Amazon reported a sharp earnings beat, with profit of $5.75 per share versus Wall Street’s $1.82 estimate. Revenue totaled $200.61 billion, up 20% year on year, while Amazon Web Services brought in $42.2 billion, up 37%. Net income reached $62.6 billion, helped by a $53 billion unrealized gain tied to its stake in AI lab Anthropic. The results underscore accelerating commercialization of the company’s cloud and AI strategy.

Meta is reportedly considering opening its AI infrastructure to third parties, including renting out unused compute and offering access to hosted models and internal chips. The company expects capital expenditures of up to $145 billion in its current fiscal year as it builds projects such as the Hyperion data-center campus. The report helped lift Meta shares 8.8% and weighed on rivals, with CoreWeave and Nebius sliding, according to Bloomberg and CNBC.

The Trump administration abruptly barred Anthropic’s latest AI models from being used overseas, prompting the company to shut off international access entirely. The move is being read as a signal to accelerate “sovereign AI,” pushing countries to speed up deployment of localized AI infrastructure. The report also points to Micron as a potential beneficiary of demand for domestically controlled AI hardware, while Amazon (AWS) and Microsoft (Azure) could face near-term risk if overseas customers shift to local AI platforms.

Amazon Web Services has launched Continuum, an AI-driven security platform designed to automatically discover, validate and remediate code vulnerabilities at machine speed with limited human involvement. The product is in gated preview and incorporates capabilities previously introduced as Security Agent, while adding threat modeling. AWS says Continuum is model-agnostic, understands business context, and supports gradual automation, and it has been shaped with customers in financial services, automotive and technology.

Data management company Hydrolix has launched on AWS Marketplace, offering an AI-ready data platform built for agentic AI with sub-second query performance. The company says the setup has been validated in work with NVIDIA, where an engineer used Hydrolix’s MCP server to pinpoint a CDN incident in nine minutes. Hydrolix also positions the platform as a natural data layer for applications built on Amazon Bedrock. The integration highlights how NVIDIA GPU compute and AWS cloud infrastructure can be paired in an AI data workflow for media, CDN and enterprise AI use cases.