To The Moon

$Amazon.com(AMZN)$ As we head in to the new month!!

AWS growth has accelerated to its fastest pace in years, but Amazon’s free cash flow has turned negative. Is this the beginning of an AI payoff, or the most expensive growth cycle in company history?

Amazon has just delivered one of the most important earnings reports of the artificial-intelligence investment cycle.

The headline numbers were strong.

Revenue exceeded expectations.

AWS accelerated.

Artificial-intelligence demand remained enormous.

Amazon shares jumped after hours.

However, underneath the earnings beat sits a much more complicated question.

Amazon now expects to invest approximately US$220 billion in capital expenditure during 2026, up from its previous expectation of around US$200 billion. At the same time, trailing free cash flow has moved into negative territory as the company funds data centres, chips, networking equipment, robotics and other infrastructure.

That means Amazon’s latest quarter was not merely an earnings announcement.

It was a test of whether investors are still willing to finance one of the largest corporate infrastructure programmes in history.

Microsoft recently passed that test by demonstrating that rapid Azure growth and visible AI monetisation could justify its enormous spending.

Amazon is now attempting to prove the same thing through AWS.

The central question is:

Is Amazon building too aggressively, or is AWS growth finally strong enough to justify the bill?

⸻

📈 1. Why the market initially liked Amazon’s result

Amazon reported quarterly revenue of approximately US$200.6 billion, supported by strength across cloud computing, advertising and its retail operations.

The biggest positive was AWS.

AWS revenue increased approximately 37% year over year to US$42.4 billion, representing its fastest growth rate in more than four years. Amazon also indicated that its AI and custom-chip businesses had reached significant revenue run rates.

That acceleration matters because AWS is the most important part of the Amazon investment case.

Retail creates enormous revenue.

Advertising adds attractive margins.

Prime builds customer loyalty.

However, AWS remains the segment investors rely upon to produce high-margin growth and finance Amazon’s broader ambitions.

If AWS growth had weakened while expenditure continued rising, the market may have punished Amazon severely.

Instead, AWS accelerated.

That suggests Amazon’s infrastructure is not sitting empty while management waits for customers to arrive.

Demand appears to be absorbing new capacity quickly.

Amazon has also indicated that it may remain unable to satisfy all expected demand for several years, which supports the argument that current spending is responding to genuine customer requirements rather than speculative overbuilding.

Adz’s perspective

This was the result Amazon needed.

The market was not going to reward another vague promise about future AI demand. It needed evidence that AWS growth was accelerating now, and Amazon delivered it.

However, strong demand does not automatically guarantee strong shareholder returns. The size and cost of the buildout still matter.

⸻

💰 2. The US$220 billion question

Amazon’s expected capital expenditure has risen from approximately US$200 billion to around US$220 billion for 2026.

That is an extraordinary amount of money.

The investment is expected to support areas including:

* AWS data centres

* AI computing infrastructure

* Amazon’s custom Trainium and Inferentia chips

* Networking and storage

* Robotics and logistics automation

* Satellites and communications infrastructure

* Rising component and memory costs

Management argues that demand remains stronger than available supply and that the company can generate attractive long-term returns from the capacity being built.

There is a logical case for spending aggressively.

Cloud customers are signing longer-term commitments.

AI models require enormous processing capacity.

Inference demand may eventually become larger than model-training demand.

Businesses increasingly need databases, security, storage, networking and application tools alongside raw computing power.

AWS can potentially earn revenue from every layer.

Amazon is therefore not merely purchasing expensive GPUs.

It is building a complete platform that can host the development, training, deployment and operation of artificial-intelligence applications.

The challenge is that Amazon must pay for much of that infrastructure before the associated revenue fully arrives.

That creates pressure on free cash flow.

⸻

⚠️ 3. Free cash flow is where the debate becomes uncomfortable

Amazon’s trailing free cash flow has reportedly fallen to approximately negative US$7.6 billion, compared with positive free cash flow a year earlier. The major reason is the rapid increase in purchases of property and equipment connected with AI and infrastructure investment.

This does not mean Amazon’s underlying operations are collapsing.

It means the company is reinvesting enormous amounts of operating cash into future capacity.

However, shareholders cannot ignore the difference between earnings and cash flow.

A company may report:

* growing revenue

* rising operating income

* stronger earnings per share

* accelerating cloud growth

while still generating weak or negative free cash flow after capital expenditure.

That can be acceptable temporarily.

It becomes dangerous when the spending never slows or when the future cash returns fail to compensate investors for the upfront sacrifice.

The key question is not whether Amazon can afford US$220 billion.

Amazon’s scale, cash generation and access to capital suggest that it can.

The real question is:

What return will shareholders eventually receive on every dollar invested?

Adz’s perspective

AWS growth gives Amazon permission to keep investing, but it does not give management unlimited permission.

The market accepted the higher spending because cloud growth accelerated sharply. If AWS growth falls back while capital expenditure remains above US$200 billion, sentiment could change very quickly.

⸻

🧠 4. Amazon’s AI monetisation model is different from Microsoft’s

Microsoft can monetise its AI infrastructure through several direct channels:

* Azure consumption

* Microsoft 365 Copilot subscriptions

* GitHub Copilot

* security products

* enterprise software

* data and analytics tools

Amazon’s pathway is more infrastructure focused.

AWS earns revenue when customers consume:

* computing capacity

* storage

* databases

* networking

* cybersecurity services

* model access

* training and inference services

* custom chips

* Bedrock and related AI-development tools

Amazon does not necessarily need to own the most popular consumer chatbot.

It needs businesses to build AI applications on AWS.

That is an important distinction.

The number of AI models may continue changing.

The leading application may change.

Consumer preferences may change.

However, nearly every serious AI product still requires computing, data storage, networking and security.

AWS wants to provide that foundation.

Amazon can therefore benefit even when another company creates the winning application.

It is similar to selling the tools, roads and electricity required for an industrial expansion.

The risk is that cloud infrastructure can become more competitive and capital intensive.

Microsoft Azure, Google Cloud and specialised AI-cloud providers are all pursuing the same opportunity.

Customers may also use multiple cloud providers, which limits the ability of one platform to dominate pricing.

⸻

⚙️ 5. Custom chips could become Amazon’s hidden advantage

One of Amazon’s most important strategic moves has been developing its own semiconductor products.

Trainium is designed for AI training.

Inferentia is designed for AI inference.

Graviton supports general cloud-computing workloads.

The purpose is not necessarily to replace $NVIDIA(NVDA)$ entirely.

The purpose is to give AWS customers additional options while reducing Amazon’s dependence on one external supplier.

Custom chips could potentially:

* lower the cost of providing AI services

* improve performance for specific workloads

* protect AWS margins

* reduce supply constraints

* create stronger customer loyalty

* give Amazon greater control over its infrastructure roadmap

This could become increasingly important as AI demand moves from training into inference.

Training a frontier model is extremely expensive, but it occurs periodically.

Inference occurs whenever a customer or application uses that model.

If billions of people and businesses begin using AI tools every day, inference could become the larger and more durable opportunity.

A lower-cost chip platform would give Amazon a stronger chance of earning attractive margins from that volume.

Adz’s perspective

Amazon’s custom chips may matter more than the market currently appreciates.

The company does not need to defeat Nvidia across every workload. It only needs to provide customers with a credible, cost-effective option that increases AWS usage and protects Amazon’s margins.

⸻

🏢 6. AWS is not Amazon’s only AI return pathway

The AWS story receives most of the attention, but artificial intelligence can create returns throughout Amazon.

Retail

AI can improve:

* product recommendations

* inventory forecasting

* pricing

* warehouse efficiency

* delivery routes

* customer support

* fraud detection

* seller tools

Advertising

Amazon can use AI to improve:

* advertisement targeting

* campaign creation

* conversion rates

* product discovery

* measurement and attribution

Advertising is particularly important because it is generally more profitable than traditional retail activity.

Robotics

Automation may reduce the cost of:

* moving inventory

* sorting packages

* operating warehouses

* preparing orders

* completing deliveries

Alexa and consumer services

Generative AI could make Alexa more useful and potentially create subscription or commerce opportunities.

These additional pathways mean Amazon’s AI investment should not be judged solely through AWS revenue.

Some of the return may appear through lower costs, faster deliveries, stronger advertisement performance or higher retail conversion.

The difficulty is measurement.

Investors need management to explain how much financial benefit AI is creating outside AWS rather than relying upon general statements about productivity.

⸻

📉 7. The bear case: strong demand can still produce weak returns

The bear case does not require AI to fail.

Artificial intelligence can become one of the most important technologies in history while certain infrastructure investments still generate disappointing returns.

Amazon could face several problems.

Overbuilding

Demand may be strong today, but infrastructure takes years to plan and construct.

If customers become more efficient or demand slows, Amazon could end up with excess capacity.

Falling prices

Cloud providers may compete aggressively to win AI workloads.

Strong usage growth would not necessarily produce strong profit growth if pricing declines.

Rapid hardware obsolescence

AI chips advance quickly.

Equipment purchased today may become less competitive before Amazon earns an adequate return.

Higher depreciation

Servers and data-centre equipment flow through the income statement over time.

Even after the immediate capital expenditure occurs, depreciation can pressure future margins.

Power and construction constraints

Data centres require enormous amounts of electricity, land, cooling and networking infrastructure.

Delays and higher costs can weaken project returns.

Persistent reinvestment

The biggest danger may be that spending never truly ends.

If every new generation of AI requires another enormous infrastructure cycle, free cash flow may remain under pressure much longer than investors expect.

Adz’s perspective

The risk is not that Amazon has misunderstood AI.

The risk is that Amazon is correct about the demand but investors underestimate the permanent cost of remaining competitive.

⸻

🐂 8. The bull case: Amazon may still be supply constrained in 2028

Amazon has suggested that demand could remain ahead of available capacity beyond the current year, potentially extending into 2027 or 2028.

If that is accurate, the investment cycle may have significantly further to run.

The bull case assumes:

* AWS continues growing above historical cloud-market rates

* businesses increase AI adoption

* inference expands rapidly

* Bedrock attracts more enterprise workloads

* Trainium and Inferentia gain adoption

* advertising margins remain strong

* robotics lowers retail costs

* capital expenditure eventually stabilises

* free cash flow rebounds once major projects begin producing revenue

Under that scenario, today’s negative free cash flow would represent a temporary investment phase rather than a structural problem.

Amazon would emerge with:

* greater cloud capacity

* more proprietary technology

* deeper customer relationships

* stronger logistics automation

* a larger advertising business

* improved operating efficiency

The US$220 billion figure would appear less frightening if it created years of high-margin growth.

⸻

🐻 9. The bear case: Amazon may be financing an endless arms race

The bearish scenario assumes:

* AWS growth slows after capacity expands

* competitors reduce cloud pricing

* custom chips struggle to gain adoption

* depreciation rises faster than expected

* energy and component costs stay elevated

* free cash flow remains negative

* investors demand stronger near-term returns

* Amazon continues increasing expenditure to avoid falling behind

Under this scenario, the infrastructure remains useful, but shareholders wait too long for the financial return.

This is similar to previous technology buildouts.

The internet created enormous value.

Fibre infrastructure became essential.

Data usage exploded.

However, some investors still suffered because companies overpaid, overbuilt or financed projects on unrealistic assumptions.

A correct long-term technology thesis does not guarantee a correct investment at every price.

⸻

📊 10. The numbers investors should watch next

Amazon’s next few quarters should be judged using a clear scorecard.

AWS revenue growth

AWS growth must remain elevated as additional infrastructure comes online.

A sharp slowdown would challenge the investment thesis.

AWS operating income

Revenue growth is not enough.

AWS must protect its profitability despite hardware, depreciation and energy costs.

Free cash flow

This may be the most important figure.

Investors will eventually expect free cash flow to recover.

Capital expenditure

The market accepted US$220 billion because demand appears strong.

Another major increase could test investor patience.

Custom-chip adoption

Greater use of Trainium and Inferentia may improve margins and reduce supply dependence.

Backlog and customer commitments

Long-term contracts would provide greater confidence that new capacity will be utilised.

Advertising growth

Advertising can help finance the infrastructure cycle and improve Amazon’s overall margins.

Retail efficiency

Investors should watch whether AI and robotics create measurable improvements in fulfilment and delivery costs.

⸻

🏆 11. Has Amazon passed the Microsoft test?

Microsoft passed the market’s AI-spending test because investors could see:

* accelerating Azure growth

* direct AI revenue

* growing Copilot adoption

* sustained enterprise demand

* confidence that new capacity would be absorbed

Amazon has now produced one major part of the same argument.

AWS growth accelerated dramatically.

Demand remains strong.

The company appears capacity constrained.

However, Amazon has not completely passed the test yet.

Its free-cash-flow pressure is more severe.

Its capital expenditure has increased again.

Its direct AI monetisation is less transparent than Microsoft’s Copilot and Azure combination.

Adz’s assessment

Amazon passed the revenue-growth test, but it has not yet passed the cash-flow test.

AWS acceleration justifies optimism. Negative free cash flow demands caution.

The next stage is not proving that customers want AI infrastructure. That has already been established. Amazon must now prove that serving this demand will create enough cash to reward shareholders.

⸻

🔍 12. Amazon versus the other Big Tech AI spenders

Microsoft

* Clearest direct AI monetisation

* Strong enterprise distribution

* Azure plus Copilot creates multiple revenue streams

* Current leader in turning spending into visible revenue

Alphabet

* Strong cloud growth

* Gemini adoption across consumer and enterprise products

* Search and advertising provide additional monetisation

* Faces disruption risk within its most profitable business

Amazon

* Strongest pure infrastructure opportunity

* AWS can benefit from growth across the entire AI ecosystem

* Custom chips provide a potential cost advantage

* Free cash flow remains the major concern

Meta

* Powerful advertising business funds investment

* AI improves recommendations and advertisement targeting

* Broader superintelligence returns remain difficult to measure

Apple

* Lowest exposure to the infrastructure arms race

* Huge device distribution

* Direct AI monetisation remains less visible

Amazon may not have the clearest monetisation model, but it may have the broadest opportunity to supply the infrastructure beneath other companies’ AI ambitions.

⸻

🎯 Final take

Amazon’s latest result strengthens the argument that AI infrastructure demand is real.

AWS growth accelerated to approximately 37%.

Customers continue consuming more cloud capacity.

Amazon remains willing to invest aggressively.

The company’s AI and custom-chip businesses are becoming meaningful.

That is the good news.

The uncomfortable part is the cost.

US$220 billion is not a normal investment programme.

Negative free cash flow is not something shareholders can ignore forever.

Amazon now needs to demonstrate that AWS growth, custom chips, advertising and operational efficiency can convert its infrastructure into durable cash generation.

Adz’s final verdict

Amazon is currently my strongest challenger to Microsoft in the AI infrastructure race.

The earnings result was impressive enough to justify the immediate share-price reaction, but I would not call the investment fully proven yet.

AWS has passed the demand test.

The next test is margins.

The final test is free cash flow.

If Amazon passes all three, the current infrastructure buildout could become one of the most valuable investments in its history.

If it does not, shareholders may discover that even extraordinary revenue growth can become expensive when maintaining leadership requires more than US$200 billion per year.

⸻

🗳️ What is your move?

Has Amazon’s AWS acceleration justified its US$220 billion investment plan?

A. Strong buy, AWS growth proves the investment is working

B. Cautiously bullish, demand is strong but free cash flow must recover

C. Hold, the earnings were good but the spending is too aggressive

D. Bearish, Amazon is overbuilding and shareholders will carry the cost

E. Microsoft remains the better AI investment

Which figure matters most to you?

* AWS growth

* Capital expenditure

* AWS operating margins

* Free cash flow

* Custom-chip adoption

Drop your view below.

Adz

Disclaimer: This post is for market discussion and education only. It does not constitute personal financial advice. Investing involves risk, and readers should conduct their own research before making financial decisions.