Cloud, Chips and Margins: Amazon, Microsoft and Meta Q1 Recap
A deep dive into Q1 results from the tech giants, breaking down cloud momentum, infrastructure bets and where AI is really driving returns.
Most of the losses post-Liberation Day have been erased across equity indices. This is good, of course but it almost feels like the calm before a storm. From my perspective, it seems to be more of a redeployment of cash rather than the consensus being that everything is going to be okay. The reason why markets fell so hard is because investors were selling and getting out/cutting position sizing. But asset managers, hedge funds, pension funds, etc their job is to make money with money. Holding cash yields no return and therefore needs to be redeployed, which is what I believe is happening, as opposed to investors turning bullish again.
Cloud Dominance
As you may or may not know already, there are three main cloud providers in the industry: Amazon, Microsoft and Google. AWS remains the largest by market share, followed by Microsoft’s Azure and then Google Cloud. These businesses sit at the center of digital infrastructure and their results offer a clear view of how enterprise AI adoption is progressing. Let’s begin with Amazon.
Amazon
AWS was launched in 2006 and originally started as a way to offer excess computing capacity to external developers. At the time, Amazon needed robust internal infrastructure to support its operations, and the team realised there was a broader opportunity to productise that infrastructure. What began with basic storage and compute services quickly expanded into a full suite of tools used by startups and large enterprises alike. By being early, reliable and aggressively priced, AWS established a dominant position in the market. It became the default choice for developers and eventually the backbone for companies ranging from Netflix to NASA.
Before getting into the AWS results, I’ll give a higher-level review of the earnings. Amazon reported Q1 revenue of $155.7bn, up 9% YoY or 10% in constant currency. Operating income rose 20% to $18.4bn, while net income reached $17.1bn, compared to $3.2bn a year ago. The jump in profit includes a $3.3bn non-operating gain related to the revaluation of Amazon’s stake in Anthropic, following the conversion of its convertible notes into preferred equity. This remains a financial investment rather than a commercial partnership, but it is a relevant signal of Amazon’s exposure to the broader AI ecosystem.
Advertising revenue grew 19% YoY to $13.9bn, subscription services, which include Prime membership and digital content like Music and Video, grew 9% YoY to $11.7bn. Physical store revenue came in at $5.5bn, up 6%, though this remains a relatively minor part of the overall mix. Capex for the quarter was $24.3bn, with the majority going toward technology infrastructure. This includes continued investment in custom silicon, generative AI training capacity and the expansion of data centre footprint. Free cash flow for the trailing twelve months was $25.9bn, down from $50.1bn a year ago, reflecting higher investment levels across the business. While spending remains elevated, Amazon’s underlying cash generation remains solid, and the investment focus has clearly shifted toward infrastructure that supports AI workloads and cloud services. Now onto the most interesting part.
AWS reported revenue of $29.27bn in Q1 2025, up 16.9% YoY and 1.7% sequentially. While the YoY figure remains solid, the slowdown in sequential growth is notable. This was the weakest quarter-on-quarter increase in over two years and follows four consecutive quarters of more consistent momentum. The deceleration stands out given the broader narrative around rising demand for AI infrastructure.
Operating income came in at $11.5bn, with operating margin reaching 39.5%. This is the highest margin AWS has ever recorded and reflects a sharp improvement from 29.6% in the same quarter last year. The scale of the increase is impressive, particularly given that investment levels remain elevated. Rather than being driven by reduced spending, the margin expansion appears to come from improved infrastructure efficiency, more effective workload distribution, and greater adoption of AWS's custom silicon, which lowers unit costs.
Graviton chips are now widely adopted for general-purpose compute, while Trainium and Inferentia are used for model training and inference. These chips help lower delivery costs and reduce dependence on external suppliers. Management noted continued customer migration towards Graviton-based instances, particularly for workloads involving generative AI. This is becoming one of the platform’s most important levers for improving economics and reinforcing control over the AI stack.
AWS also introduced its own foundation model family, Amazon Nova, and added access to Claude 3.7 and Llama 4 through Bedrock. The platform now supports a range of third-party and in-house models, aiming to offer flexibility without forcing vendor lock-in. From a technical perspective, the capabilities are competitive. Commercially, however, visibility is limited. Amazon hasn’t disclosed how much of AWS’s growth is being driven by AI. Microsoft, by comparison, Microsoft does. Without similar detail from Amazon, it’s relatively difficult to measure the impact of recent launches.
Ultimately, AWS remains a highly profitable and strategically important business. The platform is delivering record margins and investing in infrastructure that should support future growth. But in the absence of stronger top-line momentum or clearer AI contribution data, the sense is that leadership in the next phase of cloud is still up for grabs.
Outlook
Management reiterated that AWS remains in the early stages of what they see as a multi-decade transition from on-premise to cloud. The business now runs at a $117bn annualised revenue rate, and while growth has moderated, Amazon emphasised that customer demand for both foundational cloud services and generative AI remains strong. The AI segment is already a multibillion-dollar run-rate business growing at triple-digit percentages YoY. Capacity remains a constraint, but more Trainium 2 and next-generation NVIDIA chips are being brought online in the second half of the year. Amazon expects these additions, along with broader workload migrations and falling inference costs, to support continued growth through the remainder of 2025.
I think that the results show that AWS is running as efficiently as it ever has, and the margin improvement deserves credit. But the revenue story is more muted. Sequential growth is slowing, and Amazon hasn’t given the market enough to work with when it comes to understanding how much of AWS’s momentum is being driven by AI. Without clearer signals, it becomes harder to justify AWS as the lead beneficiary of enterprise AI spend. I think the platform is well-positioned, particularly with its silicon strategy and broad model support, but unless that starts to translate into accelerating revenue, the perception of leadership could very well shift elsewhere. Before looking into Azure specifically, let’s see how Microsoft performed all around, given its shares were up ~9% post earnings.
Microsoft
Microsoft delivered a strong set of results. Revenue was $70.1bn, up 13% YoY. Operating income was $32.0bn, up 16% YoY. Net income reached $25.8bn, and diluted earnings per share rose 18% to $3.46. Gross margin improved to 69%, and operating margin expanded by more than 100 basis points. The performance was broad across all segments. Productivity and Business Processes grew 10%, Intelligent Cloud increased 21%, and More Personal Computing was up 6%. Microsoft Cloud revenue grew 20% YoY to $42.4bn. These results show that AI demand is no longer just a headline story. It’s beginning to filter through to the financials and drive both revenue and margin expansion. There’s now a clear link between AI adoption and Microsoft’s commercial performance, with margin improvement and revenue acceleration moving in tandem.
Azure and other cloud services grew 33% YoY, or 35% in constant currency. AI services contributed 7 percentage points to that figure, in line with the previous quarter. Microsoft has now broken out AI’s contribution to Azure growth for three quarters in a row, giving investors clear visibility on how AI adoption is progressing. Management stated that more than 65% of the Fortune 500 are using Azure’s AI capabilities in some form, including Azure OpenAI Service and custom model deployments. This signals that enterprise demand isn’t just exploratory. Workloads are moving into production.
What differentiates Microsoft is how tightly Azure is integrated with the rest of the product stack. AI demand runs through Office, GitHub, Dynamics and Copilot. That means growth in Azure is tied not only to infrastructure spend, but also to user engagement and broader enterprise activity. The 7 percentage point AI contribution isn’t just about new capacity. It reflects usage that is being monetised across multiple channels. So, although AWS remains the most profitable cloud platform, the momentum seems to slowly be shifting. Azure is narrowing the gap, and is doing so with greater consistency, stronger integration, and clearer attribution to AI-driven demand.
Price Action
MSFT 0.00%↑ Can be pretty range-bound but when it moves, it really moves. After a massive push higher, price isn’t too far off ATH’s which sounds crazy given everything that has been going on. But like I said in my piece on tariffs and tech, you can’t put tariffs on software and data. I still stand by that, tech firms only sold off rapidly due to valuations, not because the future was bleak. I wouldn’t be surprised if price consolidates in this region above $420. And if there’s any good news for the markets in the coming weeks, I can certainly see Microsoft trading near/at ATHs!
Outlook
Management guided Azure growth to remain strong in Q4, expecting 34% to 35% in constant currency. AI continues to be the main driver, but underlying demand across non-AI services remains resilient. Capex will increase again next quarter as Microsoft adds more capacity, particularly for AI infrastructure, with management noting that in some regions, demand is already outpacing availability. Beyond Azure, the broader business remains well supported. Office commercial, Dynamics and LinkedIn are expected to sustain mid to high single-digit growth, while Windows and Surface face more modest trends. The commercial bookings backlog rose 19% YoY and total remaining performance obligations now stand at $215bn, up 18% YoY. This reflects a healthy demand environment across the enterprise base. Management noted that AI adoption is becoming more embedded in customer strategy, not just in trials but in long-term commitments. Overall, Microsoft is entering the final quarter of its fiscal year with a strong pipeline, deepening product integration, and confidence that investment in infrastructure and software will continue to compound across segments.
So it’s fair to say, in my opinion of course, that the future looks much brighter for Microsoft than it does for Amazon. Of course, Amazon isn’t solely a tech business so I’m not comparing apples to apples here. But on the cloud services front, I wouldn’t be surprised if Azure catches up with AWS in terms of revenue, within the next couple of quarters.
Meta
To have a clearer understanding of Meta’s business, please check out my previous earnings article covering them. Since that post, Meta has hit a peak drawdown of ~-30% but has since recovered strongly. I really like Meta as a business. They sit at the edge of innovation and have a highly profitable business. However, the current ruling, which threatens to break up its family of apps, is somewhat worrying, along with the recent sanction that the EC imposed. For context, the European Commission has ruled that Meta’s ad-free subscription model violates the Digital Markets Act. Specifically, the EC found that users must be given a true choice between a personalised and a non-personalised experience, and that Meta’s current implementation doesn’t offer a compliant opt-out. The company has said it will appeal the ruling but it also acknowledged that changes to the user experience in Europe may be required before the appeals process concludes. These changes could materially weaken user engagement and reduce advertising revenue in the region. In practical terms, this introduces uncertainty into Meta’s European performance as soon as Q3. While the core model remains intact elsewhere, this is a reminder that regulatory pressure isn’t easing and could lead to structural changes in monetisation strategy.
Turning to the earnings, the results were strong across the board. Revenue grew 16% YoY to $42.3bn, or 19% in constant currency. Net income rose 35% to $16.6bn, and diluted EPS increased 37% to $6.43. Operating income reached $17.6bn with an operating margin of 41%, up from 38% a year earlier. Free cash flow was $10.3bn. Advertising continues to drive the bulk of the business, with ad revenue up 16% YoY to $41.4bn. This was powered by a 10% increase in average price per ad and a 5% increase in ad impressions. Online commerce was the strongest vertical and ad spend grew across all major geographies. Rest of World and North America led at 19% and 18% respectively.
Engagement metrics also trended well. Family daily active people reached 3.43bn, up 6% YoY. Time spent on Facebook rose 7%, Instagram was up 6%, and Threads saw a 35% increase. Threads now has over 350 million monthly actives. The uplift is largely driven by ongoing improvements in recommendation systems and new features designed to increase social interaction around content. Meta also rolled out Blend, a tool that merges algorithmic preferences between users to spark engagement in direct messages, and saw early success with its AI-enhanced Edits app for creators.
The ad platform continues to evolve. Meta is now deploying its Generative Ads Recommendation Model, or GEM, trained on thousands of GPUs, across more services. Initial deployment on Facebook Reels delivered up to a 5% lift in conversions. Advertisers are also using Advantage+ creative tools at scale. Meta has broadened access to video expansion and image generation features and is now testing virtual try-on for fashion ads. These tools are increasing the adoption of automated campaign setups and pushing conversion performance higher. Incremental attribution, which aims to track conversions that wouldn’t have happened without an ad, is seeing promising results, with an average 46% lift compared to business-as-usual campaigns.
Business messaging and monetisation outside the core feed also showed momentum. WhatsApp Business and Meta Verified subscriptions helped drive a 34% increase in Family of Apps “other” revenue, which reached $510mn. Messaging in general is becoming a bigger focus. Management is testing AI-powered business agents that can assist customers across WhatsApp, Messenger, Instagram and Facebook ads. Meta is also expanding monetisation on Threads, although contribution this year is expected to remain immaterial.
On the investment side, capex came in at $13.7bn, nearly double YoY. Most of this is going into infrastructure, particularly AI compute and data centres. Despite this spend, operating leverage remains strong. The Family of Apps posted a 52% operating margin, with expenses growing slower than revenue. Reality Labs reported a $4.2bn loss, as expected, with Quest hardware soft but Ray-Ban Meta AI glasses showing traction. Sales have tripled YoY, and user engagement is improving. Meta’s broader AI efforts are also scaling. Meta AI now has nearly one billion active users, and a standalone app was launched during the quarter. While monetisation isn’t a near-term priority, management believes the long-term potential is substantial.
Price Action
Meta's price has recovered pretty well, as mentioned. It has closed above its 200-day moving average and I can see momentum carrying on into the $635 level. Just to be clear, I don’t make investment decisions based on moving averages but the 200-day MA helps paint the picture of where sentiment stands.
Outlook
Meta expects Q2 revenue between $42.5bn and $45.5bn, implying another quarter of mid-teens growth at the midpoint. Management noted strength in core engagement and continued traction from AI-driven ad tools, though flagged some softness in Asia-based e-commerce advertising. The full-year capex range was raised once again to $64–72bn for the full year. This is positive news amidst the macro uncertainty, reinforcing the company’s conviction in scaling its infrastructure footprint. While regulatory changes in Europe may create volatility later in the year, the broader tone remained constructive, with management positioning Meta as well-placed to compound through both product innovation and disciplined investment.
Final Thoughts
It has been a relatively strong quarter for cloud and AI-focused businesses, but there is a clear divergence in positioning. Microsoft continues to set the pace, with Azure growing faster than peers and a clear narrative around how AI is being monetised across its ecosystem. AWS remains highly profitable and operationally disciplined, but the growth story lacks clarity and the AI contribution is still difficult to quantify. Meta delivered solid results and continues to scale its infrastructure spend aggressively, but there is more uncertainty here. Regulatory pressure in Europe, persistent Reality Labs losses, and limited visibility on monetisation from emerging platforms all cloud the picture. The numbers are strong, but the margin for error feels narrower. AI is clearly no longer just a theme. It is now a force that is reshaping product strategy, capital allocation and revenue composition at the very top of the tech stack. Execution and visibility are what will separate the leaders from the rest. I do have full confidence that the big players will continue to deliver better than expected results.
Thank you for reading! I hope you found this post insightful. If you know anyone who may be interested in the topics I’ve covered, please share this article as it supports my work and helps the newsletter grow.
Disclaimer: This article is for informational purposes only and should not be considered financial or investment advice. The views expressed are my own and based on publicly available information, market trends, and personal analysis. I own shares Meta shares which may change at any time. Readers should conduct their own research and consult a financial professional before making any investment decisions.