An in-depth look at Alphabet, Microsoft, artificial intelligence, cloud growth, and the real test behind the latest earnings season.
Research note: October 28, 2026. This article was prepared before the companies released their latest quarterly results. The figures discussed below are published historical results and analysts’ estimates, not confirmed outcomes for the current quarter.
I've been thinking about what a real earnings beat looks like
I've been watching the development of artificial intelligence for some time, and I keep coming back to one question: at what point does growing interest in AI become a business that can reliably pay for itself?
That question feels particularly relevant as Google and Microsoft prepare to report their latest quarterly results. Both companies have spent years building businesses that reach billions of people. Both have strong positions in cloud computing, software, advertising, and digital infrastructure. And both are now trying to make artificial intelligence a larger part of how people work, search for information, create content, and run businesses.
The upcoming reports matter because they bring several important questions together. Is demand for AI computing still accelerating? Are businesses willing to pay enough for AI services to make the infrastructure costs worthwhile? Can Google protect its established search business while changing the way people find information? And can Microsoft turn its extensive cloud and software ecosystem into a lasting advantage as AI becomes part of everyday work?
These aren't questions that can be answered by looking at one earnings-per-share number.
A company can report revenue above expectations while spending more than anticipated to support that growth. It can post strong cloud numbers while struggling to bring enough computing capacity online. It can also report healthy profits without proving that its newest products will remain profitable as they reach millions of customers.
That's why I think this earnings session deserves a closer look than a simple prediction about which company will beat Wall Street.
The focus should be on the businesses underneath the headline figures, the money being spent to support them, and the evidence that customers are finding enough value to keep paying.
What the market is expecting from Google and Microsoft
Both companies are scheduled to report on Wednesday, October 28, after the U.S. stock market closes. Alphabet's third-quarter earnings call is scheduled for 4:30 p.m. Eastern Time, while Microsoft's fiscal 2027 first-quarter results are also due after the market closes.
At the time of writing, the results remain pending. The most useful starting point is therefore the set of expectations analysts have established before the announcements.
According to Reuters' October 27 reporting, which cited LSEG and Visible Alpha estimates, Alphabet is expected to report approximately $128.18 billion in quarterly revenue. Google Cloud revenue is expected to reach $29.57 billion, representing growth of about 95.1% from a year earlier. Search revenue growth, meanwhile, is expected to slow to approximately 14.7%, compared with 16.8% in the previous quarter.
For Microsoft, analysts expect approximately $90.66 billion in revenue. Azure sales are estimated at around $33.06 billion, with growth expected to accelerate to approximately 47.7% from 42% in the previous quarter, according to the same Reuters report.
These are estimates, not company-reported figures. Different data providers can publish slightly different consensus numbers, especially for earnings per share, where adjusted and generally accepted accounting principles (GAAP) figures may differ. The precise benchmark matters because an earnings beat is measured against what analysts already expect, not simply against the company's performance in the previous quarter.
Source: Reuters — Big Tech earnings to test AI boom amid market turmoil, October 27, 2026.
What stands out to me is that the two companies face different versions of the same problem.
Google must show that the strength of its cloud business can continue while its core advertising business adapts to AI-driven changes in search. Microsoft must demonstrate that the rapid expansion of Azure and AI services can support further infrastructure spending without placing too much pressure on cash flow and margins.
The market has already seen substantial growth from both businesses. That means a good result alone may not answer every question. Investors, customers, and developers will also be listening for an explanation of what comes next.
Google: the challenge of changing a business that already works
Google occupies an unusual position in the AI industry. It doesn't have to build a large audience from scratch. Search, YouTube, Android, Chrome, and Google Workspace already provide established ways to reach users. Its challenge is to introduce new AI experiences without damaging the economic foundations that support these products.
This is more difficult than it may sound.
For years, Google's search experience largely connected a question with a list of websites, advertisements, and other relevant results. AI-generated answers are changing that interaction. A person can now ask a complicated question and receive a synthesized response instead of clicking through several pages.
That makes search more conversational and can help people find information faster. However, it also changes how Google organizes information, how advertising appears, and how publishers and businesses receive visitors.
Google has to make that transition while maintaining the relevance and commercial value of Search.
The financial foundation remains important
Alphabet's second-quarter 2026 results provide a useful starting point.
The company reported $119.8 billion in revenue for the quarter, up 24% year over year. Google Services generated $94.5 billion, increasing 15%. Search and other revenue grew 17%, while YouTube advertising revenue increased 13%. Google Cloud reported $24.8 billion in revenue, up 82%. Alphabet's operating income increased 30% to approximately $40.8 billion, and its operating margin reached 34%.
Those figures show that growth isn't coming from a single product. Search and advertising remain significant sources of revenue, while subscriptions, consumer products, and cloud computing extend the business into other areas.
The original financial details are available in Alphabet's second-quarter 2026 earnings release filed with the SEC.
This breadth matters because Google can introduce AI in several different ways. It can use new models to improve the search experience, integrate them into Workspace, provide access through its Gemini app, and sell AI infrastructure and software to enterprise customers.
The company isn't relying solely on people deciding to pay for a separate chatbot. It can bring AI into services people already use.
But reach and usage should not be confused with revenue. A feature can attract millions of users without producing enough additional income to cover its operating costs. The real question is whether these new experiences improve engagement, support advertising, encourage subscriptions, or create demand for other paid services.
That is the distinction I will be watching in Google's report: whether adoption is translating into measurable business value.
AI in Search: more useful answers, but a changing economic model
Google's AI-powered search features are especially interesting because they affect its oldest and most important business.
In his remarks on Alphabet's second-quarter 2026 earnings call, CEO Sundar Pichai said that Search and AI Overviews were becoming part of a more integrated experience. He also reported that AI Mode had surpassed one billion monthly active users after expanding globally the previous October. Google said its AI-powered features were increasing overall Search queries and that it was sending billions of clicks to websites each week through AI features in Search.
Google also said engineering and hardware improvements reduced the cost of AI Mode responses to their lowest level since launch, even as the company introduced more advanced capabilities.
These are meaningful developments because AI answers require computing resources. Each response must be generated, and the cost of doing that can accumulate rapidly when a service handles enormous numbers of queries.
Source: Google's official Q2 2026 earnings-call remarks.
There is a practical lesson here. A useful AI feature isn't enough by itself. The company also needs to deliver it at a cost that makes sense at scale.
Imagine a service that becomes twice as popular but costs substantially more to operate. The increase in usage might look impressive, yet the financial benefit would depend on whether the service also creates enough additional revenue or savings.
Search has an extra complication. Google needs to maintain a useful experience for people who want quick answers, while still providing commercial opportunities for advertisers and a sustainable relationship with publishers whose material informs those answers.
Some publishers have raised concerns that AI-generated summaries can reduce the need for users to visit the original websites. Google, on the other hand, has said its AI search experiences continue to send substantial traffic to publishers.
The long-term outcome remains uncertain. If people receive more value from Search but fewer websites receive meaningful traffic, the broader information ecosystem could change in ways that affect the availability and diversity of online content.
That issue matters beyond a quarterly revenue figure. Search depends on a continuing supply of useful information, and publishers need viable ways to support their work.
For this earnings session, I'd pay attention to the growth of Search revenue, management's explanation of AI's commercial contribution, and any indication that the cost of providing AI answers is becoming easier to manage. A growing number of queries would be encouraging, but it's not the entire story.
Google Cloud: the clearest test of enterprise AI demand
If Search shows how AI is changing an established business, Google Cloud offers a more direct view of how companies are paying for AI infrastructure and services.
The division's reported 82% revenue growth in the second quarter was one of Alphabet's strongest signals of enterprise demand. Google attributed the increase to Google Cloud Platform services, enterprise AI infrastructure, and enterprise AI solutions.
There are two related parts to this business.
The first is infrastructure. Companies need computing power, storage, networking, and related services to train, deploy, and operate AI models. Some businesses purchase those resources to develop their own products. Others use them to serve customers or run internal applications.
The second is the software and services built around that infrastructure. Businesses may need tools for data analysis, cybersecurity, workplace productivity, model development, and AI agents that help employees complete tasks.
Google can serve both needs. It offers computing resources, develops its own AI models, and provides enterprise products that connect those capabilities to customers' existing systems.
That combination could help it serve customers who want to experiment with AI without assembling every part of the technology themselves.
Alphabet's CEO also said in July that Google Cloud's backlog had reached $514 billion and that nearly 90% of Fortune 100 companies were using Gemini Enterprise. These were company-reported figures at the time of the second-quarter announcement.
The backlog figure is worth examining carefully. It represents contracted demand expected to be recognized as revenue in the future; it isn't the same as money already collected or revenue that will necessarily appear immediately. The timing of delivery, customer usage, capacity availability, and contract terms all affect how quickly it becomes recognized revenue.
Still, a substantial backlog can offer evidence that enterprise customers are making commitments beyond short experiments. It suggests that at least some organizations are planning to incorporate Google's cloud and AI services into longer-term operations.
The next question is execution.
Can Google bring enough computing capacity online to meet demand? Can it deliver the services reliably during busy periods? Can it convert contracts into recognized revenue without letting the cost of infrastructure expand faster than the business can support?
Reuters' latest pre-earnings estimates point to Google Cloud revenue of approximately $29.57 billion for the September quarter, implying growth of more than 95% year over year. That would be a further acceleration from the 82% growth reported in the previous quarter.
Such a figure would be impressive if achieved, but the quality of that growth matters just as much as its pace.
Customers don't just need access to powerful models. They need predictable latency, dependable service, secure data handling, clear billing, and systems that can keep running when workloads increase. A company deploying AI into customer service, software development, or business operations cannot treat reliability as an optional extra.
Google's longer-term opportunity depends on making its infrastructure dependable enough that enterprises can build essential services on top of it.
Google's spending problem: demand can be real and still be expensive
The other side of Google's growth story is the cost of building the infrastructure required to support it.
In July 2026, Alphabet raised its forecast for full-year capital expenditure to between $195 billion and $205 billion, up from the previous range of $180 billion to $190 billion. The company cited demand for AI computing capacity that continued to exceed available supply.
Its second-quarter cash flow figures show why this spending deserves attention. Alphabet generated approximately $39.1 billion in operating cash flow during the quarter but spent about $44.9 billion on property and equipment. Under the company's definition of free cash flow, that resulted in negative quarterly free cash flow of approximately $5.9 billion.
Source: Alphabet's SEC-filed Q2 results and Reuters' July 22 report on Alphabet's higher capital expenditure forecast.
Negative free cash flow in one quarter doesn't automatically indicate a weak business. Major infrastructure investments can require cash today while supporting services that produce revenue over several years.
The difficulty is that AI hardware, data centers, networking equipment, and electricity are not inexpensive. Much of the spending occurs before the resulting capacity can generate revenue. And the market for computing infrastructure can change faster than the physical infrastructure can be built.
Google must therefore balance two risks. If it invests too little, it may be unable to serve customers who are ready to pay. If it invests too much ahead of dependable demand, it can tie up capital in capacity that takes longer than expected to earn an adequate return.
This is a question about timing as much as demand.
For the current earnings report, I would want to understand whether the company continues to expect demand to exceed available supply, how it is thinking about infrastructure investment in 2027, and whether the economics of its cloud and AI services are developing as expected.
Rapid growth tells us that customers are showing up. It doesn't yet tell us the complete cost of serving them.
Microsoft: AI is moving into the tools people already use
Microsoft approaches the AI opportunity from a different starting point.
Its business is built around software, cloud infrastructure, developer tools, security, and enterprise relationships. Customers already use Microsoft products to communicate, manage documents, build software, run workloads, and administer their businesses.
Microsoft's challenge is to make AI useful within these established activities.
Instead of asking people to adopt an entirely new way of working, the company can introduce AI into familiar tools such as Microsoft 365, GitHub, Azure, and its security products.
That creates a potentially useful distribution advantage. An AI feature may reach an employee through a product the company already uses rather than requiring a separate service, a new vendor relationship, and a completely different workflow.
However, distribution alone doesn't guarantee durable adoption. Customers still need to see measurable value, particularly when AI services involve additional subscription costs or usage-based charges.
Microsoft's fourth-quarter fiscal 2026 results provide the baseline for judging its latest performance.
The company reported quarterly revenue of $90.0 billion, up 18% year over year, and operating income of $40.6 billion, also up 18%. Diluted earnings per share were $4.81 under GAAP and $4.74 on the company's non-GAAP basis.
Microsoft disclosed that several discrete items benefited diluted earnings per share by $0.27 relative to its earlier guidance. Those items included a $3.2 billion gain from its investment in Anthropic and lower-than-expected retirement-program expenses, partly offset by other costs.
That detail matters because reported earnings can contain items that don't reflect the underlying performance of the core business in a typical quarter.
Source: Microsoft's official FY2026 Q4 earnings release.
A careful reading of Microsoft's new results should distinguish reported earnings from adjusted earnings, and should explain why the two differ instead of treating one figure as the whole story.
Azure: demand is strong, but capacity is part of the equation
Azure may be the single most closely watched part of Microsoft's upcoming report.
In the fourth quarter of fiscal 2026, Azure and other cloud services revenue grew 43% year over year. Microsoft said customer demand continued to exceed available capacity. It also said that efficiency improvements across its CPU and GPU infrastructure, along with faster deployment of new capacity, helped it deliver stronger-than-expected growth.
The company added that the extra capacity it brought online during the quarter was quickly monetized.
Azure surpassed $100 billion in annual revenue during fiscal 2026, according to Microsoft. The wider Microsoft Cloud business generated $59.3 billion in the fourth quarter alone, up 27% year over year.
Source: Microsoft's Q4 earnings release and conference-call transcript.
These figures establish a strong starting point. Microsoft is not trying to prove that customers have any interest in cloud computing. It is operating a large existing business, and its challenge is to keep expanding that business while supporting new workloads that require considerable computing resources.
That last point is important.
A business can have more demand than it can serve, but the inability to fulfill demand also places limits on revenue. It creates pressure to bring capacity online faster without compromising service quality.
For customers, the difference between a promising AI service and a dependable one often comes down to what happens during real work. Does the application respond consistently? Can the business maintain access when usage surges? Does the service integrate with existing databases, security controls, identity systems, and compliance requirements?
Microsoft has to answer those questions at enterprise scale.
The company's previous guidance projected Azure growth of approximately 45% in constant currency for the first quarter of fiscal 2027. Reuters' pre-earnings estimates put expected growth at approximately 47.7%, with Azure revenue of around $33.06 billion. These figures are not directly interchangeable because one is company guidance expressed in constant currency and the other is an analyst estimate using its own reporting assumptions.
That distinction is worth keeping in mind when assessing whether Azure has beaten expectations. A percentage that looks higher on the surface may be based on different assumptions.
Source: Microsoft's Q4 earnings-call transcript and Reuters' October 27 pre-earnings analysis.
I also think the new disclosure of Azure revenue in absolute dollars will make the discussion more useful. Growth percentages tell us how fast a business is expanding, but the revenue base helps show the scale of that expansion.
A large business growing quickly can add substantial revenue in a single quarter. That makes it easier to connect the growth story with the resources being invested to support it.
Microsoft's reporting changes deserve attention
Microsoft has also introduced a revised reporting structure that places cloud, AI, and business software under a segment called Agents and Infra, while Windows, gaming, and advertising fall under Devices and Consumer.
The new structure reflects the company's view that software, AI agents, cloud services, and computing infrastructure are becoming increasingly connected. It also means readers will need to pay attention to how the new segments are defined before comparing them with older figures.
Source: Microsoft's 2026 annual report.
I find this reporting change particularly relevant because AI agents are moving beyond simple question-and-answer interactions. They can be designed to carry out multi-step tasks, interact with software, and help people complete work over longer periods.
Microsoft has been developing this direction across its productivity and developer products. In its FY2026 reporting, the company said Microsoft 365 Copilot had surpassed 30 million paid seats. Its annual report also cited 50 million GitHub Copilot users.
Those figures show meaningful product reach, but they need to be interpreted correctly. Paid seats are a useful adoption measure; user counts show the scale of use. Neither automatically reveals how frequently customers use the tools, whether the tools improve productivity, or whether the resulting benefits justify their costs.
A business may purchase software for thousands of employees, but its long-term decision to renew will depend on whether the software becomes valuable enough to remain part of daily work.
That is where Microsoft's AI story must move beyond user counts.
If AI agents help employees complete complex tasks more quickly, reduce repetitive work, improve software development, or make business information easier to use, customers may have a reason to expand their use. If those improvements are difficult to measure, inconsistent, or outweighed by the cost and complexity of deploying the tools, adoption could become less straightforward.
The new reporting structure may help observers understand how Microsoft's software and infrastructure businesses are evolving. Yet the most useful evidence will still come from actual customer demand, recurring revenue, margins, and clear explanations of what the company is seeing in production.
Capital expenditure: Microsoft's biggest operational question
Microsoft's infrastructure spending is another major part of the story.
During the fourth quarter of fiscal 2026, the company reported capital expenditure of approximately $41 billion, including finance leases. Roughly two-thirds of that spending was allocated to short-lived assets, primarily CPUs and GPUs, while the remaining amount went toward longer-lived assets.
Microsoft also reported $55.4 billion in operating cash flow and $19.6 billion in free cash flow for the quarter, reflecting the effect of higher capital expenditure.
Source: Microsoft's FY2026 Q4 earnings-call transcript.
The distinction between short-lived and long-lived assets deserves some attention. Servers and other computing equipment can have a different replacement cycle from the buildings and physical facilities that house them. A company must therefore plan for both the immediate cost of deploying new compute resources and the continuing cost of maintaining and replacing its infrastructure.
Microsoft's capital expenditure reporting also comes with an accounting detail. The company said it was extending the estimated useful lives of data centers and office buildings from 15 to 25 years beginning in fiscal 2027. It explained that this would affect the timing of depreciation and shift some future data-center commitments from finance leases to operating leases, which are treated differently in reported capital expenditure.
Microsoft said the accounting change would have only a minimal benefit to fiscal 2027 operating income. Its conference-call commentary indicated that the change affects how some spending appears in capital expenditure figures, rather than eliminating the underlying economic cost of operating the infrastructure.
This is worth understanding because a change in accounting presentation should not be mistaken for a change in how much physical infrastructure a company actually needs.
Microsoft also expects capital expenditure in fiscal 2027 to grow year over year. Its previous guidance projected more than $50 billion in first-quarter capital expenditure, including the lease-classification effect.
The central question isn't whether Microsoft should invest in infrastructure. Without sufficient capacity, it cannot serve customers or support the AI services it is developing.
The question is whether this investment can support a healthy business over time.
Are new servers being used efficiently? Is newly available capacity generating revenue quickly? Is customer demand broad enough to support continued investment? Are cloud margins stabilizing as the business expands? Can the company maintain reliable services while bringing large amounts of new infrastructure online?
Those questions connect capital expenditure to execution. More infrastructure is useful when customers need it and when the company can operate it efficiently. The same expenditure becomes harder to justify if demand slows, costs rise sharply, or capacity takes longer than expected to generate revenue.
Reliability, security, and the ordinary experience of using AI
Earnings reports tend to concentrate on the financial measures that can be placed in tables. The user experience behind those figures receives less attention, even though it can determine whether customers continue using a service.
Consider a company that uses an AI assistant to help employees search internal documents, write software, or prepare reports. If the system is unreliable, gives inconsistent answers, or struggles under heavy workloads, employees may stop depending on it.
The same principle applies to cloud infrastructure. Enterprises need services that remain available, protect their data, and perform predictably. For a company running customer-facing applications, an outage or a serious security problem can have consequences far beyond the cost of the cloud bill.
AI can make these operational requirements more demanding. Systems may need to process sensitive information, interact with other software, or generate responses that influence important decisions.
Both Google and Microsoft have extensive experience operating large digital platforms, but scaling new AI workloads introduces additional technical and operational challenges. The performance of an established service doesn't automatically guarantee that every new workload will have the same reliability or economics.
Infrastructure also depends on physical resources. Microsoft identifies land, energy, networking equipment, servers, supply chains, and regulatory approvals as important factors in expanding its data-center capacity. It warns in its annual filing that power constraints, permitting delays, component shortages, and community opposition can delay projects or increase costs.
Source: Microsoft's FY2026 Form 10-K.
This is an important reminder that AI infrastructure isn't simply software running somewhere out of sight. It requires buildings, electricity, cooling, hardware, connectivity, and people who can maintain it.
The ability to manage those dependencies may become as important to sustained growth as the ability to develop a more capable model.
For customers, reliability isn't an abstract financial measure. It's whether the tools they depend on work when they need them. For the companies operating those tools, reliability is part of the product itself.
Regulation and trust: the longer-term questions surrounding Google
Google also faces legal and regulatory questions that matter to the future of its search and advertising businesses.
In its June 2026 quarterly filing, Alphabet described the continuing legal proceedings related to U.S. antitrust cases involving Search and advertising technology. The company reported that a final judgment in the U.S. Search case had been entered in December 2025, imposing restrictions on certain distribution practices and requiring the sharing of specified search data and the offering of syndication services to certain competitors. Alphabet said it appealed the judgment in January 2026.
The company also described separate legal proceedings related to its advertising technology business and other regulatory matters.
Source: Alphabet's Q2 2026 Form 10-Q.
These matters don't determine whether Google will beat the upcoming quarterly estimate. Their significance is broader: they can influence how the company distributes its products, conducts business with partners, and develops its search and advertising ecosystem.
AI adds another layer to the discussion. As Search becomes more capable of generating direct answers, questions arise about the visibility of publishers, how information sources are presented, and who captures the economic value created by content.
Google's long-term success will depend partly on whether it can improve the user experience while maintaining relationships with businesses, publishers, and other participants whose content and services help make Search useful.
Trust can be difficult to measure in a quarterly report, but it can shape long-term behavior. Users need confidence that the information they receive is useful. Businesses need a clear understanding of how their advertising works. Publishers need a reason to keep producing material that helps people find answers.
A strong earnings result may show that the current model is working financially. It won't resolve every question about how that model should evolve.
What would count as a meaningful earnings beat?
I don't think there's a single number that can settle this question for either company. The most useful judgment will come from looking at several indicators together.
First, the quality of revenue growth. Did the companies beat expectations because their core businesses performed well, or did the result rely heavily on unusual accounting items? For Google, the relationship between Search, YouTube, and Cloud growth deserves close attention. For Microsoft, Azure and other cloud services will be important, but the broader software business also matters.
Second, demand translated into revenue. Google's Cloud backlog and Microsoft's comments about capacity constraints suggest that both businesses have customers seeking additional computing resources. The next step is to understand how quickly those commitments and requests are becoming revenue. A backlog provides visibility, but delivery and customer usage determine when it becomes recognized income.
Third, margins and cash flow. Revenue growth can look impressive while the cost of serving customers rises. Google should be judged in the context of its expanding infrastructure spending and recent negative quarterly free cash flow. Microsoft's cloud margins, capital expenditure, and ability to generate operating cash flow will help show how its expansion is affecting the business.
Fourth, the outlook. A quarterly beat is backward-looking. Management's guidance, commentary on capacity, and explanation of customer behavior may tell us more about how sustainable the result is. If demand remains strong but capacity is difficult to expand, the company faces a different challenge from one that has abundant capacity but weakening customer demand.
Fifth, transparency. Microsoft is introducing a new reporting structure and is expected to disclose Azure's quarterly revenue in dollars, while Alphabet's AI-driven search changes make it particularly important to understand how its core business is performing. Clear definitions and consistent reporting make it easier to judge whether progress is genuine and repeatable.
There is also a final distinction worth remembering: beating expectations doesn't mean that every underlying issue has disappeared. Similarly, missing a consensus figure by a small amount doesn't necessarily mean the business is deteriorating. Estimates are forecasts, and forecasts can be wrong.
The more useful question is what the results reveal about the direction of the business.
The competition for AI revenue is also a competition for time and resources
Google and Microsoft are both investing in a future in which AI becomes part of everyday digital activity. But that future is being built under practical constraints.
Computing capacity takes time to deploy. Advanced hardware costs money. Electricity and data-center space must be available. Enterprise customers need tools that fit their existing systems. Developers need dependable platforms. And users need reasons to keep returning after the initial novelty of AI has faded.
These demands are connected.
A more capable model may attract users, but the company must still serve those users efficiently. A cloud platform may win a large contract, but it must deliver the contracted service reliably. An AI assistant may demonstrate impressive results in a trial, but the real test comes when employees use it repeatedly for ordinary work.
This is why I find the relationship between infrastructure investment and customer adoption so important. The industry cannot judge progress only by the number of models released or the size of the data centers being built. It also needs evidence that people are using these systems for tasks that matter, and that the resulting services can be operated sustainably.
Google's integrated position across Search, models, custom chips, and cloud services gives it multiple ways to connect AI development with existing products. Microsoft's established presence in cloud infrastructure, software, and developer tools gives it another route to enterprise adoption.
Both approaches have strengths. Both also require continued execution.
A business can invest heavily and still need time to discover which products customers value most, which workloads are commercially viable, and which services can grow without becoming too expensive to run.
That process rarely unfolds in a perfectly straight line.
What this session can tell us about the next stage of AI
The October 28 earnings reports will offer a useful snapshot of an industry that is still learning how to turn AI capabilities into sustainable business activity.
For Google, the central issue is whether cloud growth can continue while its search business adapts to AI-driven changes in how people find and consume information. Its results should help clarify the balance between the strength of its established advertising business, the growth of enterprise cloud services, and the cost of building additional AI capacity.
For Microsoft, the key question is whether rising demand for Azure and AI-enabled software can support continued growth while the company expands infrastructure and manages its costs. Azure's absolute revenue, cloud performance, capital expenditure, and management's explanation of demand will be important pieces of that picture.
The two companies don't need to follow identical paths. What matters is whether the choices each company makes are producing useful services, dependable revenue, and a financial structure that can support continued investment.
I would also be cautious about treating one quarter as a final verdict. Large infrastructure projects take time, enterprise adoption develops at different speeds, and customer behavior can change as products mature. One set of results can provide important clues without resolving the entire question.
The most informative outcome may be the one that helps us understand not only how much these companies earned, but also what they had to spend to earn it and what they expect customers to need next.
A personal reflection on what I'm watching
I've come to see AI earnings as a test of something more ordinary than the headlines often suggest.
People keep using a tool when it helps them do something valuable. Businesses keep paying for a service when it solves a real problem, works consistently, and fits within their budget. Developers build on infrastructure when they can trust it to support their applications without adding unnecessary complexity.
Those habits don't develop simply because a company has invested billions of dollars or introduced a more advanced model. They develop through repeated experiences, when a service works well enough that people begin to depend on it.
That's what makes this earnings season interesting to me. Google and Microsoft already have enormous reach, established products, and the resources to invest in AI at a scale that few companies can match. Their challenge now is to connect those advantages with long-term customer value.
A strong quarter would provide evidence that the connection is taking shape. It would not prove that every investment will succeed, that infrastructure costs will always remain manageable, or that today's patterns of AI adoption will continue unchanged.
I'm watching for the quieter signs: whether cloud demand converts into revenue, whether AI features become part of people's normal routines, whether the cost of serving them becomes more manageable, and whether the companies remain transparent about the trade-offs involved.
The question I keep returning to is simple. When AI becomes an ordinary part of work and daily life, will people trust these systems enough to rely on them, and will the companies providing them be able to sustain that trust while running a viable business?
The earnings figures will give us part of the answer. What customers do with these products over the months and years ahead will tell us much more.