Meta, Amazon & Microsoft Q2: Big Tech Delivers, But Unevenly
This article covers how Meta soared, Amazon wobbled, and Microsoft quietly asserted dominance and what it all means for AI, margins and portfolios.
Markets were down last week for a few reasons. Firstly, while the latest tariffs weren’t on the scale of April’s Liberation Day, they’ve served as a reminder that elevated trade barriers aren’t going away anytime soon. Along with that, softer payroll numbers came out of the US, suggesting a slowing labour market, with recent releases seeing large revisions. The revisions led to Trump firing the lead official on economic data. It was an interesting week for earnings, with Meta doing really well, but Amazon not doing great. For me, although I’m a bull and love tech, I’m struggling to justify how all this capex will translate to revenue higher than the capex being spent. At the same time, I don’t think the hyperscalers have a clear idea either, and they will figure it out as they go along. Which isn’t necessarily a bad thing. In this article, I’ll be covering Meta, Amazon and Microsoft earnings to see how the landscape is progressing. Make sure to check out the Q1 article on these three if you haven’t already.
Meta
Meta man Zuck seems to be pulling off his AI plans so far. Of course, not by himself. But it was another blowout quarter for Meta. Since the previous quarter, a lot has happened with Meta, but the most interesting headline, in my opinion, was the $100mn+ sign-on bonuses that Zuckerberg proposed to prospective new members of his AI superteam. Now, of course, I'm sure it isn't just $100m up front, and they receive the cash the next day, there will almost certainly be conditions to unlock this bonus. Away from that, Meta acquired Scale AI, the data labelling startup, for $14bn. This deal hasn’t officially closed yet, but the number is widely reported, and if accurate, it would be one of the largest AI-related acquisitions to date. Scale has been a key player in providing high-quality labelled data for training LLMs, and Meta sees this as a way to deepen its control over the entire AI stack from data to infrastructure to models.
Financials
Revenues remain strong with Q2 revenue of $47.5bn, up 21.6% YoY and 12.3% QoQ. This came in slightly ahead of expectations and was driven by strong advertiser demand, continued improvement in pricing, and higher monetisation across newer surfaces. Advertising revenue came in at $46.5bn, up 21% YoY. The ad business is clearly seeing ROI on the CapEx being poured into AI. My slight concern is at what point does ad growth slow down? This is now the second quarter in a row where pricing has driven the bulk of ad revenue growth. That’s not a problem in itself, but it suggests impressions are starting to flatten. Meta is getting more money out of each ad, but there's only so much you can keep squeezing out of the same surfaces. If Reels and messaging formats are already monetising better, where does the next leg of ad load come from? Maybe AI agents and click-to-message drive another wave, but that hasn’t broken out yet in a big way. Perhaps a new avenue will be explored next year, but key figures like revenue will have high starting points for YoY growth.
However, for now, margins remain healthy, with operating income at $20.4bn and group operating margin holding steady at 43%. Family of Apps delivered $23.1bn in operating income with a 50% margin, while Reality Labs posted a $4.5bn loss on $370mn of revenue. Meta has so far managed to absorb higher infrastructure spend without margin compression, but with depreciation stepping up next year and investment levels still rising, maintaining this level of profitability will require even more output from the same stack, either through new ad formats or greater efficiency.
Impressions across all platforms rose 11% YoY, while average price per ad increased 9%. This was a more balanced mix than previous quarters, where pricing did most of the work. Management attributed the gains to stronger demand and continued improvements in AI-driven delivery. No specific region was called out this time, suggesting growth was broad-based rather than concentrated.
Free cash flow came in at $8.5bn for the quarter, down from $12.5bn in Q1. The drop was entirely due to a sharp rise in CapEx, which hit $17bn this quarter alone. Management reiterated full-year guidance of $35–40bn, meaning over 40% of that has already been deployed. Despite the step down, trailing twelve-month FCF still stands at $46.6bn. The business remains highly cash generative, but with infrastructure spending accelerating and Reality Labs still deeply negative, maintaining this pace will require the ad engine to continue firing and AI investments to continue feeding back into revenue. Meta also repurchased $9.8bn worth of stock during the quarter and paid out $1.3bn in dividends. With a fresh $50bn authorisation added, bringing total buyback firepower to $61.1bn, capital returns remain firmly in focus. Management explicitly stated that repurchases are being used to offset equity dilution tied to AI hiring and stock-based comp, while also maintaining regular dividend distributions.
Management reiterated that the bulk of investment is going into AI infrastructure, including data centres and server buildout to support training and inference. While no new updates were given on MTIA v2 this quarter, it remains central to Meta’s long-term strategy to reduce reliance on third-party silicon and scale internal compute capacity. If you’ve never heard of MTIA v2, don’t worry. It’s the second-generation version of Meta’s in-house AI inference chip, short for Meta Training and Inference Accelerator. While v1 is already deployed in production for on-platform inference, v2 is expected to be faster, more power-efficient, and optimised for running Meta’s large language models like Llama directly in its own data centres. The goal is to reduce dependence on Nvidia and other third-party chips over time. As a reminder, inference is the next step from training a LLM and it’s all about how the model computes and gives an output based on unseen data.
Price Action
Shares rallied ~10% after results, but have given up some gains. I think the $700 area could make sense for another entry, but to be honest, I’m a bit wary of adding more at this stage simply because I can’t see a clear path as to how the CapEx will be justified in terms of ad revenue in the next few years. There is a lot of hype and I do believe in AI as a technology but I feel like growth may slow down next year, which will disappoint investors. I’m not saying that I’m selling all my shares at this stage, but more so that I’m being wary of adding more size at high levels. The valuation Meta trades at now isn’t crazy though, at ~28x P/E, so in terms of valuations, Meta isn’t overvalued at all.
Outlook
The setup into the second half looks solid. Revenue growth is still running above 20%, margins are holding despite record CapEx, and capital returns remain aggressive. Management reiterated that the majority of investment is going into AI infrastructure, with MTIA and custom-built data centres central to future product delivery. Susan Li noted that compute remains a constraint and Meta is building for long-term flexibility. On the revenue side, management called out stronger demand for messaging ads, with click-to-message continuing to scale, and early signs of monetisation improvement in Reels. But they were also clear that the benefits from AI will take time to fully materialise. Llama 3 is already integrated across platforms, but attention is now shifting to the next model iteration and what it can unlock, especially as Llama 3 was deemed somewhat of a flop. With depreciation stepping up and Reality Labs still deeply negative, the path forward depends on AI investments translating into tangible revenue growth.
Amazon
Things weren’t as great as Meta for Amazon. Revenue was still solid, AWS growth ticked up, and margins held. But there wasn’t the same sense of acceleration. E-commerce is steady, and AWS continues to improve. But the market wanted more, especially with the stock already near all-time highs.
Financials
Before diving into AWS, it's worth reviewing the broader business. Amazon reported Q2 revenue of $167.7bn, up 13% YoY and 11% in constant currency. Operating income rose to $19.2bn, up 31% from $14.7bn a year ago, driven by efficiency gains across both the retail and cloud segments. Net income reached $19.17bn compared to $13.45bn a year ago, and included a $1.2bn non-operating gain from Amazon’s Rivian stake. While these non-core items added a boost, the underlying profitability trends remain strong. Advertising revenue rose 22% YoY to $15.69bn, maintaining its position as one of Amazon’s highest-margin growth engines. Subscription services, which include Prime and digital content, reached $12.2bn, up 11% YoY. Online store revenue came in at $61.48, up 10%, which is impressive given that the Online Store segment is the largest revenue contributor by far. These segments of the business remain solid and are growing at stable double-digit rates. Now, for the most exciting part of Amazon’s business (in my opinion). AWS revenue came in at $30.87bn, a 17% YoY increase. This is great right? Well, the market didn’t think so.
Operating income retracted sequentially, falling from $11.55bn in Q1 to $10.16bn in Q2, while the operating margin dipped from 39.5% to 32.9%. It’s not a dramatic drop, but it comes at a time when investors were hoping to see the opposite, margin expansion. AWS is supposed to be the crown jewel, with the highest margin profile across Amazon’s business lines. So any slip, even modest, raises questions. Management pointed to rising energy costs and continued investment in AI infrastructure as the main drivers, but that explanation isn’t enough to offset investor disappointment. The narrative has been that all this investment would soon translate into stronger demand and higher monetisation. Instead, the growth is stalling while costs are rising, which undermines confidence in the payback timeline. Of course, it’s only one quarter, so it’s not the end of the world, but if this becomes a pattern going forward, things could get really ugly for Amazon.
That tension is most obvious when looking at CapEx and cash flow. CapEx surged to $31.4bn in Q2, a significant jump that Amazon explicitly linked to AWS buildout. Management confirmed the bulk of this is going into chips, data centres, and power infrastructure to support the generative AI opportunity. But that’s precisely what raises the stakes. When you deploy over $30bn in a quarter, the market expects more than flat sequential growth and declining margins. Free cash flow for the trailing twelve months came in at $18.29bn, down from $53bn a year ago, reflecting the investment-heavy cycle. The spending might be justified over time, but for now, the cash burn is real and the revenue payoff hasn’t materialised. Investors are watching closely to see when, or if, this AI investment starts to deliver a meaningful return.
I don’t have Amazon in my portfolio, and it’s because AWS is already so big, I don’t see materially high growth rates for it going forward. And I’m not interested in the other parts of Amazon’s business. AWS has also lost market share, but even if it were to get its market share back, the overall share price would still depend on other business segments, too. Although that’s my view, it’s still important to follow how AWS is doing, as it helps paint the picture of the cloud space.
Price Action
Following the earnings numbers, AMZN 0.00%↑ fell by ~8.8%. A contrasting drop compared to Meta. If the lull continues, I can see the price drifting down into $200. As I’ve mentioned, I have no interest in adding Amazon to my portfolio, so it’s not a chart I’ve been following, but if I were to be a bull, the $200 mark would seem like a reasonable place to enter/add.
Outlook
The outlook from management remains optimistic but cautious. They reiterated confidence in AWS’s long-term opportunity, particularly around generative AI, and signalled that investment will remain elevated as they scale out infrastructure. However, they also acknowledged that margins will remain volatile in the near term as depreciation catches up with CapEx. There was no change to guidance, which suggests Amazon is sticking to its view that this is a necessary build phase. The broader retail business looks stable, and advertising continues to grow at a healthy clip. But the story here is AWS and the next couple of quarters will need to show clear progress in monetising AI workloads. If that doesn’t happen, the disconnect between investment and return will become harder to ignore.
Microsoft
Microsoft closed out its fiscal year with another strong quarter, beating across the board and reaffirming its position as the most broadly leveraged AI beneficiary in big tech. Let’s see where things went well.
Financials
Microsoft wrapped up its fiscal year with Q4 revenue of $76.7bn, up 18% YoY. Operating income came in at $34.4bn, up 23% YoY, and net income reached $27.2bn, an increase of 24% YoY. Diluted EPS was $3.69, up 23% from $3.00 a year ago, comfortably ahead of expectations. All three segments contributed meaningfully to growth, but the headline once again came from Microsoft Cloud, which delivered $46.7bn in revenue, up 27% YoY. That now makes up over 60% of the company’s total revenue base.
Breaking it down properly, Productivity and Business Processes revenue was $33.1bn, up 16% YoY. Intelligent Cloud revenue came in at $29.9bn, up 26%, with Azure and other cloud services growing 27% YoY or 30% in constant currency. More Personal Computing generated $13.5bn, up 9%, helped by a rebound in Windows OEM (+10%) and strength in Xbox content and services (+61%) from the Activision Blizzard acquisition. Surface remained a drag, with revenue down 17%, while Search and News advertising revenue rose 3%.
Margins remain strong across the board. Gross margin expanded by 1pt YoY to 70%, while operating margin increased 2pts to 45%. CapEx hit $14bn in the quarter, up sharply from $11.5bn last quarter, reflecting continued investment in AI infrastructure. Free cash flow for the quarter was $22.7bn, up 31% YoY. Microsoft also returned $9.0bn to shareholders via buybacks and dividends, maintaining its disciplined capital return strategy. Despite elevated AI spending, profitability is not just holding, it’s expanding. Microsoft continues to deliver both operating leverage and growth at scale, with cloud and AI monetisation firmly in motion.
What’s striking is how broad-based this is. Azure gets the headlines, but everything’s working. Office is compounding. LinkedIn is growing steadily despite a soft hiring backdrop. Even Windows bounced, which no one was really expecting. Gaming’s contribution looks big, but that’s mostly Activision. The real story is that Microsoft is squeezing more revenue from every corner of the portfolio without breaking the margin model. They’re executing with discipline and showing clear operating leverage, even while ramping AI investment. There’s very little fat in this print, and if anything, the mix keeps improving. It's hard to poke holes in this quarter. And that’s why, if an investor wants AI exposure without betting on a single product or vertical, Microsoft is the cleanest way to get it. It’s not just about Azure; it’s AI showing up across Office, GitHub, Copilot, and even infrastructure. This breadth gives Microsoft a structural advantage in turning hype into recurring revenue, without relying on any one breakout use case.
Price Action
Microsoft shares have performed exceptionally well. So well that it’s hard to really point out an entry point, as Microsoft doesn’t typically surge outside earnings prints. It generally just grinds higher, or sideways, slowly but surely. And as such, I don’t feel the need to highlight a specific level to enter. The business is very well positioned going forward, and even if I were to enter now, if the current trajectory continues, I don’t see why ~$580-$600 wouldn’t be hit by the end of this year.
Outlook
Looking ahead, the setup is favourable. Management guided to $66bn–$67bn in cloud revenue for Q1 (this will be Q3 of 2025, don’t let the fiscal years confuse you), implying steady momentum, especially in Azure. CapEx will continue rising, but Microsoft has shown it can absorb that without compromising margins. AI services are expected to scale further across the stack, with Copilot adoption likely to pick up in Office and GitHub. The big question is whether Azure can reaccelerate or if this ~30% pace is the new baseline. Either way, as long as margins hold and AI monetisation keeps broadening, Microsoft looks well-positioned to keep delivering. There’s execution risk, but it’s hard to find a better risk-adjusted AI compounder in the market right now.
Final Thoughts
Big tech is still delivering, but not all at the same pace, and not with equal clarity. Meta is executing well and keeping margins high while pushing hard into AI infrastructure. Microsoft is showing that you can invest aggressively in AI and still grow profitably across every major line. Amazon, on the other hand, is still in the heavy investment phase, and while AWS remains critical, the lack of margin progression is starting to raise questions. The broader takeaway is that AI is real, the opportunity is massive, but monetisation remains uneven. CapEx is soaring everywhere, and the market is willing to wait for now. But patience has limits. If we don’t see a step-change in revenue or product rollout over the next couple of quarters, the mood could shift quickly. Until then, quality and execution will matter more than raw ambition.
Thank you for reading. I hope you found this article insightful. For more articles like this, subscribe (please)! And if you think any friends/family/colleagues would find this article interesting, please share it with them.
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 both META shares, which may change at any time. Readers should conduct their own research and consult a financial professional before making any investment decisions.