Amazon’s $220bn AI bet puts cash flow under pressure as spending outpaces returns

Amazon’s plan to spend about $220 billion on capital investments in 2026 is putting pressure on its cash flow as the technology giant races to expand its artificial intelligence infrastructure to meet growing demand for cloud computing services.

The company’s free cash flow swung to a $7.6 billion outflow in the 12 months to June 30, 2026, compared with a positive $18.2 billion in the corresponding period a year earlier, highlighting the financial cost of its accelerating investment programme.

Amazon raised its 2026 capital expenditure forecast by 10 percent in July, from approximately $200 billion to $220 billion, citing higher memory chip costs and demand for AI and cloud infrastructure.

The spending covers the company’s broader capital investment programme, including infrastructure beyond AI, rather than AI alone.

The spending increase comes as major technology companies pour unprecedented sums into data centres, specialised chips, servers and networking equipment to secure a competitive advantage in AI.

For Amazon, the challenge is to convert that investment into sustained revenue and cash generation before the costs begin to weigh more heavily on its financial flexibility.

Amazon’s cash flow pressure reflects a timing gap between building the infrastructure and earning revenue from it.

Andy Jassy, CEO of Amazon, has said data centres require substantial investment roughly two years before they begin operating, while the company expects the facilities to generate revenue over a much longer period.

AWS growth offers a defence for the spending

Despite the cash flow strain, Amazon’s cloud computing business is delivering strong growth, providing evidence that at least part of its AI investment is generating commercial returns.

Amazon Web Services (AWS) reported second-quarter revenue of $42.2 billion, up 36.7 percent from a year earlier, its fastest growth in 18 quarters.

Operating income at the cloud division increased to $16.6 billion from $10.2 billion in the same quarter of 2025, according to Amazon’s earnings report.

The company also reported an AWS contract backlog of $496 billion at the end of June, up from $364 billion in the previous quarter. The backlog represents contracted business yet to be recognised as revenue, providing visibility into future demand, although it does not guarantee that every dollar will translate into profit or immediate cash flow.

Jassy has argued that Amazon is investing to meet existing customer demand rather than building infrastructure speculatively. He said the company expected capacity constraints to persist through 2027, with demand for cloud infrastructure extending into 2028.

The company’s strategy rests on the expectation that revenue from newly deployed infrastructure will eventually outpace additional capital spending, allowing cash generation to recover.

However, the timing remains critical as capital expenditure is incurred before new data centres become operational, while the revenue generated by those facilities arrives later. Higher equipment costs can further increase the amount Amazon must commit before it begins recovering its investment.

Rising AI costs raise questions about financial sustainability

Amazon’s position reflects a broader challenge facing technology companies competing to build AI infrastructure which is strong demand does not automatically translate into strong free cash flow.

Free cash flow measures the cash remaining after capital expenditure and is an important indicator of how much a company can use for debt repayment, acquisitions, shareholder returns and other investments without raising additional financing.

Amazon’s operating cash flow increased 33 percent to $161.4 billion over the 12 months to June 30, 2026, according to its financial results. Yet free cash flow fell into negative territory, primarily because purchases of property and equipment, net of sales proceeds and incentives, increased by $66.1 billion year on year. Amazon attributed the increase primarily to AI investments.

The figures illustrates the distinction between generating cash from operations and retaining cash after funding expansion. Amazon continues to generate substantial operating cash, but its investment requirements are absorbing more of that money.

The company is also operating in an environment where access to external financing is becoming an important consideration for the technology sector.

In June, Amazon secured a $17.5 billion loan facility from a group of lenders, including Citibank, JPMorgan Chase, HSBC and Wells Fargo. The facility was for general corporate purposes, and the announcement did not establish that the entire amount would fund AI infrastructure.

On October 19, S&P Global warned that the largest cloud and AI infrastructure companies could have more AI-related liabilities than publicly disclosed because of complex financing arrangements involving data centres and smaller AI firms.

The ratings agency estimated that six major technology companies, including Amazon, could spend more than $7 trillion collectively on AI-related capital expenditure between 2025 and 2030, with negative free cash flow expected across the group through at least next year.

The warning points to a growing concern for investors that whether the projected returns from AI will materialise quickly enough to justify the scale of investment and the associated financing risks.

Amazon maintains that its AI investments are sustainable

Amazon has pushed back against concerns that the AI spending boom is becoming financially unsustainable.

On October 16, John Felton, AWS chief financial officer told Axios that the company’s AI infrastructure investments were sustainable and were producing meaningful innovation.

He cited AWS’s nearly $500 billion order backlog as evidence of demand from customers ranging from large enterprises and AI laboratories to small and medium-sized businesses.

The argument is that Amazon has advantages that many smaller AI infrastructure providers lack, including a diversified business, an established cloud platform, substantial operating cash flow and a large base of existing customers.

Regardless, the investment carries risks because if demand slows, customers reduce spending or competing cloud providers drive down prices, Amazon could find it harder to generate the returns needed to justify its capital commitments.

Delays in bringing infrastructure online could also postpone revenue while financing and operating costs continue.

For now, AWS’s growth suggests that Amazon is not investing in an AI market without customers. The unresolved question is how efficiently it can convert that demand into cash after accounting for the cost of building and maintaining the infrastructure.

The next test is cash generation

Amazon’s $220 billion spending plan reveals the scale of its ambition to strengthen its position in AI and cloud computing but the negative free cash flow shows that the financial benefits of the strategy have not yet kept pace with the capital being deployed.

The company’s substantial backlog and rising AWS revenue offer reasons for confidence, while its growing infrastructure costs leave little room for prolonged delays in monetisation.

For investors, the key measure will not simply be how quickly AWS grows, but whether that growth eventually produces enough additional cash to support Amazon’s investment programme without sustained pressure on its balance sheet.

As the AI race intensifies, Amazon must demonstrate that its spending can deliver durable returns, rather than simply securing more computing capacity at an expensive price.