The investment case for small caps over large caps is based on two attributes: they grow faster and are more undervalued. While this is true for the majority, not all large companies are created equally, at least not big tech.
As a theme and ecosystem, I follow the big technology companies closely and so far, I’ve published deep dives on two of them, Alphabet and TSMC. In this article, I’ll analyse the broader big tech earnings, and I’ll cover:
A snapshot of big tech’s H1 2025 earnings
An overview of their big techs sorted by divisions
Analyse their earnings centred on 10 key topics:
Digital advertising
SAP’s turnaround
Tesla. Cars or robots?
Semiconductors (TSMC and ASML)
Microsoft’s expanding moat
Big tech’s entertainment market
Cloud computing
Apple’s Services division dependence
Costs and Capex
Valuations
Charlie Munger’s advice - The big tech case
In one of Charlie Munger’s last interviews, he laid out the investment case for these big techs:
“What everybody has learned is that everybody needs some significant participation in the 12 companies that do better than everybody else. You need two or three of them, at least” - Charlie Munger.
While the quote doesn’t mention “big tech”, he proceeded to discuss Berkshire Hathaway’s investment in Apple and why these big companies get bigger.
Traditional value investing has broadly shied away from tech and large caps, but upon reflecting on Munger's statement and my own studies into these companies, I fear this has been a mistake when put into context.
Compared to the 1970s - 2000s, the largest companies were then dominated by companies with depleting assets, oil and auto companies (Exxon Mobil, General Motors, General Electric and Ford).
Unlike them, big tech's core asset, data, becomes more useful and valuable the more it's used by consumers, which creates a natural compounding machine over time, driven by technological advancements and further innovations. This is why initial tech leaders like Microsoft, at nearly $4 trillion market cap and $280 billion in revenue, can still grow above 15% per year.
I. A snapshot of Big Tech’s H1 2025
The table above highlights our big tech H1 2025 revenue, operating earnings and margins. For comparative purposes, I replaced three big techs - Nvidia, Broadcom, and Oracle with SAP, Netflix, and ASML due to their different fiscal years, as they haven't yet reported their full H1 2025 results.
Also included are their current forward P/E and three other proprietary metrics; the price I'm willing to pay (Jenga IP P/E), how I rate their business quality - moats, barriers to entry, test of time (Jenga IP Quality score out of 100) and how I view the current sentiment on their stock (Jenga IP sentiment analysis). In most cases, I prefer to purchase shares below my earnings multiple, as seen in my recent purchases of Alphabet and TSMC shares (highlighted in green).
Growth: On average, the big tech companies grew their revenue by 19% and EBIT by 29.7% during the first half of 2025, certainly exceeding the majority of small caps.
Profitability: The big tech companies achieved an average operating profit margin (EBIT margin) of 31.3%, well ahead of the broader market.
Jenga IP Quality: All ten companies have Jenga IP high moat scores (colour coded in green), ranging from Tesla’s 75.6 to Microsoft’s 90.7.
Sentiment: The market remains quite optimistic on all ten companies, with all excluding Alphabet’s sentiment score below 6 (colour coded in red).
II. Analysis of company segments
Every six months, I track the revenue performance of the major divisions across big tech, and the table below presents their segment's revenue from Q1 2021 to Q2 2025, grouped by:
Digital advertising (Amazon advertising, Meta platforms family of apps, Google search, YouTube ads, Google Network and Microsoft search & news)
Digital entertainment (Amazon subscriptions, Netflix revenue and Google subscriptions, platforms & devices)
Amazon commerce (online stores, physical stores and third-party sellers)
Cloud (Amazon Web Services, Microsoft Intelligent Cloud, Google Cloud Services)
Apple (iPhone, Mac, iPad, Wearables/home/accessories and services)
Tesla (automotive sales and energy generation & storage)
Other segments (Amazon other, Meta other, Meta reality labs, Google other)
Jenga IP Big Tech Sheet. Source: Company filings.
The big tech sheet examines the respective companies sorted by their business segments; digital advertising, digital entertainment, Amazon commerce, Cloud, Apple products, Tesla and other segments.
You can access the Big Tech spreadsheet here.
Examining their results with this format helps put their performance in perspective:
How are YouTube ads performing relative to Netflix? Slower.
Are Amazon's subscription services keeping up with Apple's services division? Both have an 11% CAGR.
How many years behind is Google Cloud Division when compared to Amazon's AWS? 4 years.
While it would be fun to go through each of these numbers, it would be more productive to highlight some stats I found surprising from the table:
Stats from the big tech sheet
Amazon Advertising is rapidly catching up to peers in the digital ad market, compounding at 20.5% annually over 4 years, well ahead of Google Search (10.9% CAGR) and Meta Platforms (13% CAGR)
YouTube Ads 4-year 8.8% CAGR trails Google Search 10.9% CAGR
Google’s cloud services (31% CAGR 4-year CAGR) revenue trajectory is roughly Amazon’s AWS T-4 years (20.2% 4-year CAGR)
Amazon Physical Stores (7.4% CAGR) are currently growing twice as fast as Amazon’s online stores (3.7% CAGR)
Tesla Automotive sales have shrunk by 4.9% per year (3-year CAGR), while Tesla’s total revenue has grown only by 9.9% per year, slower than all other big tech companies, excluding Apple.
Apple’s iPad, Mac, and Wearables/Home/Accessories divisions have all contracted over the past 4 years.
Amazon Subscriptions remain consistently larger than Netflix revenue and have grown slightly faster over 4 years (11.4% vs. 10.8% CAGR)
Fastest-growing major divisions (> $5B quarterly revenue, 4-year CAGR):
Google Cloud Services 31%
Amazon Advertising Services 20.5%
Amazon Web Services 20.2%
Slowest-growing major divisions (> $5B quarterly revenue, 4-year CAGR):
Apple Wearables, Home & Accessories -4.2%
Apple iPad -2.8%
Google Network -0.8%
III. Ten big tech discussion topics
Digital advertising
Advertising is quite crucial for big technology. At least 27% of all big tech revenue comes from digital advertising, and given its faster growth, it's likely its share will increase in the mid-term. Among the seven (excluding SAP, TSMC and ASML), advertising is either a core part of their business model (Alphabet and Meta Platforms) or part of their growth strategy (Netflix and Amazon).
The chart below presents the digital advertising market structure across six major big tech properties, led by Google search.
Amazon has been the clear winner in the digital advertising revenue over the past four years, growing its share from 8.1% to 11.4% over the past four years while achieving an incremental $8.2 billion in quarterly revenue between Q2 2021 and Q2 2025. While advertising revenue took a back seat during the earnings call, it's easy to underestimate its role in the overall Amazon group.
Over the past 1, 3 and 4 years, Amazon's quarterly advertising revenue has outpaced its AWS Q2 revenue growth rate. Amazon's advertising strategy is multifaceted, but an increasingly important area is the integration with streaming apps (Roku, Disney, Prime Video), which certainly increases its direct competition with both Netflix and YouTube.
Moving to the big two advertising companies, Alphabet and Meta, both companies remain resilient despite several competitive forces disrupting their business models. Between both, Meta is certainly coping far better. Its family of apps' EBIT margins reached a recent Q2 high of 53% (see chart below) with revenue accelerating by 21.5% YoY during the quarter. How do apps with 3.48 billion daily active users grow?
Artificial Intelligence. An under-discussed area of AI is its applications in core advertising and social media business models. Both Google and Meta discussed this and the value creation experienced in recent months:
"Over 2 million advertisers now use Google's AI-powered asset generation tools to run ads, a 50% increase on this time last year" - Philipp Schindler, Chief Business Officer, Google.
"Our new AI-powered recommendation model for ads to new surfaces and improved its performance by using more signals and longer context. It's driven roughly 5% more ad conversions on Instagram and 3% on Facebook." - Mark Zuckerberg, CEO, Meta.
Finally, keeping track of the reported Google and Meta KPIs serves as a reality check when assessing the impact of LLMs like ChatGPT on their core business models.
While 4% is certainly at the lower end for Google's historic paid clicks change, it's impressive that they've managed to achieve this in the face of the various challenges. Moreover, Meta's more impressive impressions growth shows its business model is more protected from the ongoing challenges in AI.
SAP’s turnaround
At a current market cap of $315 billion, SAP is currently ranked as the 31st largest publicly listed company, and thus deserving a place among the big tech peers, which likely comes as a surprise to most readers like myself.
Five years ago, many predicted SAP’s downfall due to its sluggish adoption of cloud software. Since then, its fortunes have reversed, and its share price performance highlights the potential of investing in turnarounds. Its share price return of 170% outperformed all but one (Meta Platforms) of the other big tech companies over the same period.
There are a few things I liked from SAP’s H1 update, but the two key areas are its efforts in cloud solutions and the resulting margin expansion generated from this focus.
Q2 2025 marked the 14th quarter in which SAP's cloud ERP suite grew by over 30% per year, while its broader cloud revenue increased by 28% during the quarter. There were several customer wins across its different operations and resources, ranging from GSK, Balmain, Alibaba, and the German Armed Forces.
This transition to cloud has incremental gains on margins, and since Q2 2021, SAP’s EBIT margins have increased from 17% to 28.5% in its latest quarter. Management expects this trajectory to continue in the coming quarters and, in its CFO’s words:
“The increase in total expenses versus the increase in revenues will be contained in the range of 80% to 90%. And that is the kind of yardstick for coming years.” - Dominik Adam, SAP CFO.
In terms of controlling costs, its CFO provided SAP’s priorities, starting with selling expenses, followed by R&D and then G&A for the coming quarters and from my view, these do make sense given its current capital allocation framework.
While SAP's 40x earnings multiple suggests the market is fully valuing its turnaround efforts, the company's consistent execution and disciplined cost control make it a rare turnaround story that still has room to run. I would still watch this story closely, given the opportunity the European market often presents during volatile macroeconomic periods.
Tesla. Auto or energy?
Among all the big tech companies, Tesla is the one I struggle with. I’m neither a bull nor a bear and prefer to watch how things unfold from the sidelines. On one hand, it's solved a near-impossible mission, led by a once-in-a-generation visionary, but on the other hand, its core product, cars, doesn’t enjoy the tech data benefits as with Microsoft or Alphabet.
From the quarterly revenue chart below, you can see the revenue volatility. Despite its rich multiple at 11x revenue, Tesla has barely grown its automotive sales, excluding regulatory credits, in the past three years.
From a profitability lens, Tesla’s energy generation and storage division ($846 million) earned nearly as much as its auto gross profits excluding regulatory credits ($888 million) despite only being 18% of the auto division's revenue.
This is a stark contrast from where Tesla was three years ago, when its auto division was experiencing operating leverage.
So is Tesla today, an automotive car company? Or an energy company?
It's both, but certainly requires more for the future, to truly cement a moat. It’s clear Elon Musk’s priorities are increasingly shifting to its autonomous ride-hailing and general-purpose robots divisions. Both its CFO, Vaibhav Taneja, and Elon Musk spent significantly more time discussing its robots and autonomous ride-hailing efforts.
During the first half of the year, its robotaxi services were rolled out in Austin, and they plan to have the robotaxi service in roughly half of the U.S. population. A Tesla car was also delivered to a customer completely autonomously, highlighting a use case for automation and its Optimus bots:
“Optimus 3 is an exquisite design, in my opinion, and will be, as I’ve said many times before, I predict will be the biggest product ever” - Elon Musk, CEO, Tesla.
“We are going to scale Optimus production as fast as humanly possible…. We’ll try to get to 1 million units a year… We think we can get there in 5 years.” Elon Musk.
If Tesla does achieve Musk’s goal of 1 million units in 5 years, that’s roughly $30 billion in additional revenue, meaning it's a division one shouldn’t take lightly.
Rarely do I hear companies mention competitors directly during earnings calls, but as we know, Musk isn’t the norm and he left an interesting commentary on the Tesla versus Waymo.
“If you compare, say, Tesla, to Waymo, the car is festooned with God knows how many sensors. And yet, isn’t Google good at AI? Yes, but they’re not good at real-world AI.” Elon Musk.
Semiconductors (ASML and TSMC)
Both TSMC and ASML provided updates on their fundamentals and how they continue to play an instrumental role in AI alongside broader technological innovations. I was impressed by the companies as they both grew revenue and operating profits quite well. However, where they diverged was in the forward guidance, and here, TSMC is more optimistic and certain about growth than ASML.
There are a few reasons behind the diverging results. First, ASML’s exposure to China has become highly political. Over the past two years, China has grown from 14% to 36% as a share of ASML’s revenue, significantly higher than Europe and North America’s 20% share, which puts ASML under intense scrutiny.
Second, the transition from 3nm to 2nm chips has created some headwinds. Several industry experts reported that the 2nm chips don’t require as much EUV layer growth as expected, which, in effect, reduces the incremental growth for ASML’s tools in the near term.
These challenges are beyond ASML’s control, and if we look closer at the factors within its control, ASML’s innovation continues to look as strong as ever. It delivered the first of the next-gen High-NA EUV system to Intel during the quarter:
“As a reminder, the EXE:5200B system is capable of achieving at least 175 wafers per hour, which is approximately a 60% productivity improvement compared to the EXE:5000” - Christophe Fouquet, ASML CEO.
Source: ASML. A High-NA EUV System
Overall, its long term 2030 revenue guidance of €44-60 billion (9.7% - 15.5% CAGR) with a gross margin expansion from its current 51.3% to 56 - 60% by 2030, puts ASML at potentially 13x its 2030 net profits. With ASML’s virtual monopoly in EUV, an exit multiple of 24x doesn’t seem expensive, indicating a potential upside of 11.4% per share through the decade.
Admittedly, I'm more excited about TSMC's potential, though, and that was reconfirmed with its Q2 results. Unlike ASML, TSMC doesn't have a virtual monopoly, but it continues to outperform its peers across different KPIs, at both the product service and fundamentals lens.
Its CEO, C.C. Wei, reconfirmed revenue targets and discussed additional investments in the US, now at $165 billion; 6 advanced wafer manufacturing fabs in Arizona, 2 advanced packaging fabs and 1 major R&D centre, which will support its U.S. customers, notably Apple, who are already receiving N4 chips from the first Arizona fab.
With its 3nm chips already at 24% of its total revenue, my focus for the second half is seeing further 3nm share of revenue gains, the 2nm enrolment and updates on its efforts in advanced packaging.
5. Microsoft’s expanding moat
In my view, Microsoft is undoubtedly the highest moat big tech company, and impressively, it continues to find multiple ways of deepening its moat and its societal embeddedness. Markets had questioned whether growth was still attainable post its cloud transition, but in Microsoft fashion, it continued to deliver, growing its revenue by 15% and operating profits by 17% for its 2025 FY (June end).
From its earnings call, I gained some insights into key trends and factors driving this growth, and it seems the key areas are its cloud solutions and tools to support businesses.
While Microsoft doesn’t disclose official revenue for Azure, its CEO, Satya Nadella, revealed that Azure surpassed $75 billion in annual revenue, growing 34% in FY 2025, which suggests it gained more market share from its larger rival, AWS.
Furthermore, its approach to AI continues to be multi-faceted with the goal of delivering tools or services that improve and simplify business productivity. Take Microsoft Fabric, an end-to-end SaaS platform for data and analytics, launched in 2023. Microsoft has quickly grown this into a mission-critical service for its customers.
“It [Microsoft Fabric] continued to gain momentum with revenue up 55% YoY and over 25,000 customers. It’s the fastest-growing database product in our history.” - Satya Nadella, Microsoft CEO.
Central to the AI strategy is Copilot, and Satya Nadella shared some updates here. The Copilot family of apps now have 100 million MAUs. Among the customer wins mentioned during the call, there was a clear bias towards financial institutions; Barclays agreed to roll out Microsoft 365 Copilot across its 100,000+ employees globally, and they announced further deployments with UBS, KPMG and Wells Fargo with over 25,000 seats in Q2 2025.
This signals that one of the more profitable customer segments for its cloud solutions is financial institutions and likely highlights an edge relative to both AWS and Google's cloud services.
In my view, there’s a small number of companies that deserve valuations above 30x and Microsoft is among the handful of these companies. It continues to guide revenue growth of at least 13%. Its EBIT margins of 45% are well ahead of peers, and as we’ll discuss later, its Capex is less intense than peers like Alphabet and Meta.
Admittedly, I’ve been slow to recognise just how deep its moat truly is and will likely pick up some shares in the event of market disruptions.
Big tech’s entertainment market
Entertainment services continue to be a key segment for big tech, and they’ve established their paths to stable, subscription-like revenue in a service that is typically volatile for consumers. Among the ten, Netflix is the only company wholly dependent on entertainment revenue, so it’s more meaningful to examine its execution during the quarter.
Test of time and the post-founder transition are critical factors when assessing moats, and a question many had on Netflix was whether they could survive this. So far, it has and I’m impressed with the bold mentality shift led by its co-CEOs Gregory Peters and Theodore Sarandos to a “never say never” culture:
“I’ve learned to never say never. So I would say we remain open to evolving our consumer-facing model.” - Gregory Peters, Netflix, co-CEO.
This point was in direct response to a question on Netflix’s approach with live sports, gaming and advertising revenue, transition from solely TV shows and movies. In Q2 2025, among the four highlighted entertainment segments in the chart below, Netflix kept its 28% market share, maintaining its revenue with Google’s subscription and platforms division (Play Store, YouTube ads).
Impressively, its growth continued to outperform YouTube ads, although one could argue YouTube's subscriptions are cannibalising its ads, leading to slower growth here. This is still quite impressive given its profit margins on both earnings (34.1% EBIT margin) and cash flow (20.4% FCF margin), despite the continued competition from bigger rivals like Apple, Disney and Amazon.
There’s also some evidence of pricing power. Netflix members watched 95 billion hours in H1 2025, up 1% in volume of hours, but its revenue grew 14%. Its pricing power is supported by its many investments in quality content, such as:
Local-for-local content: Growth in Asia-Pacific continues to outpace USA sales and flagship shows like Squid Games Season 3 (122 million views, already Netflix’s 6th biggest TV series) and movies like K-pop Demon Hunter (80 million views) outperformed expectations.
Partnerships with traditional TV: In France, Netflix partnered with TF1, bringing its local French channels into Netflix.
Award-winning shows: 44 individual shows were nominated for the Emmys with 120 total nominations, far ahead of Apple’s 79, indicating further investments in quality content.
Source: Netflix. The Squid Games series by Netflix.
Over the past four and a half years, its operating margins have increased from 18.3% to 29.5% and the question today is how high can this get in steady state? 30%?, 40%? Or Meta’s Family of apps 53%?
The answer to this determines whether it's worth paying its current forward P/E of 42x, and while I’m very impressed with its delivery, I don’t think its valuation today presents enough margin of safety.
Cloud computing
Across the ten companies, the most surprising worry among investors post earnings release was Amazon, with the AWS segment reporting some growth slowdown. During these moments, it's essential to put these into context:
Taking Nadella's comments on Azure revenue growth for FY 2025, in the table above, I compared Microsoft Azure's reported revenue in FY 2024 and 2025 relative to both Google and Amazon and the table highlights some worry for Amazon. Azure's revenue growth of 34% represents more dollar revenue (est. $19 billion) than AWS $18 billion.
On one hand, I certainly see room for an oligopolistic market structure with all three companies achieving outsized profits, but the key justification for Amazon's earnings multiple premium relative to peers was the market dominance of AWS relative to Azure and GCP.
Its CEO, Andy Jassy, did address this issue:
“I think the second player is about 65% of the size of AWS….when we look at the results over the last number of quarters, there are times we’re growing faster than others and sometimes others are growing faster than us.” - Andy Jassy, Amazon, CEO
Another interesting point is that it seems Amazon's focus has slightly shifted from a complete emphasis on market share to service quality, a strength they continue to highlight over Microsoft Azure:
“If you look at what matters to customers, they care a lot about what the operational performances are, what the availability is, what the durability is, what the latency and throughput is of the various services. And I think we have a pretty significant advantage in that area.”
“The security and the privacy of that data matters a lot, and there are very different results in security in AWS than you’ll see in other players.”
It's unclear who Andy Jassy was referencing in the last statement regarding the security outperformance relative to peers, but it will be an interesting data point to review going forward.
All three companies also discussed recent client wins, new services and how their latest tools are delivering value for customers:
Alphabet: The number of $250 million cloud deals doubled YoY. They signed the same number of $1 billion deals in H1 2025 as in the whole 2024. New GCP customers increased nearly 28% QoQ. The Gemini app now has over 450 million MAUs with 50% daily request growth versus Q1.
Amazon: Signed new agreements with PepsiCo, Airbnb, NASDAQ and LSE. Its custom AI chip, Trainium 2, continues to be the backbone of Anthropic's newest generation Claude models. Released Strands, an open-source way to build agents and now has over 300,000 downloads.
Microsoft: Opened new data centres on all six continents with 400+ data centres in 70 regions. Azure AI Foundry was launched and is seeing traction in helping customers manage their AI applications and agents at scale. 80% of Fortune 500 companies already use Foundry.
In the chart above, I compare AWS, Google Cloud services and Microsoft's Intelligent Cloud divisions. While this isn't a like-for-like comparison, given that both Microsoft and Google include other cloud services beyond their core cloud platform, it does present a snapshot of the current state of the cloud market.
The key edge for Microsoft and Google is their distribution and integration with a wide breadth of other tools and services to support customers, and I wouldn't be surprised to see further share gains in the coming quarters.
Apple’s Services division dependence
Relative to its valuation, I haven't been too impressed with Apple. Over the past 4 years, Mac (-0.6% CAGR), iPad (-2.8%), and its wearables/accessories (-4.2%) divisions have each declined in sales, while iPhone sales have slowed to just 3% per year. The growth in Apple sales predominantly comes from its services division (app store, cloud and revenue from Google), a highly profitable and 12% grower.
Between Q2 2021 and 2025, Apple's Services has grown by $9.9 billion while the rest of its four hardware divisions (iPhone, Mac, iPad and wearables) have grown by just $2 billion, meaning four-fifths of the total dollar revenue growth is services-led.
This isn't necessarily a problem; not all companies need double-digit growth, but at 28x forward earnings, it's difficult to justify this.
That said, there's no question about Apple's deep moat and room for further enablements from AI. Tim Cook and the Apple management continue to view Apple services as the key earnings growth driver and are doubling down on more investments here rather than just the traditional hardware divisions.
Of course, hardware will always be necessary for Apple; they sold their three billionth iPhone since its 2007 launch during the quarter, but the dilemma of "better product quality, longer shelf life" caps the growth potential here.
They launched over 20 Apple Intelligence features, from visual improvements to writing tools
A more personalised Siri will be released next year
Deeper investments in Apple Silicon with $600 billion in committed investments in the US over the next 4 years
Record high iCloud paying accounts across its installed base of active devices
Source: Apple. Apple Intelligence
Its services division now has over 1 billion paid subscriptions across all of its platforms and unlike hardware, these services don't have marginal costs for new users, providing further room for profit growth.
A key challenge for Apple is the current geopolitical and trade war issues, and so far, Tim Cook has navigated this proactively. For the next quarter, management has guided a $1.1 billion hit to costs (7.5% of its Q3 2024 net profits, which is better than initially feared.
Overall, Apple's cash profile, minimal Capex intensity, and well-managed shareholder value creation through its buybacks set it apart from the other big techs. That said, if I could either buy Apple at 28x earnings or Microsoft at 30x earnings, I'd go with the latter—growth across all divisions matters.
Costs and Capex
I've mainly examined these big tech companies from a revenue lens, but an equally important aspect is their costs, and the key risks that impact their investment cases are more projected here.
Over the past few years, big tech, particularly Amazon, Alphabet, Microsoft and Meta, have changed their depreciation accounting, reflecting what they believe to be more accurate forecasts of the useful life of server and network equipment. From a profitability lens, these changes increase short-term earnings, so it's essential to review these from a cash flow lens to understand their true cash earnings.
In the table below, I added their Capex, depreciation, share-based compensation (SBC) and cash flow from operations (CFO) between Q2 2021 - Q2 2025 (4 years) and divided these by their cash flow from operations during the four years.
From a Capex standpoint, a few things seem worrying:
Meta’s Capex/cash flow from operations is nearly as high as TSMC. One wouldn’t expect this when comparing a social media app versus a semiconductor manufacturing company.
Both Meta and Alphabet’s Capex/depreciation have significantly increased over the past 4 years (see chart below). While these could mean several things, one potential outcome is that both companies might be potentially "over-earning"
From a Capex standpoint, a few things seem worrying:
Meta’s Capex/cash flow from operations is nearly as high as TSMC. One wouldn’t expect this when comparing a social media app versus a semiconductor manufacturing company.
Both Meta and Alphabet’s Capex/depreciation have significantly increased over the past 4 years (see chart below). While these could mean several things, one potential outcome is that both companies might be "over-earning"
Over the past four years, Amazon has delivered very little free cash flow, just $33.7 billion, less than ASML’s free cash flow during the period, despite being 8x bigger in market cap.
Amazon’s share-based compensation is roughly a quarter of its total free cash flow during the past four years.
These are certainly problems and potential strains for Meta, Alphabet and Amazon and with the ever-increasing forward Capex guidance, a potential AI slowdown could pose financial challenges for each of them. Compared to these three, Microsoft and Apple (see chart above) have been more disciplined in terms of managing their Capex intensity while preserving cash.
Again, this doesn't necessarily mean a great thing for the business case; Apple's minimal Capex intensity could reflect its minimal reinvestment opportunities available in iPhone and other hardware businesses.
In the future, it's crucial to monitor their free cash flow, especially with Meta and Alphabet's increasing diversion between Capex and depreciation growth (chart above).
Valuations and final thoughts
To conclude, it's important to bring the quality and growth analysis and learnings from the Q2 2025 earnings calls into a valuation perspective. The table below highlights the profitability margins, return on capital, forward earnings growth (2-year EBITDA & EPS) and the key earnings multiples, EV/EBIT and P/E on both a trailing and forward basis.
For disclosure, I own both Alphabet and TSMC, and I'm closely watching their two overarching risks and red flags. Alphabet, as I mentioned, has a growing cash flow generation challenge, given by its Capex intensity, while TSMC faces a direct impact on chip demand from any AI demand slowdown.
Outside my holdings, I'm most excited about Microsoft and Amazon. Microsoft, as we discussed, is in its own league in terms of quality and, in my view, will be less impacted than most on any potential downturn. From its call, the growth target of 15% EPS (table above) doesn't seem unrealistic, and I could potentially purchase its shares if the market presents an opportunity to buy at or less than 30x earnings.
Amazon, on the other hand, had enjoyed a higher earnings multiple relative to peers, but with online sales growing more slowly than its physical store operations and its cloud business potentially losing share to Azure, it's important to reflect on what earnings multiple its business is worth going forward.
On one hand, it's pleasing to see Amazon generate over $120 billion in cash flow from operations over the past 12 months. However, if nearly all of its cash goes into Capex, then one must scrutinise the capital allocation within its Capex to ensure it's spent on projects and investments that deliver better services for customers and profits.
Tesla, as I analysed, is the one big tech company on the list I remain uncertain about. I'm neither bullish nor bearish, but with increasingly intensive competition in the electric vehicle market, slowing car sales, cuts in regulatory credits, and profit margin compression, 11x revenue does optically seem high.
That said, as shared from the earnings call, its future is not in existing businesses but in ride-hailing and its general-purpose robots, Optimus. If Musk does achieve his target of 1 million robots per year in 5 years, that's nearly $30 billion in additional revenue and could materially change Tesla's earnings projections.
For me, though, it would be more fun to watch this on the sidelines as a tech enthusiast, rather than as a Tesla shareholder at current valuations.