The Big Four released their earnings reports simultaneously, with Amazon and Google emerging as the big winners
Amazon delivered an impressive earnings report: revenue growth, a significant profit increase, AWS regaining momentum, and continued expansion in its advertising business.
But the most important aspect of this report is that Amazon is repositioning itself as an infrastructure company for the AI era using AWS.
On the day of the earnings release, Amazon closed up about 1.29% during regular trading hours. After the earnings announcement, the stock price once turned negative in after-hours trading, with losses exceeding 3% at one point, as the market was mainly concerned about high AI and cloud infrastructure capital expenditures and a significant decline in free cash flow; however, AWS's growth exceeded expectations, and management’s comments on AI chip demand helped restore market sentiment, causing the stock to turn positive, rising nearly 4% at one point. As of this writing, Amazon’s after-hours increase is approximately 1.7%.
On April 29 local time, Silicon Valley was bustling as four giants – Alphabet, Microsoft, Meta, and Amazon – reported their results on the same day. In a direct comparison, the differences were stark – Google surged, Amazon rose, Microsoft reacted indifferently, while Meta fell.
All four companies are investing heavily in AI, but market reactions have varied.
Alphabet has Google Cloud and TPU, Amazon has AWS and Trainium, Microsoft has Azure and Copilot. Their capital expenditures are substantial, but cloud revenue, corporate orders, and large customer demands have absorbed the high spending. Meta also has a strong advertising business, but after raising its full-year capital expenditure forecast again, the market first saw a more expensive future bill.
Amazon is currently among the more favored companies, and it is trying to prove that AI infrastructure is a business that AWS can handle.
However, Amazon also faces challenges, as AI infrastructure is not a light-asset business. The stronger Amazon’s story becomes, the more costly it gets. The future remains uncertain and will depend on whether AI demand continues to rise.
A
First, let’s look at the fundamentals.
In the first quarter of 2026, Amazonreported net sales of $181.519 billion, a year-over-year increase of 17%。
Among them,AWS revenue reached $37.587 billion, growing 28% year-over-year, marking the fastest growth rate in the past 15 quartersBased on this quarter's calculations, AWS's annualized revenue scale is approaching $150 billion.
The advertising business is also continuing to expand. In the first quarter, Amazon's advertising services revenue reached $17.243 billion, a year-over-year increase of 24%. Andy Jassy, CEO of Amazon, mentioned in the earnings release that Amazon Advertising's revenue over the past 12 months has exceeded $70 billion.
In terms of profit performance, Amazon's operating profit for the first quarter was $23.852 billion, compared to $18.405 billion for the same period last year, representing a year-over-year increase of approximately 29.6%. Net profit was $30.255 billion, compared to $17.127 billion for the same period last year, a year-over-year increase of approximately 76.7%. Diluted earnings per share were $2.78, compared to $1.59 for the same period last year, a year-over-year increase of approximately 74.8%.
Amazon provided a relatively positive outlook for the next quarter: projected net sales for the second quarter of 2026 are expected to be between $194 billion and $199 billion, representing a year-over-year increase of 16% to 19%. Operating profit is projected to be between $20 billion and $24 billion.
This is an impressive earnings report, but it should not be simply interpreted as 'a surge in net profit'.
It is important to note that in Amazon's first quarter,the net profit includes a pre-tax gain of $16.8 billion from its investment in Anthropic,which is categorized under non-operating income, not the operating profit of its core business. Therefore, the increase in net profit should not be entirely equated with natural growth in Amazon's main operations.
If we only look at the operational side, Amazon remains a highly efficient machine; however, when looking at net profit, it also includes financial gains brought by AI investments.
While Amazon is showcasing its quarterly performance, it is also demonstrating its bets on the AI era, which have started to simultaneously impact revenue, profit, cash flow, and capital market narratives.
Among these, the most critical part is AWS.
In the past few years, the cloud computing industry has undergone a round of 'cost reduction and efficiency enhancement.' Many enterprise customers began to scrutinize their cloud expenditures, cutting unnecessary computing and storage resources. Amazon AWS, Microsoft Azure, and Google Cloud have all faced pressure from slowing growth rates.
However, AI has reignited the cloud business. Large model training requires computing power, inference requires computing power, and Agent operations require computing power. Internal corporate data integration, permission management, model evaluation, and application deployment also rely heavily on the infrastructure provided by cloud vendors.
In the past, the core selling point of AWS was elastic computing and storage.
Now,In this earnings report, Amazon repeatedly emphasized chips, models, Agents, inference, and enterprise applications. AWS has been rebranded as an AI infrastructure platform.
The first layer of AI infrastructure is chips.
Amazon disclosed this quarter that the annualized revenue of its chip businesses, including Graviton, Trainium, and Nitro, has exceeded $200 billion. Over the past 12 months, AWS deployed more than 2.1 million AI chips, with more than half being Trainium.
Meanwhile, AWS still plans to deploy over 1 million NVIDIA GPUs by 2026. Rather than focusing on how to bypass NVIDIA, Amazon emphasizes that its self-developed chips are beginning to scale up and keep customers within AWS's ecosystem of chips, networking, storage, model services, and Agents.
Anthropic is the deepest tied card, committing to invest over $100 billion into AWS in the next decade and securing up to 5GW of Trainium capacity. OpenAI has also committed to using approximately 2GW of Trainium capacity starting from 2027.
With Bedrock’s token processing volume in Q1 surpassing the total of all previous years combined and customer spending increasing 170% quarter-over-quarter, Amazon aims to prove: From training, inference to Agent scheduling, every layer of AI infrastructure can be built on AWS.
B
Today's Silicon Valley is particularly bustling, not only because Amazon released its latest earnings report.
On April 29 local time, four giants released their earnings reports: Alphabet, the parent company of Google, Microsoft, Meta, and Amazon.
Alphabet: Shares rose over 6% in after-hours trading, mainly driven by high growth in Google Cloud and progress in AI commercialization.
Amazon: After-hours saw fluctuations, with a drop of more than 3% at one point, then turning around to rise nearly 4%, before settling at a gain of over 1%. AWS's growth exceeded expectations, and management’s comments on AI chip demand helped restore market sentiment.
Microsoft: Shares fell over 2% in after-hours trading. Azure’s growth met expectations but was not impressive enough.
Meta: Shares dropped over 6% in after-hours trading, mainly due to raising the full-year capital expenditure forecast.
Interestingly, in terms of performance, each company demonstrated its own strengths, but the market responded differently.
One clear observation is that the three companies with cloud businesses did not perform poorly.
Alphabet received the best response in this round of earnings releases. Google Cloud's Q1 revenue reached $20 billion, a year-on-year increase of 63%, which is a very strong growth rate.
More importantly, Google has now demonstrated that models, cloud services, TPUs, and enterprise customers are forming a connected business chain, rather than just focusing on Gemini's model capabilities. Apple has announced that its next-generation foundational models (Apple Foundation Models) will be based on Google's Gemini models and cloud technology. In April, Anthropic expanded its TPU computing power partnership with Google and Broadcom, with future multi-gigawatt capacity to be used for Claude model training and services.
In other words, Google is not only developing its own models but has also started to become an underlying supplier for other large model companies and terminal giants.
Reuters cited analysts' views stating that Google Cloud’s performance shows that Google is transforming its AI research advantage into cloud business growth that can be verified by financial reports. The market can see that people are already paying for Google’s AI investments.
Amazon's situation also looks positive. After the earnings report, Amazon shares rose nearly 4% in after-hours trading. AWS revenue for the first quarter was $37.6 billion, a year-on-year increase of 28%, surpassing market expectations.
This figure is crucial for Amazon because over the past few quarters, the market had been concerned about AWS’s slowing growth and losing AI cloud demand to Microsoft and Google. This re-acceleration of AWS is akin to giving the market a vote of confidence.
Moreover, Amazon now holds two key cards: Anthropic and OpenAI. Anthropic is deeply tied to AWS, and OpenAI has also started using AWS’s Trainium capacity. Investing.com senior analyst Jesse Cohen believes that the re-acceleration of AWS sales is the most prominent highlight of this earnings report, indicating that customers are running more new computing tasks, especially AI-related ones, on AWS.
Microsoft appears relatively mediocre among the three cloud providers.
Azure and other cloud revenues increased by 40% year-on-year, a figure that is not bad, but not particularly 'surprising,' nor does it exceed expectations as obviously as Google Cloud did.
The issue with Microsoft is not that its AI narrative is weaker than its competitors’. On the contrary, Microsoft has been at the center of the AI narrative over the past year, with Copilot, Azure, and OpenAI’s binding relationships all repeatedly priced by the market. Precisely because of such high expectations, Microsoft needs to deliver equally outstanding results.
Reuters cited Valoir CEO Rebecca Wettemann’s statement that, against the backdrop of Google significantly exceeding expectations and the market worrying about Microsoft’s AI infrastructure expenditures, Microsoft needed to produce numbers that could impress the market—but this earnings report failed to do so.
Meta, on the other hand, was the most awkward player in the day’s earnings competition.
Meta's core business performed strongly, with revenue in the first quarter growing 33% year-over-year. Both ad impressions and ad prices are on the rise. However, the issue is that Meta has raised its full-year capital expenditure forecast from $115 billion–$135 billion to $125 billion–$145 billion, and the market is primarily seeing a more expensive AI bill.
Compared to Google, Amazon, and Microsoft, it is harder to separately extract Meta’s AI returns. Meta uses AI to improve ad recommendation efficiency, enhance content distribution and conversion, while also relying on AI to support AI assistants, AI glasses, and future personal super-intelligent interfaces. For now, the area with significant returns remains advertising.
Moreover, Meta's recent moves in AI adjustments have been aggressive, but the results are coming out relatively slowly.
In June last year, Meta invested $14.3 billion in Scale AI and recruited its CEO, Alexandr Wang, to lead the newly formed Super Intelligence Lab (MSL). This move was seen at the time as Zuckerberg's major overhaul of the AI team, driven by the pressure of Llama 4 underperforming and the flagship model Behemoth being delayed.
However, from June last year to April this year, Meta only recently officially released Muse Spark, the first model from MSL. Meta officially stated that MSL had rebuilt the AI technology stack over the past nine months; but in the eyes of the market, this means that Meta’s spending and restructuring pace has been ahead, while verifiable models and product outcomes are just starting to emerge.
The annual planned expenditures of the other three companies are also very high: Alphabet is expected to spend $180 billion–$190 billion, Microsoft is projected at $190 billion, and Amazon spent $44.2 billion in the first quarter, with an estimated $200 billion for the full year.
All four companies released their earnings reports on the same day, and the market is not entirely unwilling to accept high expenditures set for AI.
Google performed the strongest, with Gemini, TPU, and Google Cloud forming a closed loop. AI investments are turning into cloud revenue and enterprise orders. Amazon also passed the test, with AWS regaining speed, and Anthropic and OpenAI endorsing its AI infrastructure. Microsoft appeared slightly mediocre, with Azure still performing strongly but too much in line with expectations, leaving the market unwilling to pay extra for 'normal performance.'
Meta is at a disadvantage, as the money it has poured in points more towards the future: larger data centers, more expensive chips, stronger models, and AI assistants and glasses interfaces.
Cloud vendors burning cash feel like they’re expanding production; Meta burning cash feels like it’s placing bets.
Amazon is on a path that is currently more favored.
C
Let's turn our attention back to Amazon.
AI infrastructure is not a lightweight asset business. The figure that best reflects the pressure of investment in this earnings report is free cash flow.
Notably, Amazon's operating cash flow is not weak. In the past 12 months ending March 31, 2026, operating cash flow increased by 30% year-over-year, reaching $148.5 billion. However, free cash flow dropped from $25.88 billion in the same period last year to $1.194 billion.
The company explained that the decline in free cash flow was mainly due to a year-over-year increase of $59.259 billion in property and equipment purchases, with this increase primarily reflecting AI investments.
This is the cost of the AI infrastructure business.
Amazon can talk about Trainium, Bedrock, OpenAI, Anthropic, Meta, Uber, and Cerebras. But to support these stories, it must first invest substantial capital in building data centers and networks, developing and deploying chips, while also providing corresponding power and cooling capabilities.
AI infrastructure is more like a new generation of industrial infrastructure. It consumes cash in the early stages but may release economies of scale later on.
The good news for Amazon is that it has high-margin businesses like AWS, and e-commerce and advertising provide cash flow. The bad news is that the scale of AI infrastructure investment is enormous, and the return cycle is not entirely within Amazon’s control.
If corporate AI demand continues to surge, Amazon will be one of the biggest beneficiaries. The more customers it has, the heavier the workload, the more valuable its data centers, chips, model platforms, and Agent management tools become.
However, if enterprise AI adoption and payments fall short of expectations, or the cost of model inference declines too rapidly, or customers frequently switch between multiple cloud providers, these heavy asset investments could turn into financial statement pressures.
In other words, Amazon's AI narrative is more solid than many AI application companies but also much heavier. It isn't as glamorous as Tesla's Robotaxi and humanoid robots, nor does it easily draw public attention like OpenAI’s model releases.
It's akin to building roads, bridges, and power plants in advance. If traffic volume is sufficient, this becomes the best business; however, if traffic is insufficient, the initial investment will appear extremely costly. $Amazon (AMZN.US)$$Star Tech Companies (LIST2518.US)$$Alphabet-C (GOOG.US)$$Alphabet-A (GOOGL.US)$$Microsoft (MSFT.US)$$MICROSOFT-T (04338.HK)$$Meta Platforms (META.US)$$Virtual Reality (LIST2139.US)$
Risk Disclaimer: The above content only represents the author's view. It does not represent any position or investment advice of Futu. Futu makes no representation or warranty.Read more
Comment (1)
to post a comment
1