Recommended Free Tools
Amazon did plan to spend roughly $100 billion in capital expenditures during 2025, but it did not create a separately disclosed $100 billion AI budget. The forecast covered Amazon’s entire business. Most of the money was expected to fund technology infrastructure for AWS, increasingly driven by artificial-intelligence demand, while some also supported fulfillment capacity and other long-lived assets.
Amazon ultimately reported $128.3 billion in cash capital expenditures for 2025, compared with $77.7 billion in 2024. For investors, the central question is not simply how much Amazon spent, but whether AWS can turn that infrastructure into durable revenue, profit, and acceptable returns.
As an Amazon Associate I earn from qualifying purchases.
What Amazon actually announced
In February 2025, Amazon said it expected approximately $100 billion in capital expenditures across the company during the year. Chief Executive Andy Jassy framed the spending as a response to a major technology shift: the rapid growth of generative AI and the computing capacity required to train and run increasingly capable models.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Amazon said the majority of its capital spending would go toward technology infrastructure, particularly infrastructure supporting AWS growth. That is why the announcement was widely described as a $100 billion AI investment. The shorthand captures the strategic direction, but not the accounting reality.
#1 Best Overall
- Your favorite music and content – Play music, audiobooks, and podcasts from Amazon Music, Apple Music, Spotify and others or via Bluetooth throughout your home.
- Alexa is happy to help – Ask Alexa for weather updates and to set hands-free timers, get answers to your questions and even hear jokes. Need a few extra minutes in the morning? Just tap your Echo Dot to snooze your alarm.
- Keep your home comfortable – Control compatible smart home devices with your voice and routines triggered by built-in motion or indoor temperature sensors. Create routines to automatically turn on lights when you walk into a room, or start a fan if the inside temperature goes above your comfort zone.
- Do more with device pairing – Fill your home with music using compatible Echo devices in different rooms, or create a home theatre system with Fire TV.
- Say goodbye to drop-offs and buffering - With eero Built-in, Echo Dot doubles as a mesh wifi extender, adding up to 1,000 sq. ft. of wifi coverage to your existing eero network.
The total was a companywide capex forecast, not an itemized AI-only fund. It included data centers, servers, networking equipment, chips, power and cooling systems, and some fulfillment infrastructure. Amazon’s public filings do not identify a precise, audited percentage labeled “AI spending.”
Capex, operating expenses, investments, and customer commitments are different
Several kinds of spending are easy to blend together in AI coverage:
- Capital expenditures: Money spent on long-lived assets such as data centers, servers, networking equipment, custom chips, and fulfillment buildings. These costs are generally capitalized and depreciated over time.
- Operating expenses: Recurring costs including employee compensation, electricity, software development, research, leases, and maintenance.
- Strategic investments: Equity or convertible-note investments in companies such as Anthropic. These are not the same as building Amazon’s own infrastructure.
- Customer commitments: Contracts or usage commitments under which AWS customers agree to buy cloud services. They can support future AWS revenue but are not Amazon’s own capex.
This distinction matters. Saying that Amazon “spent $100 billion on AI” overstates what the company disclosed. A more accurate description is that Amazon launched a roughly $100 billion-plus companywide investment program, with most capital directed toward AWS technology infrastructure that supports AI and conventional cloud workloads alike.
Amazon’s final 2025 spending figure was higher than the forecast
Amazon’s later 2025 Form 10-K reported $128.3 billion in cash capital expenditures, up from $77.7 billion in 2024. The company said most technology-infrastructure spending supported AWS growth, while some capital spending went toward fulfillment capacity.
The spending was already accelerating during the year. Amazon reported $31.4 billion of cash capex in the second quarter of 2025 and $55.6 billion during the first six months, primarily reflecting technology infrastructure and fulfillment capacity, according to its Q2 filing.
The difference between approximately $100 billion and $128.3 billion illustrates why forecasts and reported results should not be treated as interchangeable. The first figure was management’s expectation in February; the second was the eventual full-year cash-spending figure. Neither figure represents AI-only spending.
Where the money goes
Data centers and the physical AI stack
AI infrastructure is not just a collection of expensive graphics processors. Large model-training and inference systems require clusters of servers connected by high-speed networking, along with storage, memory, power, cooling, backup systems, and software.
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Scan for outdated or missing drivers - takes under a minute3Clear out junk files and repair common Windows errorsBuilding that capacity also requires land, permits, fiber connections, electricity generation and transmission, construction labor, and equipment supply chains. A data center can take years to plan and build, so cloud providers often need to commit capital before customer demand is fully realized.
Rank #2
- Your favorite music and content – Play music, audiobooks, and podcasts from Amazon Music, Apple Music, Spotify and others or via Bluetooth throughout your home.
- Alexa is happy to help – Ask Alexa for weather updates and to set hands-free timers, get answers to your questions and even hear jokes. Need a few extra minutes in the morning? Just tap your Echo Dot to snooze your alarm.
- Keep your home comfortable – Control compatible smart home devices with your voice and routines triggered by built-in motion or indoor temperature sensors. Create routines to automatically turn on lights when you walk into a room, or start a fan if the inside temperature goes above your comfort zone.
- Do more with device pairing – Fill your home with music using compatible Echo devices in different rooms, or create a home theatre system with Fire TV.
- Say goodbye to drop-offs and buffering - With eero Built-in, Echo Dot doubles as a mesh wifi extender, adding up to 1,000 sq. ft. of wifi coverage to your existing eero network.
For AWS, the same facilities may support several types of workloads:
- Training generative-AI models.
- Running AI inference when users submit prompts.
- Hosting databases, storage, and enterprise applications.
- Running conventional computing workloads unrelated to generative AI.
- Providing networking, security, and other cloud services.
Custom chips: Trainium, Inferentia, and Graviton
Amazon is also trying to control more of the computing stack through its own silicon:
- Trainium is designed for AI training and inference.
- Inferentia focuses on inference workloads.
- Graviton is an Arm-based CPU family for general-purpose cloud computing.
Custom chips can potentially improve performance, availability, or unit economics for supported workloads. But they are not automatically cheaper or better than third-party accelerators. The result depends on the model, software optimization, utilization, instance pricing, region, and the customer’s engineering requirements.
Free tools Windows power users keep installed
One-click scans. No signup required.
Amazon later said Trainium and Graviton had surpassed a combined annualized revenue run rate of $10 billion, with triple-digit year-over-year growth, in its Q4 2025 results. That is evidence of commercial demand for Amazon-designed chips, but it does not establish that every AI infrastructure investment is profitable.
NVIDIA and other third-party accelerators
AWS is not replacing outside chip suppliers with Amazon silicon overnight. It also offers instances using NVIDIA and other vendors’ hardware. Amazon’s Q2 2025 results highlighted the general availability of EC2 instances powered by NVIDIA Grace Blackwell superchips.
Offering multiple accelerator types gives customers a choice, but it also increases the complexity and cost of AWS’s infrastructure, software support, and capacity planning.
Project Rainier and Anthropic’s role
Anthropic is an important part of Amazon’s AI strategy. Amazon has invested billions in the AI-model company and uses it as an anchor customer for AWS infrastructure, including large-scale Trainium systems associated with Project Rainier.
Amazon’s 2025 annual filing reported $2.7 billion in additional investment activity in Anthropic during 2025. That investment is separate from Amazon’s companywide capex.
Rank #3
- Your favorite music and content – Play music, audiobooks, and podcasts from Amazon Music, Apple Music, Spotify and others or via Bluetooth throughout your home.
- Alexa is happy to help – Ask Alexa for weather updates and to set hands-free timers, get answers to your questions and even hear jokes. Need a few extra minutes in the morning? Just tap your Echo Dot to snooze your alarm.
- Keep your home comfortable – Control compatible smart home devices with your voice and routines triggered by built-in motion or indoor temperature sensors. Create routines to automatically turn on lights when you walk into a room, or start a fan if the inside temperature goes above your comfort zone.
- Do more with device pairing – Fill your home with music using compatible Echo devices in different rooms, or create a home theatre system with Fire TV.
- Say goodbye to drop-offs and buffering - With eero Built-in, Echo Dot doubles as a mesh wifi extender, adding up to 1,000 sq. ft. of wifi coverage to your existing eero network.
A later announcement added another potentially confusing figure: Amazon said it would invest $5 billion immediately in Anthropic, with the possibility of up to $20 billion more. Anthropic, in turn, committed to spending more than $100 billion over ten years on AWS technologies and securing up to five gigawatts of Amazon Trainium capacity. That was Anthropic’s planned AWS spending commitment—not Amazon’s 2025 capex and not money Amazon had already recognized as revenue.
In short:
- Amazon’s roughly $100 billion figure referred to its projected 2025 companywide capex.
- Amazon’s Anthropic investment was a separate strategic investment.
- Anthropic’s more-than-$100 billion figure referred to a planned ten-year commitment to AWS technologies.
How AWS expects to make the investment pay
AWS can monetize AI infrastructure through more than raw access to accelerators. The broader revenue opportunity includes:
- EC2: Compute instances using NVIDIA, Trainium, Inferentia, and general-purpose processors.
- Amazon Bedrock: Managed access to foundation models and tools for building generative-AI applications.
- Amazon SageMaker AI: Services for developing, training, tuning, deploying, and monitoring machine-learning models.
- Amazon Q Business and Q Developer: Enterprise and software-development assistants.
- Supporting services: Storage, networking, databases, security, monitoring, and data-transfer services consumed by AI applications.
- Large customer commitments: Long-term arrangements with AI companies and enterprises that can provide visibility into future cloud usage.
Amazon reported that AWS reached a $142 billion annualized revenue run rate in the fourth quarter of 2025, with revenue growing 24% year over year, according to Jassy’s 2025 shareholder letter. Those figures show strong AWS demand. They do not, by themselves, reveal the return on each data center, chip program, or AI customer.
Investors should separate four questions:
- Revenue growth: Are customers buying more AWS services?
- Profit growth: Is AWS retaining attractive operating profit after hardware, power, staffing, and depreciation costs?
- Free cash flow: How much cash remains after the enormous capital program?
- Return on invested capital: Does the infrastructure base produce returns that justify its cost and risk?
The investment case for spending so aggressively
Amazon’s logic is straightforward. AI workloads require unusually large amounts of computing capacity, memory, networking, electricity, and cooling. AWS must add capacity faster than ordinary cloud growth would require if it wants to serve model developers and large enterprises.
Building ahead of demand can help AWS secure major workloads, offer customers reliable capacity, and prevent them from moving to Microsoft Azure, Google Cloud, Oracle Cloud, specialized GPU providers, or their own data centers. Amazon also has an incentive to develop custom silicon so it is not entirely dependent on outside accelerator suppliers.
Jassy has described the opportunity as a major technology inflection point and argued that Amazon needs to “bet big” when it identifies one. The strategy could strengthen AWS’s position if AI becomes a large, durable source of cloud consumption.
The risks for Amazon and its investors
Demand may disappoint
AI adoption could grow more slowly than Amazon expects, or customers could discover that some workloads are too expensive to run at scale. Underused servers and data centers would still create depreciation, maintenance, and energy costs.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →Hardware can become obsolete quickly
AI accelerators evolve rapidly. If a newer generation delivers substantially better performance or lower cost, older equipment may produce weaker returns than expected. Amazon could face accelerated depreciation or write-down risk if assets become economically outdated before the end of their planned useful lives.
Rank #4
- Meet Echo Dot Max: Experience rich room-filling sound that automatically adapts to your space and fine-tunes playback. Features a built-in smart home hub and Omnisense technology for highly personalized experiences.
- Music to your ears: With nearly 3x the bass versus Echo Dot (2022 release), it fits beautifully in any space, delivering your personal sound stage with deep bass and enhanced clarity. Listen to streaming services, such as Amazon Music, Apple Music, Spotify, and SiriusXM. Encore!
- Do more with device pairing: Connect compatible Echo smart speakers and smart displays in different rooms, or pair with a second Echo Dot Max to enjoy even richer sound
- Simple smart home control: Set routines, pair and control lights, locks, and thousands of smart home devices that work with Alexa without needing a separate smart home hub. With Omnisense technology, you can activate routines via temperature or presence detection.
- Say goodbye to drop-offs and buffering - With eero Built-in, Echo Dot Max doubles as a mesh wifi extender, adding up to 1,000 sq. ft. of wifi coverage to your existing eero network.
Power and permitting can constrain growth
Data centers need reliable electricity and may face grid constraints, long interconnection queues, local opposition, water-use concerns, and lengthy permitting processes. Money alone cannot guarantee that capacity becomes operational on schedule.
Custom chips require an ecosystem
Designing a chip is only part of the challenge. Amazon must also support compilers, frameworks, libraries, developer tools, manufacturing, and customer migration. A chip can be technically competitive yet commercially limited if developers find the software ecosystem less familiar than NVIDIA’s.
Large customers create concentration risk
AI companies can generate substantial cloud demand, but a major customer may change providers, build its own infrastructure, renegotiate terms, or experience financial difficulties. A large commitment is not the same as cash revenue already collected.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Growth can pressure margins
AWS may report rapid revenue growth while near-term margins face pressure from depreciation, power, chip purchases, construction, and staffing. The investment thesis depends on utilization and pricing eventually rising enough to earn attractive returns on the enlarged asset base.
Competition is intense
Amazon competes with Microsoft Azure and Google Cloud, as well as Meta’s infrastructure investments, Oracle Cloud, CoreWeave and other specialist GPU clouds, and customers building private capacity. Competition can limit pricing power or force providers to keep investing simply to maintain their position.
Amazon’s AI investments, cloud contracts, chip strategy, and relationships with AI-model companies may also attract regulatory or antitrust scrutiny. That possibility should not be confused with a specific legal finding.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What the spending means for investors
The spending program creates a clear trade-off: higher capex may support stronger long-term AWS growth, but it reduces near-term free cash flow and raises the amount of future demand needed to reach attractive returns.
Investors evaluating Amazon should monitor:
- AWS revenue growth and operating margin.
- Cash capex compared with operating cash flow.
- Free-cash-flow conversion.
- Depreciation growth and the useful lives of infrastructure assets.
- Utilization and availability of new AI capacity.
- Adoption and revenue from Trainium and Inferentia.
- Customer concentration and the quality of long-term commitments.
- Whether AI expands total AWS consumption or merely shifts workloads between AWS services.
- Whether Amazon continues raising capex forecasts faster than revenue forecasts.
Amazon later said it expected approximately $200 billion of capital expenditures during 2026 across AI, chips, robotics, satellites, and other areas. The figure was subsequently reported by the Associated Press as having risen to $220 billion. Those are 2026 developments, not revisions to Amazon’s original 2025 announcement. See Amazon’s 2026 results announcement and the AP report.
Best Value
- MEET ECHO SPOT - A sleek smart alarm clock with Alexa and big vibrant sound. Ready to help you wake up, wind down, and so much more.
- CUSTOMIZABLE SMART CLOCK - See time, weather, and song titles at a glance, control smart home devices, and more. Personalize your display with your favorite clock face and fun colors.
- BIG VIBRANT SOUND - Enjoy rich sound with clear vocals and deep bass. Just ask Alexa to play music, podcasts, and audiobooks. See song titles and touch to control your music.
- EASE INTO THE DAY - Set up an Alexa routine that gently wakes you with music and gradual light. Glance at the time, check reminders, or ask Alexa for weather updates.
- KEEP YOUR HOME COMFORTABLE - Control compatible smart home devices. Just ask Alexa to turn on lights or touch the screen to dim. Create routines that use motion detection to turn down the thermostat as you head out or open the blinds when you walk into a room.
What the spending means for AWS customers
More infrastructure can improve capacity and product choice, but Amazon’s corporate investment does not guarantee lower prices or better economics for every customer.
Businesses considering AWS AI services should ask:
- Are the required GPU or Trainium instances available in the necessary region?
- Does the chosen model support the intended AWS hardware?
- Will storage, retrieval, networking, logging, and data-transfer costs materially affect total cost?
- Is the workload predictable enough for reserved capacity or a long-term commitment?
- Is AWS lock-in acceptable?
- Does the team need a managed model API or control over training and deployment?
- Are data residency, compliance, identity, and logging requirements satisfied?
Amazon Bedrock generally suits teams that want managed access to foundation models and do not want to manage model infrastructure. Its pricing is consumption-based and varies by model, modality, usage, and inference tier; see the official Bedrock pricing page.
Amazon SageMaker AI is better suited to data-science and machine-learning teams that need more control over training, tuning, deployment, and monitoring. Charges can apply to compute, storage, processing, deployment, and related services; see SageMaker’s pricing page.
Amazon Q Business is designed for enterprise assistants connected to internal information and systems. AWS lists Lite at $3 per user per month and Pro at $20 per user per month on its pricing page, along with index-capacity charges. Amazon Q Developer offers coding and AWS-development assistance, with free and paid tiers whose current terms should be checked before purchase. Both are described on the official Amazon Q pricing page.
Alternatives include Microsoft’s Azure AI Foundry, Google Vertex AI, Anthropic’s direct API, the OpenAI API, and specialist GPU clouds. The best choice depends on model access, software compatibility, regional capacity, support, data requirements, and total cost—not on Amazon’s spending headline alone.
The bottom line on Amazon’s AI spending plan
Amazon’s $100 billion 2025 plan was real, but the headline needs an accounting footnote. It described a companywide capital-expenditure forecast, with most spending directed toward AWS technology infrastructure that supported both AI and broader cloud growth. Amazon’s final reported cash capex was higher, at $128.3 billion.
The bet is strategically ambitious: Amazon is investing in data centers, networking, third-party accelerators, custom chips, managed AI software, and relationships with model companies such as Anthropic. The payoff will depend on whether AI demand remains durable, AWS can monetize capacity quickly, and the resulting revenue and profits justify the added depreciation, power costs, competitive pressure, and execution risk.
Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallQuick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
