

From Google’s earnings call

Finally, as it relates to CapEx, in Q2, the largest component was for servers, which included a meaningful increase in our investments in AI compute.

The sequential step-up in the second quarter was lower than anticipated for two reasons. First, with respect to office facilities, we continue to moderate the pace of fit-outs and ground-up construction to reflect the slower expected pace of head count growth. Second, there were delays in certain data center construction projects. We expect elevated levels of investment in our technical infrastructure increasing through the back half of 2023 and continuing to grow in 2024. The primary driver is to support the opportunities we see in AI across Alphabet, including investments in GPUs and proprietary TPUs as well as data center capacity.

From Meta’s earnings call

Turning now to the CapEx outlook, we expect capital expenditures to be in the range of $27 billion to $30 billion, lowered from our prior estimate of $30 billion to $33 billion. The reduced forecast is due to both cost savings, particularly on non-AI servers, as well as shifts in CapEx into 2024 from delays in projects and equipment deliveries rather than a reduction in overall investment plans.

Looking ahead, while we continue to refine our plans as we progress throughout the year, we currently expect total capital expenditures to grow in 2024 driven by our investments across both data centers and servers, particularly in support of our AI work.

The rush towards GenAI picks and shovels is well and truly here, and two of the largest hyperscalers aren’t able to meet their need for GPUs. These comments from the companies drove consensus to increase their forecasts for aggregate capex of the three by $12.8bn in 2024 and 2025 combined. There are still questions up in the air on whether these investments in GPUs make sense from a financial return perspective, and if the end demand is there to support it. It comes at a time where data from SimilarWeb in June suggested that monthly visits to ChatGPT has peaked, and if the history of pace of enterprise adoption of technology was any guide (see: on-prem to cloud transition), the returns could take longer to bear fruit than expected (apart from maybe Meta, where AI is already driving ROI in the form of better ad tools and content discovery). To be clear, 72% of the $12.8bn in aggregate capex is coming from Microsoft alone, so it seems that the street is expecting them to monetise it quicker than the other two (or it could be a simple case of using yoy growth rates on higher than normal sequential build outs of capex)

Related: This tweet suggested that Microsoft was pricing co-pilot (at $30/month) at breakeven, which may be too aggressive on cost estimates given ChatGPT plus is at $20/month. Ben Thompson had more insight on the additional revenue streams Microsoft would capture (that Nadella mentioned on the call) in his earnings review.

Bottlenecks to enterprise adoption of new technologies: Banks’ Cautious Approach to Generative AI Is More Internal Than Customer-Facing

How Salesforce, AT&T, and Workday are putting generative AI to work

Who’s behind in GenAI?

This interview with Adam Selipsky, CEO of AWS was a good read. Various industry participants have stated that AWS is generally behind on GenAI (though there’s a lot of conflicting noise on this), so it was interesting to get the thoughts from the leader of the biggest industry player.

RW: Microsoft has been working with OpenAI for four years. With that amount of experience, and that head start, they have a real advantage in perfecting or optimising the platform for this kind of technology. Are you behind?

AS: No, I think that’s just flat out untrue. I’ve seen numbers recently published that suggested AWS is at least twice as large as [Microsoft’s cloud computing service] Azure. So our experience in running technology and systems at massive scale is unparalleled. We have far more customers across every industry, every use case, every country . . . So I think, just in terms of overall cloud capabilities, it’s really not even close. We have by far the broadest and deepest set of capabilities and of customers.

Then, specifically in machine learning, we released SageMaker in 2017. So we’ve had a machine learning platform since 2017, and have over 100,000 customers using it already. We’ve even been building our own LLMs. And not all of our large competitors are choosing to build their own models. Some have just outsourced the models that they’re going to use to other companies, which is an interesting choice, but not the choice we’re making. It’s interesting that somebody who’s not running their own models [like Microsoft, which has outsourced to OpenAI] would argue they’ve got such deep expertise in this area.

Related: Google on its earnings call said that >70% of GenAI unicorns are GCP customers, including Cohere, Jasper, Typeface and many more. Take this with a grain of salt though, as it seems at least Cohere is going multi-cloud for its infrastructure needs.

We don’t want your cheap cross-border goods

From TechInAsia

The Indonesian Ministry of Trade plans to limit the sales of imported goods on ecommerce platforms as it is revising its regulations on online trade.

Isy Karim, the director general of domestic trade, said that the new rules will serve to protect Indonesia’s MSMEs.

“There are several [policies] being revised, one of which involves setting a minimum limit of US$100 per unit of goods traded on marketplaces by foreign merchants,” said Karim, as reported by Katadata.

This is more punitive to cross-border ecommerce operators than the end of tax exemption that Brazil announced earlier this year on international orders up to $50, as it’s an outright ban. The policies are looking to address complaints from local sellers and merchants who have to compete with cheap goods from Asian giants such as AliExpress, Shein and Temu (owned by PinDuoDuo). The restriction effectively prevents the entry into the regions from Temu and TikTok Project S (its white-label goods vertical), which has been a key concern for investors in Sea over the past year. Their (PDD’s) strategy historically has been to connect factories in China directly to consumers, bypassing the middlemen and providing large bulk orders at a fraction of the price. Over time they would use the demand that they’ve built to sign up local merchants in the countries and move up-market in product SKUs (a playbook that Shopee has used to penetrate Brazil).

Related: Shein Makes an Aggressive Pitch to Woo US Amazon Sellers

That’s all for this week. If you’ve made it this far, thanks for reading. If you’ve enjoyed this newsletter, consider subscribing or sharing with a friend

I welcome any thoughts or feedback, feel free to shoot me an email at portseacapital@gmail.com. None of this is investment advice, do your own due diligence.

Tickers: GOOG 0.00%↑, MSFT 0.00%↑ , META 0.00%↑ , V 0.00%↑ , AMZN 0.00%↑ , PDD 0.00%↑ , BABA 0.00%↑ , SE 0.00%↑