Tension was running high on Thursday afternoon in the US following a rough day for Meta, which unnerved investors with its cost outlook and capex plan, both of which underscored just how expensive building the future’s likely to be.
That was the set up for results from AI bellwethers Microsoft and Alphabet.
Naturally, both companies suggested the sky’s the limit for their AI-enabled products and services. At Microsoft, Copilot is “driving better business outcomes across every role and industry” as part of what Satya Nadella described as “a new era of AI transformation.”
Over at Alphabet, the Gemini era’s “well under way” and has “great momentum.” Contrary to media reports and generalized concerns that Google might be playing from behind for a change, Sundar Pichai said the company’s a leader in AI research and infrastructure, and remains “well positioned… for the next wave of AI innovation.”
Now to the numbers. Microsoft beat on the top line, but not by a lot. Revenue of $61.90 billion was up 17% YoY. Consensus was looking for $60.90 billion.
Fiscal Q3 was the second straight quarter during which sales grew 17% or better YoY.
Operating income was better than expected at $27.60 billion, up 23%. Consensus there was $26.22 billion. EPS of $2.94 beat by a dime.
As usual, it was all about Azure and on that front, 31% YoY growth marked an acceleration from the prior two quarters’ 28% pace.
Fiscal Q3 was the best quarter for Azure growth in a year. Don’t prejudge the market reaction. It’s all about the guide and, more to the point, what kind of mood traders are in by the time Amy Hood gets around to it.
Intelligent Cloud revenue was $26.70 billion, a 21% increase, but not a huge beat. The estimate was $26.25 billion.
Capex was $14 billion all in. I’m not sure if that’s comparable to consensus, but if it is, it’s high. From what I can see on Bloomberg, the estimate was around $11.30 billion.
As for the house of Pichai and Porat, revenue of $80.54 billion was a beat. The ex-TAC print was too.
Top-line growth at Alphabet was the briskest in two years and now looks to have “normalized,” where that just means we’re moving beyond COVID-era volatility.
At the risk of giving short shrift to this or that nuance, everything looked decent from Alphabet. YouTube ads brought in $8.09 billion versus $7.73 billion expected. Cloud revenue, which investors are starting to watch more keenly, was $9.57 billion, up 28% and $200 million (give or take) ahead of consensus.
Operating margin, at 32%, was much better than expected. 340bps better, in fact. Operating income of $25.47 billion handily topped estimates. Consensus there was $22.40 billion. EPS of $1.89 was a big beat (the Street expected $1.53).
For whatever this is worth (probably a lot considering how much Google employees make), the company had 4,000 fewer workers than expected at quarter-end, and 10,000 fewer versus a year ago.
Alphabet also announced a first-ever dividend and lifted buybacks, taking a page out of Meta’s playbook from last quarter.
Ultimately, both Nadella and Pichai turned in decent reports from where I’m sitting. Google’s — sorry, Alphabet‘s — was pretty clearly the stronger of the two. My guess is the stock will reflect as much, particularly given the decision to return cash to shareholders.
GOOG revenue growth +15%, a +200bp acceleration from 4Q growth. Search +14%, much higher growth in YouTube +21% and Cloud +26%. Search is still ~60% of revenues but company says YouTube + Cloud will exit 2024 at combined $100BN annual run-rate.
Testing AI content in Search, claims good results and that cost will be manageable, claims advertisers supposedly using AI tools to strong effect. Gross and EBIT margins up, much emphasis on consolidating teams and “durably re-engineering cost base” to accommodate “headwinds” of higher depreciation and technology investment. Speaking of, capex $12BN doubled YOY and will stay at or above $12BN/qtr this year – so GOOG sees and raises META’s $35-40BN with at least $48BN of its own. The company doesn’t explicitly guide, but CFO seemed to emphasize will lap increasingly tough comps, including twice noting Search growth led by “APAC-based retailer” starting 2Q last year.
So – growth accelerated this quarter, tougher comps coming, monster AI spend growing . . . but GOOG wisely said “efficiency” a lot, played the dividend card, and investors are grateful to hear no musing about multi-year investment cycles and revenue-light product scaling. Not so long along, some were saying GOOG looked like AI roadkill; those cat-calls are probably silenced, and least I checked, GOOG was the “cheapest” valued of the Mega names.
Nice plug…..
MSFT rev +17%, slight deceleration from Dec qtr but comps were +500bp tougher. Azure +31% with +700bp growth from AI, claims Azure ubiquitous “for pretty much anybody who is doing any AI project”, similar to GOOG’s rhetoric.
Unclear to me – maybe others know the details better – exactly where else AI growth is showing up, but I think mostly in commercial Office. Cloud is the largest and fastest-growing segment, Productivity is growing well (“mature” Office is growing low-teens), Personal Computing’s underlying growth is low excluding Activision but that segment is a collection of businesses investors don’t care as much about. Maybe we briefly cared about Search and Ad but its revenue incl TAC has not accelerated off +LSD so Bing, not eating Google’s lunch after all, has retreated to “don’t care” status (my view).
Under the covers, gross margin looks (to me) under some pressure in each segment, but that is offset by mix shift to higher-margin segments and leveraging opex, so overall margins are improving – but are guided to decline -100bp in FY2025. Capex $14BN and will increase “materially” sequentially in Jun F4Q24 then increase again in FY2025 as company is still capacity-constrained – so MSFT’s CY 2024 capex plan looks equal or higher to GOOG’s?
Guidance implies a June qtr revenue slowdown to +13-15% with slightly lower margins. MSFT faces much harder revenue comps starting Sep F1Q25. The valuation here is – well, none of these names look set to tolerate a marked slowdown, and I’d think MSFT is no exception.