Amazon has raised $25 billion in the investment-grade bond market, drawing demand that reportedly climbed to about $62 billion, even though the company ended the first quarter with roughly $143.1 billion in cash and marketable securities.
The bond sale highlights a central question now facing one of the world’s largest technology companies: why borrow heavily when cash reserves remain enormous? The answer lies in Amazon’s rapidly expanding spending plans for artificial intelligence infrastructure, where the company is moving aggressively to build data centers, specialized chips, and network systems designed to support long-term cloud demand.
The offering also shows that credit markets remain highly confident in Amazon’s ability to meet future interest and principal payments. Some of the new debt matures in about 40 years, and the longest-dated portions were priced at roughly 145 basis points above comparable U.S. Treasury securities. That pricing suggests bond buyers are willing to lend to Amazon for decades at relatively modest premiums, despite the scale of the company’s spending surge.
For Amazon, the transaction provides long-term funding at a time when capital expenditures are rising much faster than near-term free cash flow. The company’s operating cash flow stood at about $148.5 billion over the previous 12 months, but free cash flow fell to around $1.2 billion after a sharp increase in spending on property, equipment, and infrastructure. Capital expenditures rose by roughly 95% over the past year, driven largely by the buildout required for artificial intelligence workloads inside Amazon Web Services.
The scale of that buildout is becoming one of the defining financial stories in the technology sector. Amazon is preparing for capital expenditures of about $131 billion in 2025 and roughly $200 billion in 2026, according to current estimates. Purchases of property and equipment already exceeded $43 billion in the first quarter, reflecting the heavy cost of expanding AI capacity before all of the related revenue arrives.
Why Amazon is borrowing despite its cash reserves
Amazon’s decision to issue debt while holding more than $140 billion in cash is not unusual for a company managing long-term assets. Data centers, networking equipment, power systems, and custom chips are not short-term expenses. They are infrastructure assets expected to support revenue over many years.
By using long-dated debt, Amazon can better match the life of its financing with the life of the assets it is building. In simple terms, the company is borrowing over decades to fund infrastructure that may also generate business for decades. This approach allows Amazon to preserve cash for other priorities, including acquisitions, research and development, supply chain needs, operating flexibility, and protection against economic uncertainty.
The bond sale also gives Amazon access to capital while demand for high-quality corporate debt remains strong. Investment-grade issuance has been robust this year, with companies taking advantage of deep liquidity in the bond market. For large technology firms with strong cash flow, dominant market positions, and long-term revenue visibility, lenders have shown a continued willingness to provide funding on favorable terms.
That support contrasts with the more cautious tone seen in parts of the stock market, where traders are focused on how quickly AI infrastructure spending can convert into revenue growth, operating profit, and free cash flow. Bond buyers are mainly assessing whether Amazon can service its debt. Stock traders are also asking whether the spending will produce returns quickly enough to support earnings expectations and share valuations.
AWS remains the center of confidence
Amazon Web Services remains the foundation of market confidence in the company’s AI spending strategy. The cloud division has long been one of Amazon’s most profitable businesses, and it is now at the center of the company’s push into artificial intelligence.
As of March 31, 2026, AWS had about $364 billion in remaining performance obligations, with most of those commitments extending beyond one year. These obligations represent contracted revenue that has not yet been recognized, offering visibility into future sales. For credit markets, that backlog is important because it provides evidence that customers are committed to long-term cloud spending.
AWS is also the business unit most directly positioned to benefit from demand for AI computing. Companies developing generative AI tools, machine learning models, enterprise automation platforms, and data-intensive applications need large amounts of computing power. Amazon is trying to ensure that AWS has enough capacity to meet that demand as customers expand from testing AI services to deploying them at scale.
Chief executive Andy Jassy has described AI infrastructure as a strategic capacity race. The logic is straightforward: if Amazon waits for demand to fully materialize before building data centers, it may lose business to competitors with available capacity. If it builds ahead of demand, it takes on higher near-term costs but may be better positioned to capture long-term cloud revenue.
That strategy carries both opportunity and risk. Building early can secure customer relationships and strengthen AWS’s competitive position. But if demand takes longer to develop than expected, the company could face a period in which depreciation, interest costs, and operating expenses rise faster than AI-related revenue.
Custom chips are part of the strategy
Amazon is also investing heavily in proprietary processors, including Trainium and Graviton. These chips are designed to reduce reliance on external semiconductor suppliers and give AWS more control over performance, cost, and system design.
Trainium is aimed at AI training workloads, while Graviton supports a broader range of cloud computing tasks. By designing more of its own hardware, Amazon can integrate chips, software, data centers, and cloud services more tightly. That integration may improve efficiency and margins over time if customers adopt the systems at scale.
The custom chip strategy is especially important because AI computing is expensive. Advanced graphics processors and related hardware remain in high demand across the industry, creating supply constraints and high input costs. If Amazon can shift more workloads to internally designed chips, it may reduce part of that cost pressure and improve the economics of AI services.
Still, the payoff is not automatic. Customers must be willing to use Amazon’s chip ecosystem, software tools must be mature, and performance must compare favorably with alternatives. The success of the strategy will depend not only on capital spending but also on execution across engineering, sales, developer adoption, and customer support.
A wider technology infrastructure boom
Amazon is not alone in this spending cycle. Microsoft, Google, and Meta are also committing enormous sums to artificial intelligence infrastructure. Together with Amazon, these four companies are expected to direct approximately $725 billion toward capital expenditures in 2026, a 77% increase from the previous year.
That level of spending marks one of the most concentrated corporate infrastructure expansions in modern business history. The largest technology platforms are moving from an era dominated by software efficiency toward one increasingly shaped by physical capacity: land, power, data centers, chips, cooling systems, fiber networks, and specialized engineering.
The shift is reshaping the digital economy. For years, cloud computing was often discussed as a high-margin, asset-light growth business, even though it always required major infrastructure. AI is making the physical side of the cloud much more visible. Training and running large models require vast computing clusters, steady power supplies, and advanced networking systems that can move data quickly and reliably.
This creates a new competitive dividing line. Companies with the ability to finance, build, and operate massive infrastructure may gain advantages over smaller rivals. At the same time, the capital intensity of the AI race raises the financial stakes. The companies spending the most must show that the revenue opportunity is large enough to justify the buildout.
Stock traders focus on timing and returns
While bond markets have responded favorably to Amazon’s debt sale, stock traders are paying close attention to the timeline for returns. The central concern is not whether Amazon can raise money or pay interest. It is whether AI spending will improve margins and free cash flow soon enough to support the company’s market value.
The gap between infrastructure spending and immediate revenue remains one of the biggest issues across the AI sector. Some AI businesses are already reporting annualized revenue run rates in the tens of billions of dollars, showing that demand is real. But that revenue is still much smaller than the hundreds of billions being spent across the industry on data centers, chips, and related infrastructure.
For Amazon, the next several quarters will be especially important. Analysts following the company expect AWS revenue growth to accelerate sharply, with some projections pointing to growth of about 34% in the second quarter and potentially 42% in the third. Results near or above those expectations would support the argument that capital spending is beginning to translate into faster sales.
A weaker result would raise fresh questions about whether Amazon is building capacity ahead of demand by too wide a margin. It could also increase concern that depreciation and interest expenses may weigh on profitability before AI revenue reaches the necessary scale.
Interest rates add another complication
The broader macroeconomic backdrop adds another layer of uncertainty. At the beginning of the year, many market participants expected the Federal Reserve to move toward interest rate cuts. Persistent inflation has complicated that outlook, and markets have considered the possibility that rates could stay elevated for longer or even rise again before year-end.
Higher rates matter for companies spending heavily on infrastructure. They increase the cost of capital, raise the hurdle rate for new projects, and can pressure equity valuations. Even for a company as financially strong as Amazon, the level of interest rates affects how traders judge long-term cash flows and future earnings.
Amazon’s recent bond pricing suggests lenders view the company as a high-quality borrower, but the macro environment still matters. If interest rates remain elevated, the cost of future borrowing could rise. If inflation pushes up construction, energy, labor, and equipment costs, the total price of the AI buildout could also increase.
At the same time, Amazon’s large cash position and strong operating cash flow give it more flexibility than most companies. The bond sale strengthens that flexibility by locking in long-term capital now, rather than forcing the company to rely only on internal cash generation during a period of unusually heavy spending.
The key test is future cash flow
The success of Amazon’s $25 billion bond sale will ultimately be measured by future cash flow recovery. The company has secured funding on terms that suggest strong confidence from credit markets. But the larger question is whether the AI infrastructure being built today will generate enough revenue and profit to justify the scale of spending.
If AWS can convert AI demand into higher growth, stronger margins, and durable cash generation, the debt sale may be seen as a well-timed financing move that helped Amazon bridge into the next phase of cloud computing. In that scenario, long-dated borrowing would support assets capable of producing returns over many years.
If revenue growth lags or margins weaken, the same spending could become a source of pressure. Free cash flow may remain constrained, depreciation may climb, and stock traders may demand clearer evidence that AI services can produce returns matching the scale of the capital deployed.
For now, the bond market has delivered a clear message: Amazon is still viewed as one of the strongest corporate borrowers in the world. The stock market’s message is more conditional. Traders want proof that the AI capacity race will not simply produce larger data centers, but also faster growth, better profitability, and a stronger cash flow profile.
The next AWS results will therefore carry significance beyond one quarterly earnings report. They will offer an early reading on whether Amazon’s aggressive infrastructure strategy is moving in step with customer demand, or whether the company is entering a longer and more expensive payoff cycle. For Amazon, Microsoft, Google, and Meta, the AI buildout has become a defining test of capital discipline in the new era of cloud computing.
To explore how AI transforms finance and trading, check out this AI in banking deep dive next.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
