AWS's New MicroVM Service Tackles AI Security Risks as Amazon Bets $200 Billion on Cloud Demand

Amazon launches AWS Lambda MicroVMs to safely run AI-generated code, boosting enterprise AI adoption amid $200B capex and strong Q1 cloud revenue growth.

Amazon’s cloud unit has unveiled a new tool aimed at one of the biggest headaches in enterprise AI: running machine-generated code safely. AWS Lambda MicroVMs, launched this month, isolate automatically produced software in lightweight virtual machines, letting developers execute foreign or AI-written code without exposing the broader system. The service, built on Amazon’s own Firecracker technology—which already handles trillions of calls each month—runs on the ARM64 architecture and offers up to 16 virtual CPUs, 32 gigabytes of memory and storage per session, with a maximum runtime of eight hours. Crucially, idle environments can be paused via an API to conserve resources.

The timing is no coincidence. AWS chief Matt Garman, in an interview on June 23, argued that enterprise artificial intelligence has moved beyond pilot phases into measurable deployment. He said he had recently spoken with roughly 100 chief information officers, and about 90 percent of them reported either seeing positive returns from AI projects already or expected to within months. That message is central to Amazon’s broader narrative as it defends a capital expenditure plan of $200 billion—the largest in the company’s history—which has become the dominant point of contention for investors.

Garman insisted the spending is demand-driven, pointing to customer data, pipeline visibility, and shorter lead times for chips and servers as inputs into Amazon’s capital planning. Still, the free cash flow figures tell a more cautious story. For the twelve months ended March 2026, free cash flow dropped to $1.2 billion, driven by a $59.3 billion year-over-year increase in property and equipment investments, predominantly for AI infrastructure.

AWS remains the engine under the hood. In the first quarter of 2026, the segment’s revenue jumped 28 percent to $37.6 billion, while operating profit rose to $14.2 billion from $11.5 billion a year earlier. That gave the cloud business an operating margin of 37.7 percent, and it contributed more than half of Amazon’s total quarterly operating income of $23.9 billion, despite representing only about a fifth of group revenue.

Yet the stock has felt the weight of the capex story. Amazon shares closed at €206.00 (or €206.20 in the secondary market), roughly 13 percent below the 52-week high of €238.05 touched in May. Over the past 30 days, the stock has shed more than 11 percent, though it still holds just above its 200-day moving average of €199.87. The modest positive reaction to the Lambda MicroVMs launch suggests investors are waiting for hard evidence that new serverless features translate into customer wins.

Garman also pushed back against the narrative that AI eliminates entry-level jobs, noting that Amazon is hiring 11,000 interns and early-career professionals this year. He highlighted automated workflows already in production across software development, telecommunications, and financial services as examples of where AI agents are driving real-world efficiency.

For the second quarter of 2026, Amazon has guided net sales in a range of $194 billion to $199 billion, implying year-over-year growth of 16 to 19 percent. Whether AWS can sustain its margin while rebuilding free cash flow will be the key test in the second half. The new MicroVM service gives the company a fresh argument for why enterprises should trust AI-generated code—and why Amazon’s $200 billion bet is more than just a brute-force infrastructure play.