Coverage: Material developments reported October 29–31, 2026. No material, verified update identified in this window for platforms, applications, robotics, or cooling/networking; those sectors are omitted.
Compute and infrastructure: investment grows, so does cash-flow pressure
Amazon reported third-quarter AWS sales growth of 39% year over year to $46.0 billion. At the same time, free cash flow for the trailing 12 months was an outflow of $42.4 billion, which the company attributed primarily to increased property-and-equipment purchases reflecting AI investment. Amazon maintained its expectation of about $220 billion in 2026 cash capital expenditure. ([aboutamazon.com](https://www.aboutamazon.com/news/company-news/amazon-earnings-q3-2026-report?trk=organization_guest_main-feed-card-text))
Why it matters economically: AWS growth provides evidence that infrastructure demand is translating into cloud revenue, while the cash-flow outflow shows how much capital must be committed before the returns are fully realized. The investment pace makes utilization, customer demand, and eventual returns important variables for both Amazon and its suppliers.
Power and grid: utilities raise spending plans as large loads expand
American Electric Power raised its five-year capital plan to $100 billion, up from $78 billion, citing expected generation and transmission investment. The utility reported 84 GW of contracted large load through 2031, up 22% from its previous-quarter update; a 10.4-GW Ohio data-center campus was a major contributor to the increase. The load figure covers large customers, not AI alone. ([utilitydive.com](https://www.utilitydive.com/news/aep-transmission-data-center-earnings/873189/?.tsrc=rss))
Why it matters economically: AI-linked data-center demand is helping pull forward substantial spending on generation and transmission, creating opportunities for utilities and infrastructure suppliers. But this expansion also raises questions about project timing and who ultimately pays for grid upgrades. An October 31 report noted that more than half of states have taken or considered measures to protect residential customers from data-center-related infrastructure costs; it described policy activity, not a uniform new rule. ([axios.com](https://www.axios.com/2026/10/31/who-pays-ai-power-boom))
Cyber: open-weight AI lowers the cost of a demonstrated exploit chain
In research published October 29, Epoch AI reported that the open-weight model GLM 5.3 exploited the patch gap in six recent Firefox releases in controlled tests, at a marginal API cost of about $1,000 per version after the researchers had built their testing setup. The demonstrated attack could read and send a local file after a user visited a webpage. Epoch said Mozilla had released Firefox 159.0 on October 27, patching the additional, previously undisclosed vulnerability used in the tests; the researchers had not confirmed whether similar attacks still worked against that release. ([epoch.ai](https://epoch.ai/publications/open-weight-models-patch-gap))
Why it matters economically: Lowering the expertise and cost required to develop exploits could increase potential incident and remediation costs for software vendors and organizations. The result is a controlled research finding, not evidence of widespread attacks, and Epoch reported that the tested Firefox vulnerability was patched.
This is a news summary, not investment advice. Capex, demand, utility plans, and policy outcomes may change; AEP’s contracted-load figure is not AI-only. Epoch AI’s cyber results came from controlled tests and do not establish widespread exploitation; Mozilla had patched the additional vulnerability used in the reported Firefox tests.