Amazon Completes AI Infrastructure Financing with $100 Billion Program: Microsoft and Google Under Pressure

In June 2026, Amazon closed a $17.5 billion credit line with a banking consortium to further expand its AI infrastructure. The credit facility allows the company to draw down funds as needed and flexibly align the timing of capital raising with actual infrastructure expansion. In July 2026, a $25 billion bond issuance followed, which CEO Andy Jassy called the final debt capital raise for 2026.

These two financing rounds mark the conclusion of a comprehensive capital strategy that Amazon has been pursuing since November 2025. At that time, the company had already issued a $15 billion bond to strengthen AWS and AI infrastructure. Added to this in 2026 was a pound-denominated bond of £4.25 billion, which for the first time tapped the British bond market for this purpose. In total, Amazon has mobilized over $60 billion in debt capital in less than a year.

Two Million GPUs by 2028: AWS Gears Up Massively

The capital raised flows directly into expanding AWS cloud infrastructure. Amazon plans to install two million additional GPUs in 2027 and 2028 to serve AI workloads. The collaboration with NVIDIA is being expanded: Blackwell Ultra, Rubin, and Rubin Ultra GPUs are to be rolled out gradually over the coming two years. For 2026 alone, Amazon had planned around one million GPUs.

AI workloads place fundamentally different demands on data centers than traditional cloud services. They require specialized AI processors (GPUs or accelerators), enhanced cooling systems, and significantly higher power capacity. These technical particularities drive both the capital intensity and time pressure at AWS to quickly provision capacity.

Amazon raised its capital expenditure guidance for 2026 to around $220 billion by September. This amount encompasses AI and cloud infrastructure as well as logistics and manufacturing, such as a $100 million advanced manufacturing facility in Indiana. Concurrently, Amazon has shown negative free cash flow since September 2026, underscoring dependence on capital markets.

$220 Billion in Cloud Sector Debt: The Race for Capacity

Amazon is not alone in its financing strategy. The four major hyperscalers—Amazon, Microsoft, Alphabet, and Meta—have collectively issued approximately $220 billion in debt in the twelve months through September 2026. The total planned spending by the four companies for the current period is around $725 billion (midpoints of ranges) according to analyses, roughly double the spending in the prior completed fiscal year. The transition from high technology budgets to industrial-scale infrastructure buildout is now complete.

Alphabet currently leads in AI-specific annual spending, with management indicating that approximately 60 percent of funds flow into servers and 40 percent into data centers and network technology. In the second quarter of 2026, Alphabet reported $44.9 billion in capital investments. Amazon ranks second in narrowly defined AI infrastructure spending but leads in already-deployed infrastructure in the last completed fiscal year and most recent reporting quarter.

Microsoft's current annual spending is roughly five percent below Alphabet's average. About two-thirds of recent quarterly investments went into chips and shorter-lived equipment. For fiscal year 2025, Microsoft committed to approximately $80 billion for AI-capable data centers. While Meta has the lowest absolute spending of the four, it proportionally commits most heavily to AI: its planned investments represent roughly two-thirds of annual revenue, a ratio Microsoft also exhibits.

Which Cloud Providers Face Pressure

AWS defends its leadership in traditional cloud business but faces the sharpest competition in the AI segment. This area grows fastest and is equally most capital-intensive. The risk lies in insufficient data center capacity driving customers to competitors offering guaranteed availability for AI projects. Competition centers not on developing proprietary language models but on computing power to serve enterprise customers at scale.

Smaller cloud providers without comparable capital market access face structural pressure from the hyperscalers' investment offensive. Hardware partners and data center operators, however, benefit from sustained capital deployment: providers of AI accelerators, network technology, and power infrastructure record stable demand. At the same time, the risk of market consolidation rises for providers unable to keep pace.

Revenue Growth and Return Questions

In the second quarter of 2026, Amazon achieved total revenue of $200.6 billion, up 20 percent year-over-year. These figures demonstrate the underlying strength of the business despite high capital expenditures. However, investors increasingly focus on whether infrastructure investments translate into revenue and earnings growth that justify capital intensity. AWS must demonstrate that revenue and operating income grow in line with GPU capacity expansion.

The debt strategy comes under scrutiny: Amazon and other hyperscalers have issued substantial debt in a short time, which capital markets must absorb. Rising financing costs could make future funding rounds more expensive. Operational challenges also loom, such as pressure on retail margins, changing consumer spending, and rising hardware prices. These factors demand precise execution with little room for error should demand weaken.

Amazon's combination of bond markets and traditional bank financing reflects both unprecedented capital needs and management confidence in AI-driven revenue growth. Credit lines provide flexibility through optionality: Amazon pays commitment fees but draws funds only as needed. This structure enables AWS to expand data center capacity for AI services without burdening operating cash flow, and supports the competitive position in the "AI arms race" against Microsoft and Google.