This site is part of the Informa Connect Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 3099067.

Data Center World
April 20-23, 2026
Walter E. Washington Convention CenterWashington, D.C.
Data Center Digest: Cloudflare Fallout, Mega Builds, and Quantum Networks

From global outages and government deals to massive spending and hyperscaler developments, it’s been another eventful stretch in the data center sector. Plus, which tech giants are partnering to make quantum networks a reality? Here’s a roundup of the latest news shaping the industry.

Major Outages and Global Connectivity

A major Cloudflare outage on November 18th took down an enormous slice of the web for more than three hours, affecting high-profile websites and apps like ChatGPT, X, Zoom, and Canva. The company attributed the incident to a misfiring configuration file intended to manage threat traffic – a failure that “triggered a crash” in its software. Roughly 20% of the world’s websites touch Cloudflare in some way. This marked the third major infrastructure outage in a month, following disruptions at AWS and Microsoft Azure. Experts say the increasing frequency and scope of these incidents lays bare the industry’s concentration risk and the fragile networks propping up everyday digital services.

Meanwhile, a different outage hit financial markets. A cooling failure at a CyrusOne data center forced the CME Group – the world’s largest exchange operator – to halt futures trading across key global markets for several hours.

Elsewhere in global news, Google announced two ambitious subsea cable projects—Dhivaru, linking Oman, the Maldives, and Christmas Island, and TalayLink, connecting Australia and Thailand. Both are designed to “future-proof” regional connectivity through new colocation hubs and content-caching capabilities.

Mega-Investments and Market Moves

The industry’s spending streak doesn’t appear to be slowing. Nvidia is reportedly set to spend $26 billion over six years renting cloud capacity. Ironically, many of the cloud providers Nvidia is paying are also major buyers of Nvidia GPUs.

But that number pales next to the $207 billion funding gap HSBC analysts say OpenAI must fill by 2030 to finance its massive data center commitments, including its Stargate mega-campuses. In recent months, OpenAI has committed to more than $1.4 trillion in deals.

On the government side, HPE secured a $931 million contract with the U.S. Defense Information Systems Agency – the Department of Defense’s combat support division – to build a distributed, on-prem hybrid cloud environment using HPE GreenLake.

Meanwhile, AWS committed $50 billion to expand AI and HPC capacity across its classified U.S. government regions, planning three new data centers and nearly 1.3GW of added compute. On the networking side, Nokia announced a $4 billion U.S. investment to expand its R&D and accelerate AI networking technologies.

In other news, Jeff Bezos is returning to the CEO seat as co-chief of Project Prometheus, a newly unveiled AI startup backed by $6.2 billion in funding and already staffed with talent from OpenAI, DeepMind, and Meta, according to The New York Times. The company plans to build AI tools targeting engineering-heavy sectors from aerospace to automotive.

Power Plays and Energy Policy

Constellation has reportedly secured a $1 billion DOE loan to restart the Three Mile Island nuclear plant, which will supply Microsoft through a long-term PPA. The loan is expected to reduce rates across the regional PJM grid and cover most of the plant’s $1.6 billion startup cost.

In Virginia, regulators approved a new rate class for data centers and other 25MW-plus loads, requiring 14-year contracts to ensure cost recovery for grid infrastructure even if projects aren’t built. The move signals growing regulatory scrutiny in the nation’s densest data center market.

Meanwhile, Vertiv and Caterpillar announced a partnership to create pre-designed energy-optimized architectures aimed at speeding deployments and improving power efficiency.

U.S. Development and Pushback

Major U.S. site plans and expansions are also making headlines. In Michigan, a local official confirmed that Meta is the company behind a proposed $1 billion, 1,000-acre data center campus near Howell – though the project continues to face vocal community opposition.

Amazon had a busy week, too. The company announced a $3 billion campus in Vicksburg, Mississippi, and a $15 billion investment in Northern Indiana that will include the construction of a new data center in Hobart and the addition of 2.4GW of capacity.

Elsewhere, Equinix filed plans for two new Dallas data center buildings totaling more than $835 million in investment, and Applied Digital opened the first 100MW building of its North Dakota campus. In Kalkaska County, Michigan, a 1,440-acre data center proposal was withdrawn after sustained local resistance.

And if you needed a reminder of the scale of Amazon’s reach, new reporting suggests AWS now operates more than 900 data centers globally—far beyond previous estimates—according to SourceMaterial. AWS does not disclose its exact data center footprint.

Next-Gen Tech: Quantum Networks, Batteries, and Unified Ops

Future-facing innovations were also in the news. IBM and Cisco unveiled a collaboration to build the first distributed quantum computing network by the early 2030s—laying the groundwork for a quantum internet by merging fault-tolerant quantum systems over not-yet-invented microwave-optical links. The companies aim to demonstrate their initial results within five years.

On the power side, ZincFive introduced a nickel-zinc battery cabinet designed specifically for AI-heavy UPS systems with extreme power-spike patterns. The battery cabinet is reportedly designed to withstand multi-millisecond GPU power spikes that can hit 15x idle levels.

Lastly, Schneider Electric launched EcoStruxure Foresight Operation, a first-of-its-kind unified management platform claiming up to 50% efficiency improvements and 90% faster issue resolution for electrical-mechanical interdependencies.