Q1 Update: Hyperscalers Had A Goldilocks Quarter
Our Hyperscaler Pick Of The Year Keeps Winning.
«The 2-minute version»
As the tariff tantrum engulfed market sentiment post-Liberation Day, there was growing speculation of hyperscalers pulling back their capex plans for the year in anticipation of weaker-than-expected AI demand.
Then the unexpected happened: Not only did all four hyperscalers stick to their full-year capex guidance, but Meta even raised theirs from $62.5B to $68B as capacity constraints continue to persist across all hyperscalers amid strengthening demand for their AI products and services. For the full year, total projected AI capex now stands at $323B, up from $317.5B in the previous quarter.
Will the Real AI MVP please stand up? Without a doubt, Microsoft stole the show in Q1 as its Azure cloud reaccelerated from the previous quarter at 35%. Not only that, Microsoft’s management doesn’t see any signs of slowdown in the growth rate of Azure in the coming quarter, estimating it to grow in the 34-35% range. OpenAI is a big reason why, while customers such as Coca-Cola and ServiceNow also expanded their Azure footprint during the quarter.
Don’t think we are back to the ZIRP era. With the AI train showing no signs of slowing down, it’s not like we are back to a “growth at all costs” mindset. On the contrary. With capex depreciation lurking, Amazon and Google are doubling down on pursuing efficiencies with custom silicon chips to drive better price performance for running AI workloads, while Microsoft is expected to decrease its pace of capex spending next fiscal year, which should be free cash flow accretive assuming a stable demand environment.
The award for the best-positioned hyperscaler in 2025 goes to… Read below to find out.
🎥Let’s set the stage: A “Goldilocks” quarter for AI capex
As the tariff tantrum engulfed market sentiment post “Liberation Day” on April 2, there was growing speculation that Big Tech would likely cut back on their capex (capital expenditure) plans for the year, as many cloud and AI initiatives in the US were likely to be pushed out during this period of macroeconomic uncertainty.
As a result, the four big hyperscalers that include Microsoft MSFT 0.00%↑, Amazon AMZN 0.00%↑ , Alphabet’s Google GOOG 0.00%↑, and Meta Platform META 0.00%↑ saw their market caps decline 10%, 17.8%, 11%, and 17.4%, respectively, between April 2 and April 7.
Furthermore, news of Microsoft cancelling up to 2 GW of data center projects in the US and Europe added even more fuel to investor pessimism before Q1 earnings season.
Then something unexpected happened.
First, ASML ASML 0.00%↑ and TSMC TSM 0.00%↑ reported their Q1 earnings, where they not only beat their revenue and earnings estimates but also held their full-year guidance steady.
Note that both ASML and TSMC sit at the top of the global AI compute stack, where TSMC is the most systemically important manufacturer on the planet, while ASML is its enabler.
What was particularly impressive was that TSMC confirmed that their AI revenue will double despite the US ban on AI chip exports to China, while they simultaneously raised their full-year capex outlook to cross $40B as they deploy resources towards advanced process nodes.
Then, we heard from the hyperscalers.
To all the doubters, these four companies did not back away from their previous capex guidance.
In fact, Meta even raised the midpoint of their full-year capex guidance range from $62.5B to $68B to support both their Gen AI and core business needs by putting more compute against their ads and recommendation engine.
So, now for the full year, the estimated cloud capex is estimated at around $323B, up from $317.5B in the previous quarter.
This is how it breaks down.
In the meantime, all four CEOs discussed the rapidly growing adoption of their AI products and services while continuing to face capacity constraints at the same time. A “Goldilocks” scenario indeed.
As Oppenheimer said, “Against a noisy/tariff backdrop, we view AI as the best/safest growth vector.”
But, let’s face it, we are not in a “growth at all costs” environment like in the ZIRP era.
On the contrary.
So, which hyperscaler is the best positioned to drive the maximum upside in this environment?
Let’s find out.
Microsoft: The Runway Winner of Q1’s AI Cloud Wars
Starting at a high level, Microsoft Azure grew their revenues at the fastest rate of 35% YoY, up from 31% in the previous quarter. Altimeter Capital’s Jamin Ball believes Azure's annualized revenue run rate is clocking ~$77B.
This was followed by GCP (Google Cloud Partners), which saw its revenue grow 28% YoY, down 200 basis points on a sequential basis to $12.26B, or an annualized rate of $49B.
Finally, Amazon AWS (Amazon Web Services) grew their revenues by 17% YoY, down from 19% in the previous quarter, to $29.2B, or an annualized rate of $117B.
While Amazon AWS grew at the slowest pace compared to Azure and GCP, we will point out that Amazon is still the market leader, with an estimated market share of 29%.
However, both Microsoft’s Azure and Google’s GCP are quickly gaining traction in market share over the last three quarters in FY24 against AWS, with their respective market shares at 22% and 12%.
Particularly, when it comes to Microsoft, its outperformance can be tied to one of its critical clients, OpenAI, which continues to scale and grow Azure usage, as noted by the 17% YoY growth rate in its Commercial bookings.
We were (extremely😀) pleasantly surprised by management’s next quarter Azure guide numbers for Azure growth projected to remain in the 34-35% growth range in Q2. That is a sign of strong & sustained yet resilient growth by Microsoft’s management despite the current uncertain narrative prevalent in the markets which is very encouraging.