News & Reading
A rolling reading list across the things I run programs in — data centers, infrastructure, RFID and hardware, networking, AI program management, and observability.
Speed to power is now the primary siting constraint. Hyperscalers are measuring site viability in megawatts available within 18 months, not square footage — a meaningful reframe for anyone running infrastructure programs at scale.
Data Center KnowledgeGov. Abbott's pause on new ERCOT grid connections could delay nearly 20% of the U.S. development pipeline, while a 2.5-GW gas-plus-nuclear project in Victoria, TX advances past concept into licensing. Two sides of the same power-availability bottleneck.
POWER MagazineJune 2026 construction starts reached $22.3B, against an 81.5B YTD total, with costs up 17–22% compounded over two years. Amazon raised its AI infrastructure spend to $220B with demand already stretching into 2028. Program leads in this space are managing scale, cost escalation, and schedule compression simultaneously.
ConstructConnectMicrosoft, Google, Amazon, and Meta have collectively committed to nearly 10 GW of nuclear capacity through more than a dozen agreements in 2026. Microsoft is investing $16B to restart Three Mile Island's sister reactor by 2027, solely for AI data centers.
ForbesA measured counterpoint from the Bulletin of the Atomic Scientists — AI companies are not spending anywhere near what's needed to bring SMRs online on announced timelines. Worth reading alongside the Forbes piece for a balanced program-level view of nuclear as a power strategy.
Bulletin of the Atomic ScientistsThe LDB milestone is a meaningful proof point for RFID at national-logistics scale: 10 crore EXIM containers tracked across ports, terminals, ICDs, SEZs, and railways with full visibility. The operational challenge at that volume — read accuracy, integration with customs and carrier systems, data normalization — maps directly to the complexity in enterprise retail rollouts.
LinkedIn / Supply Chain & Logistics NewsFull compliance is required across apparel, home goods, toys, electronics, and most general merchandise as of this year. The operational problem — encoding at scale, supplier onboarding, and read-rate accuracy in store — is where the real program complexity lives.
RFID HY / Walmart Supplier GuideHow Decathlon is using RAIN RFID across its global retail operation. The interesting detail isn't the tag — it's how they handle encoding throughput and integration with inventory systems at volume. A solid reference point for enterprise-scale rollouts.
RAIN AllianceEnterprise observability has shifted from experimental tooling to critical infrastructure. As mature observability platforms underpin incident detection, problem root-cause, and SLA reporting, ITIL 4's service management practices — incident management, problem management, and continual improvement — are being operationalized through the same telemetry pipelines rather than separate ITSM tools.
BankInfoSecurity / Enterprise Observability ReportGrafana Labs named a Gartner Magic Quadrant Leader for Observability Platforms in 2026, with Fleet Management enabling collector deployments at scale. For ITIL practitioners, this represents the practical stack behind continual service improvement — OTel-native pipelines feeding Grafana dashboards that drive incident, change, and problem workflows rather than siloed monitoring tools.
Grafana LabsThe shift to AI-assisted and chat-native incident workflows directly reinforces ITIL 4's emphasis on speed of detection, swarming models for major incidents, and post-incident review as a continual improvement driver. Worth reading for anyone sitting the ITIL 4 Managing Professional or Strategic Leader tracks alongside real operational work.
Incident.ioThe methodology behind the PMI-CPMAI credential continues to iterate, with closer alignment to data-science lifecycle stages and enterprise governance patterns. Useful for anyone managing AI programs who wants a structured delivery model behind the work.
CPMAIPractical writing on identity-based mesh networking. Reshapes what home lab and small-team infrastructure looks like — and increasingly what zero-trust enterprise access looks like too.
Tailscale BlogBefore July 2026
At least 16 GW of capacity is slated for 2026 across ~140 projects, yet only about 5 GW is currently under construction. The analysis maps where the gap lies — power interconnection queues, permitting timelines, and capital deployment pace — and what it means for program leads managing vendor and site delivery.
Currence AI / Data Center OutlookThe convergence of RFID with AI is enabling predictive insights and real-time tracking across retail, logistics, and healthcare. Useful framing for anyone managing RFID programs at scale — the technology story has moved well past compliance mandate into operational intelligence.
CYBRAHow much of the application layer belongs at the edge, how much still belongs in origin. Useful mental model for anyone running a hybrid platform.
Cloudflare BlogAutomatic HTTPS, sane defaults, config that reads like a config instead of a program. A good example of infrastructure done with taste.
CaddyResearch from PMI on how the TPM role has widened across engineering organizations. Closer to product and platform, further from Gantt.
PMIWhere metrics live, how long they live, and what you give up by moving them out of your own environment. Applies to home lab and enterprise both.
PrometheusNotes on how the tool most of us reach for first has evolved. Worth tracking if your programs ship containers to distributed environments.
DockerHow volunteer receiver networks continue to shape open access to aviation telemetry. Interesting as an open-data pattern, useful as a hobby.
FlightAwareBaseline labor and rate data for anyone running a personal underwriting model on a small real-estate portfolio.
U.S. Bureau of Labor StatisticsCurated by hand — no algorithm. Suggestions welcome via the contact page.