AWS CEO Matt Garman isn’t worried about an AI bubble

Amazon Web Services chief executive Matt Garman, speaking on an Andreessen Horowitz podcast, pushed back firmly against fears of an AI spending bubble, arguing that infrastructure investment is broadly diversified and that core cloud workloads — compute, storage and inference — are built to last, underpinning what he described as a durable growth cycle for the cloud industry.

The remarks land at a moment when Amazon.com, Inc. (AMZN) is riding a wave of investor confidence: the company's second-quarter 2026 results showed AWS growing at 37% year over year, far ahead of the 31% consensus estimate, and total revenue of $200.61 billion beating the $197.03 billion forecast. For markets and investors watching whether the AI infrastructure boom has legs, Garman's argument offers one of the most direct rebuttals yet from inside the hyperscaler world.

What happened

Matt Garman, chief executive of Amazon Web Services, made the case against an AI bubble in a conversation published on the Andreessen Horowitz podcast. His central argument rested on two pillars: first, that spending on AI infrastructure is not concentrated in a single bet but spread across a wide range of use cases and customers; second, that the workloads driving that spending — compute, storage and inference — are not speculative but foundational to how businesses operate in the cloud.

Garman's comments represent a significant public stance from the leader of the world's largest cloud platform. Rather than hedging on the sustainability of AI-related capital expenditure, he framed the current investment cycle as structurally different from past technology bubbles, where spending was concentrated and the underlying demand was thin.

The numbers behind the confidence

Amazon Web Services delivered second-quarter 2026 revenue growth of 37% year over year, a figure that exceeded the 31% consensus estimate by a wide margin. Amazon.com's total revenue for the quarter reached $200.61 billion, against an estimate of $197.03 billion, while earnings per share came in at $5.75, dramatically above the $1.82 estimate. The advertising business also exceeded expectations alongside AWS, pointing to broad-based outperformance rather than a single-division surprise.

The market's reaction was swift: shares of Amazon.com jumped 11% following the second-quarter report. The trailing price-to-earnings ratio stands at 20.23, with the forward price-to-earnings ratio at 20.42, according to the latest available market data.

• AWS revenue growth: 37% year over year in Q2 2026, vs. 31% estimate
• Total Amazon revenue: $200.61 billion in Q2 2026, vs. $197.03 billion estimate
• Earnings per share: $5.75 in Q2 2026, vs. $1.82 estimate
• AMZN shares: rose 11% after the Q2 2026 earnings report
• Trailing P/E: 20.23; forward P/E: 20.42

Why it matters

The AI infrastructure debate has become one of the defining questions for technology and cloud investors in 2026. Sceptics have argued that the pace of capital expenditure by hyperscalers — Amazon, Microsoft and Alphabet among them — is outrunning near-term monetisation, raising the spectre of overcapacity and eventual write-downs. Garman's rebuttal, delivered through one of Silicon Valley's most influential venture platforms, directly challenges that narrative.

His argument about workload durability is particularly significant. Compute, storage and inference are not experimental features layered on top of existing cloud services — they are the core of what enterprise customers pay for. If those workloads are indeed expanding structurally rather than cyclically, the revenue base supporting AWS's growth is more resilient than bubble analogies suggest. That distinction matters for how the market prices not just Amazon but the entire hyperscaler cohort.

The choice of Andreessen Horowitz as the venue also carries weight. The firm is among the most prominent backers of AI startups and infrastructure companies in the United States, and its audience spans founders, limited partners and institutional observers who shape capital flows into the sector. A message delivered there reaches decision-makers across the AI supply chain.

How did markets react?

Wall Street's response to Amazon's second-quarter 2026 results was broadly positive, with the 11% share-price gain reflecting confidence in both the AWS growth trajectory and the company's wider AI strategy. The outperformance of AWS relative to consensus estimates was cited as a particular driver of that reaction, reinforcing the view that cloud demand tied to AI workloads is translating into measurable revenue rather than remaining a forward-looking promise.

Amazon's advertising segment also beat expectations in the same quarter, suggesting that the company's AI investments are beginning to generate returns across multiple business lines simultaneously. The combination of a revenue beat, an earnings-per-share figure nearly three times the estimate, and AWS growth six percentage points above consensus gave analysts a broad set of data points to work with when assessing the durability of the current cycle.

Risks and counterpoints

Not everyone shares Garman's sanguine view. The broader debate over AI infrastructure spending centres on whether enterprise adoption of AI workloads will scale fast enough to justify the capital being deployed. Critics point to the long lead times between data-centre investment and revenue recognition, and to the possibility that a handful of large customers are driving a disproportionate share of current demand — a concentration risk that diversification arguments do not fully address.

There is also the question of competition. Microsoft Azure and Alphabet's Google Cloud are investing at comparable rates, meaning that even if overall cloud demand holds up, pricing pressure and market-share battles could weigh on margins across the sector. Garman's confidence in workload durability does not resolve the question of which platform captures the largest share of that demand over time.

Macroeconomic conditions add a further layer of uncertainty. Enterprise technology budgets are sensitive to interest rates and corporate earnings cycles, and a deterioration in either could slow the pace at which businesses migrate workloads to the cloud or expand their AI usage — regardless of the structural case for those workloads.

What to watch next

The next concrete data point for Amazon.com and AWS will be the company's third-quarter 2026 earnings report, scheduled for 29 October 2026.

• 29 October 2026: Amazon.com Q3 2026 earnings report — the next AWS growth figure and EPS print
• Ongoing: enterprise AI workload adoption trends, particularly in compute, storage and inference, as a gauge of structural versus cyclical demand
• Ongoing: capital expenditure announcements from rival hyperscalers Microsoft Azure and Google Cloud, which will shape the competitive and pricing environment for AWS
• Ongoing: macroeconomic data from the United States, including corporate earnings and interest-rate signals, which influence enterprise cloud budgets