Amazon is accelerating its artificial intelligence ambitions, with CEO Andy Jassy outlining a massive $200 billion investment plan focused on AI infrastructure, custom silicon, and cloud expansion in his 2025 shareholder letter.
The strategy, shared by Andy Jassy in a letter to employees and Amazon shareholder, signals a decisive shift toward becoming an AI-first company, with Amazon Web Services (AWS) emerging as the central engine powering this transformation.
$200 billion AI investment backed by real demand
Amazon expects its capital expenditure to reach about $200 billion, largely directed toward building AI data centers, expanding compute capacity, and developing proprietary chips. Jassy emphasized that this level of spending is not speculative but driven by strong and growing enterprise demand.
He noted that a significant portion of the planned infrastructure for 2026 is already backed by customer commitments, with monetization expected in 2027 and 2028. This forward visibility is helping Amazon justify one of the largest AI investment programs in the technology sector.
AWS AI business hits $15 billion annual revenue run rate
A major highlight from the disclosure is AWS’s AI business, which has crossed a $15 billion annualized revenue run rate for the first time. This represents roughly 10 percent of AWS’s overall $142 billion revenue run rate, underscoring how quickly generative AI is becoming a core revenue driver.
Jassy said AI revenue is “ascending rapidly,” adding that AWS growth would be even stronger if not for ongoing capacity constraints across the industry. AI workloads are also boosting demand for complementary services such as storage, analytics, and compute, creating a multiplier effect across AWS.
Infrastructure expansion struggles to keep up with AI demand
To address surging AI workloads, Amazon added 3.9 gigawatts of power capacity in 2025 and aims to significantly expand its infrastructure footprint by 2027. Despite this aggressive buildout, demand continues to outpace supply.
Jassy revealed that some customers have requested to secure all available capacity for certain AI chips, highlighting the intensity of competition for compute resources. The imbalance between demand and supply remains one of the biggest constraints on AWS growth.
Custom AI chips cross $20 billion revenue run rate
Amazon’s in-house silicon business is emerging as a critical pillar of its AI strategy. Products such as Graviton CPUs, Trainium AI chips, and Nitro networking components have collectively surpassed a $20 billion annual revenue run rate, doubling from previous disclosures.
These custom chips are designed to reduce reliance on external suppliers like Nvidia while lowering infrastructure costs. Jassy indicated that Amazon could eventually sell these chips or entire server racks to third-party customers, opening a new revenue stream.
AI could drive AWS toward $600 billion opportunity
Internally, Amazon sees AI as a transformational growth engine. Jassy has indicated that AI could help AWS scale toward a $600 billion annual revenue opportunity over time, significantly higher than earlier projections.
He also dismissed concerns about an AI investment bubble, stressing that demand is grounded in real-world enterprise use cases and long-term contracts rather than hype.
Cost discipline alongside AI expansion
Even as Amazon ramps up AI spending, it has taken steps to streamline operations, including reducing approximately 30,000 jobs in recent months. The move is aimed at cutting bureaucracy, improving efficiency, and reallocating resources toward high-growth AI initiatives.
AI becomes the core of Amazon’s long-term strategy
Amazon’s latest shareholder communication makes it clear that AI is no longer a side business but the foundation of its future growth. With a $200 billion investment pipeline, a fast-growing $15 billion AI revenue stream, and expanding chip and cloud capabilities, Amazon is positioning itself at the forefront of the global AI infrastructure race.
RAJANI BABURAJAN
