Coverage window: October 28–31, 2026. This update includes material developments verified in sources published during the window; quiet sectors are not padded.
Infrastructure: Alibaba Cloud opens an Amsterdam region
Alibaba Cloud says its first Netherlands public-cloud region is live, with two data centers in Amsterdam. The company says the region offers local data residency and lower-latency cloud and AI services, with agent-focused services planned. Economically, local capacity may help Alibaba compete for European enterprise workloads where data location and response times influence vendor selection. The announcement establishes a regional foothold, but does not quantify customer commitments or utilization. ([alibabacloud.com](https://www.alibabacloud.com/blog/alibaba-cloud-launches-first-public-cloud-region-in-the-netherlands_603704))
Power: FirstEnergy’s data-center pipeline grows, but much of it is prospective
FirstEnergy reported a 26-GW pipeline of contracted and potential data-center load, up 4.8% quarter over quarter; 7.9 GW was contracted through 2036. The utility also described plans for a proposed 1.2-GW gas-fired plant in West Virginia, turbine reservations, and a request for proposals covering another 3 GW of turbines. It estimates data-center demand could add at least $4.75 billion to its transmission investment plan. The economic significance is the scale of potential spending on generation and grid connections—and the opportunity for power providers—alongside a long lead time: the first wave is expected around 2030, and pipeline figures are not equivalent to operating load. ([utilitydive.com](https://www.utilitydive.com/news/firstenergy-data-center-pipeline-maidsville/873149/))
Platforms and applications: OpenAI and ServiceNow deepen workflow integration
OpenAI and ServiceNow announced an expanded partnership to connect ServiceNow workflows with ChatGPT and Codex and bring OpenAI models into more ServiceNow products. The announced use cases include IT requests, application development, and security work; some capabilities are planned for later release. The commercial importance is distribution: embedding AI in existing enterprise workflows may lower adoption friction and support paid software usage, while also increasing competition over who controls the interface, business context, and execution layer. Announced plans should not be mistaken for generally available products. ([openai.com](https://openai.com/index/openai-servicenow-partnership-expansion/))
Platforms: OpenAI backs a shared AI-training preference signal
OpenAI said it will honor a common robots.txt signal for website operators to communicate preferences about use of their content in foundation-model training. The company says implementation is planned in the coming months and that existing controls remain in place. A common signal could reduce the operational burden of managing separate crawler rules and give publishers a clearer mechanism to express their preferences; its economic effect will depend on adoption by other AI providers and how the signal is implemented. ([openai.com](https://openai.com/index/helping-websites-express-ai-training-preferences/))
Robotics: Kodiak signs a 75-truck driverless deployment agreement
Kodiak AI announced a binding agreement for DTL Transport to deploy its driverless system in 75 carrier-owned trucks, with a per-mile fee model and service planned on Sun Belt routes. Separately, Kodiak said its company-defined long-haul autonomy-readiness measure reached 99% and that it is targeting a driverless launch by year-end. If deployed at scale, a per-mile model could turn autonomy software into recurring transport revenue and give carriers a way to add capacity beyond driver-hour limits. The agreement and readiness metric are not proof that the planned fleet or launch has already entered service. ([kodiakrobotics.gcs-web.com](https://kodiakrobotics.gcs-web.com/news-releases/news-release-details/kodiak-ai-and-dtl-transport-reach-historic-agreement-deploy-75))
Cyber: AI-assisted exploitation reported against exposed firewall interfaces
SecurityWeek reported that attackers used an AI-assisted exploit chain involving PHP2Shell flaws and a Linux privilege-escalation vulnerability to gain root access to at least 25 Palo Alto Networks firewall or management systems, citing researcher findings confirmed by Palo Alto Networks. The company said patches were available and urged customers to restrict internet access to management interfaces and update affected systems. Economically, successful compromise can create incident-response costs, downtime, and exposure of internal networks; the immediate priority for affected operators is patching and access restriction. The report describes a limited number of confirmed compromises, not widespread successful exploitation across all exposed systems. ([securityweek.com](https://www.securityweek.com/php2shell-attacks-root-palo-alto-firewalls-but-fail-against-wordpress/))
Quiet sectors
No material compute-chip or cooling/networking development was verified for this update’s coverage window.
Several infrastructure figures describe proposed projects or prospective demand, not completed capacity or guaranteed revenue. Product availability and Kodiak’s launch timing remain forward-looking. Cyber reporting confirms a limited number of compromises; it does not establish broad successful exploitation. This is an informational sector update, not investment advice.