With $143 billion in cash and securities on hand, why is Amazon borrowing for AI?

TL;DR

· Amazon has issued approximately $25 billion in investment-grade US dollar bonds, with market demand peaking at around $62 billion.

· The borrowing is not due to a cash crunch but rather a proactive investment in AI data centers, squeezing free cash flow.

· Related Tickers: AMZN, MSFT, GOOGL, META, NVDA, data centers, power, and AI infrastructure chain.

Amazon completed around a $25 billion investment-grade US dollar bond issuance on July 7, however, as of the end of the first quarter, the company still held approximately $143.1 billion in cash and marketable securities.

These numbers may seem unusual. Why would a company with strong operating cash flow and ample current assets take on long-term debt? The answer lies not in a short-term lack of funds but in the payment cadence for AI infrastructure. Investments in data centers, chips, networks, and power capacity need to be made upfront, while revenue realization will happen over the next few years.

This is also why the market is reevaluating Amazon. In the past, investors were more accustomed to viewing AWS as a stable cash cow. Now, AI is pushing the cloud business into a heavier asset cycle. Amazon CEO Andy Jassy has referred to AI as a "once-in-a-lifetime opportunity," but analysts are more concerned about when these investments will translate back into free cash flow.

AI Factory Spends First, Earns Later

Let's first clarify three financial concepts. Operating cash flow is the money earned from daily operations, capital expenditures involve upfront investments in building data centers, buying servers, and expanding networks, while free cash flow is what's left after deducting capital expenditures.

Amazon's issue is not a depletion of operating cash flow. The company's first-quarter financials show that over the past 12 months, operating cash flow was approximately $148.5 billion, still at a high level. However, free cash flow for the same period dropped to around $12 billion, a 95% decrease year-over-year, primarily due to increased property and equipment spending.

Based on the company's prior disclosures and management guidance, capital expenditures for 2026 are estimated at around $200 billion, higher than the 2025 level of about $131 billion. Net property equipment purchases in the first quarter have already exceeded $43 billion, mainly directed towards data centers, custom chips, network equipment, and AI infrastructure.

So, $143.1 billion in cash and marketable securities does not equal "no need for financing." This portion of the assets serves as a strategic buffer for acquisitions, investments, operational turnover, and uncertainty. The AI data center is a long-term asset, better matched with long-term debt.

The Bond Market Still Believes in AWS's Credit Anchor

Another signal of this bond issuance is that buyer demand remains strong. Market orders peaked at around $62 billion, higher than the approximately $25 billion issue size. The initial price guidance for the longest, about 40-year maturity bonds, was about 145 basis points higher than U.S. Treasury bonds, roughly 1.45 percentage points.

The implication of investment-grade bonds is straightforward. The credit market is willing to lend money to Amazon for a long time at a relatively modest spread. This does not prove that AI investment will be successful, but it indicates that bond investors do not currently view the financing as a credit deterioration event.

The core supporting this is still AWS. The cloud business has high customer stickiness, with strong visibility into future contract revenue. Regulatory filings show that as of March 31, 2026, future obligations with terms exceeding one year are approximately $364 billion, mainly related to AWS. This is not current revenue, but it can provide clues to future cash flows.

The bond market is buying into a timeline: Amazon first finances the construction of AI capacity, and in the future, AWS converts this capacity into revenue and cash flow through training, inference, enterprise AI services, and in-house chip ecosystem. As long as the chain continues, the increase in debt looks more like a capital structure adjustment.

Bezos Bets on Capacity Window, Stock Market Focuses on Cash Flow

Bezos's logic is clear. AI is a platform-level opportunity, and if cloud providers miss the capacity window, the cost may be higher than a short-term decline in free cash flow. Within this framework, the approximately $200 billion in capital expenditures are a ticket to preempting the next generation of cloud demand.

This argument is grounded in reality. Generative AI training and inference both require a large amount of computing power, and the earlier customers lock in capacity, the more cloud providers need to build data centers in advance. Amazon's in-house Trainium, Graviton, and other chips are also aimed at reducing reliance on external GPUs and binding hardware and cloud services more tightly.

However, the cautionary scenarios from Bloomberg Intelligence and some Wall Street analysts cannot be ignored. They are not concerned about whether Amazon can borrow money but whether capital expenditures will continue to exceed market expectations and if free cash flow will remain under pressure from 2026 to 2027.

This is precisely the difference between stock investors and bond investors. As long as bond investors believe that Amazon can pay interest and principal, a spread of around 145 basis points is attractive. Stock investors also look at earnings per share, profit margins, and valuation multiples. If AI revenue realization lags behind depreciation and interest rate hikes, stock price pressure may arise sooner.

Cloud Giants Shifting Valuation Anchor to Capacity Payoff

The larger implication of Amazon's recent bond issuance is that the valuation anchor of tech giants is changing. In the past, cloud providers' advantage mainly came from software economies of scale, where the more customers they had, the lower the marginal cost, leading to better cash flow. AI has shifted this model back toward a heavy asset cycle.

This is not a strategic choice for Amazon alone. Other major cloud and platform companies like Microsoft, Google, Meta, and others are also increasing their AI infrastructure investment, with industry capital expenditures already reaching the hundreds of billions of dollars. The difference lies in who can translate these expenditures into steady utilization rates and high-margin revenue.

For Amazon, the debt itself is not the most dangerous variable. The key concern is whether AWS's AI-related revenue can keep up with the capital expenditure slope. If capital expenditures continue to rise over the next few quarters, and contract conversion, cloud revenue growth, and profit margins do not improve simultaneously, the market will demand a higher risk premium.

Free Cash Flow Determines Market Response to This Debt

This $25 billion bond issuance has shifted Amazon's AI investment cycle onto the balance sheet. The company can currently finance at a lower cost, and the credit market is willing to participate. However, issuing bonds only solves the funding duration matching issue and does not automatically address the capacity payoff problem.

The validation points will be reflected in the financial statements. Investors will need to see if capital expenditures continue to increase, if AWS revenue growth accelerates, if profit margins can hold, and if free cash flow can recover from its low levels. If these indicators fail to improve significantly, the market discussion on Amazon will shift from "can they build" to "can they make money from what they've built."

The cruelest aspect of the AI infrastructure cycle is that ramping up capacity too early can drag down cash flow, while having insufficient capacity may result in missing out on customer demand. Amazon has now chosen to borrow money upfront and will need to prove over the next few years that these AI factories can turn compute power into free cash flow.

Welcome to join the official BlockBeats community:

Telegram Subscription Group: https://t.me/theblockbeats

Telegram Discussion Group: https://t.me/BlockBeats_App

Official Twitter Account: https://twitter.com/BlockBeatsAsia