Getting your news
Attempting to reconnect
Finding the latest in Climate
Hang in there while we load your news feed
Anthropic
Data center news, project activity, and monthly briefings for Anthropic.
Editor's picks
-
Liz Kendall's speech to London Tech Week
Liz Kendall, Secretary of State for Science, Innovation and Technology, announced expanded UK AI industrial policy at London Tech Week, including Sovereign AI (SovAI) details and a new AI Hardware Plan.
- Main announcement: SovAI (launched in April) will invest £500m in British AI companies to start up, scale up and win globally; the government is also mobilising fully funded access to the UK’s largest super computers, super-priority visa decisions and free visas for R&D, and working with the British Business Bank (which runs a £2bn annual investment programme) to back companies. The government also published an AI Hardware Plan that
-
Von Cloud bis KI: Strategie und Governance im Spannungsfeld von Souveränität und Innovation | Konferenz "Finanzdienstleister der nächsten Generation", Frankfurt School of Finance and Management
The Deutsche Bundesbank has announced its AI strategy and adjusted cloud strategy.
- Main announcement: The Bundesbank has adopted an internal AI strategy and adjusted its cloud approach from “Public Cloud First” to “Cloud First Smart Placement”, operating an in-house AI platform TIA (in use since 2024) that runs on the Bundesbank’s own data-centre servers while also using LLMs (including Mistral hosted on Microsoft Azure servers in Europe). The bank is migrating its integrated data and analytics platform (IDA) to a hyperscaler for advanced analysis tools, keeps a Private Cloud for highly sensitive workloads, and is developing exit- and multi-cloud strategies and a procurement framework that will consider EU-native sovereign cloud providers.
- Background and other facts: The Bundesbank runs its own Rechenzentrum and European payment infrastructures (TARGET), has internal teams for quantum computing, and colleagues filed a patent on quantum-safe encryption with a start-up. Internal NLP tooling (an NLP platform and the RegBot) was used to consult more than 27,000 pages of rulebooks. The bank cited that over 80% of digital infrastructure in Europe is imported from non-European providers, referenced the EU AI Act as a regulatory framework, and noted the BSI informed on risks around Anthropic’s Claude Mythos.
-
Bereits da, aber noch nicht überall: Wie lassen sich die wirtschaftlichen Auswirkungen von künstlicher Intelligenz beeinflussen? | Keynote-Rede beim International Economic Symposium
Joachim Nagel of Deutsche Bundesbank delivered a keynote in Rome outlining AI’s likely effects on growth, inflation and financial stability and describing concrete institutional responses (Bundesbank AI strategy) and EU investment initiatives.
- Main announcement/action: The Deutsche Bundesbank has adopted a comprehensive AI strategy and reports that two out of three staff now use AI regularly in a secured environment; the speech also highlights the EU’s InvestAI initiative to mobilise €200 billion (including a €20 billion fund for AI gigafactories) and an EU energy-efficiency package for data centres expected in 2026.
- Background and concrete details: The keynote compares AI diffusion to historical general-purpose technologies, cites private AI investment totals (US$109.1bn in the USA, US$19.4bn in Europe, US$9.3bn in China) and estimates of state support (China ~US$62bn, US ~US$3.3bn, EU ~US$1.2bn); it notes data-centre constraints (e.g., 3–5 year grid connection delays in Dublin and Frankfurt) and references harmonised central-bank surveys in Italy and Germany (2024–25).
-
Already Here, Not Yet Everywhere: Shaping the Economic Impact of Artificial Intelligence | Keynote speech at the International Economic Symposium
Bundesbank President Joachim Nagel delivered a keynote at the International Economic Symposium in Rome on 21 April 2026 about shaping the economic impact of artificial intelligence.
- Main announcement/action: The Bundesbank has approved a comprehensive strategy for artificial intelligence and reports that two out of three Bundesbank colleagues use AI regularly; Nagel called for Europe to accelerate financing, scaling, infrastructure, skills, energy and market integration, highlighting EU-level initiatives such as the AI Act (entered into force 1 August 2024) and the InvestAI initiative (launched February 2025) aiming to mobilise €200 billion, including a €20 billion fund for AI gigafactories.
- Background and details: Nagel summarised cross-country AI investment figures — US private AI investment USD 109.1 billion (2024), China private USD 9.3 billion; China government AI investment USD 62 billion; Europe private USD 19.4 billion; EU government AI investment about USD 1.2 billion — and flagged energy/infrastructure constraints including data centres consuming about 415 TWh/year (≈1.5% of global electricity in 2024) and the forthcoming EU Data Centre Energy Efficiency Package expected in 2026.
-
The energy and environmental impact of AI and how it undermines democracy
Greenpeace and allied campaigners have published reports and actions warning about AI’s rising energy, water and emissions footprint and the democratic risks of concentrated corporate power.
- Main announcement / findings: Greenpeace Germany’s 2025 report warned that AI data centre electricity demand could be 11 times higher in 2030 than in 2023 unless governments intervene, and a February 2026 report backed by Beyond Fossil Fuels found 74% of industry claims about AI’s climate benefits were unproven. The reports document rapidly rising electricity use, water consumption and raw material demands tied to chips and data-centre buildout.
- Context and concrete actions/details: Community and local government pushback is documented with multiple cases: New Brunswick, New Jersey removed data centres from a redevelopment plan; San Marcos, Texas blocked a proposed data centre at a 5-2 vote; South Dublin County Council (Sept 2025) called for a nationwide ban/moratorium or strict conditions (e.g., 100% renewables). The article also cites corporate and contractual figures (e.g., Nvidia revenue US$215.9 billion, Palantir–ICE $30m contract) and legal or policy actions such as a UK legal challenge to a 90MW hyperscale data centre in Buckinghamshire.
-
Powering AI Centers with AI Spines
Arista has announced the 7800R4 as its Universal AI Spine platform for large-scale AI deployments.
- Main announcement: Arista unveils the 7800R4 AI Spine (part of the 7800-series) as an integrated alternative to disaggregated leaf–spine fabrics, offering predictable RDMA performance, centralized control-plane and management, and support for 400G/800G/1.6Tbps interfaces; the portfolio includes 20+ Etherlink switches and is positioned for AI clusters ranging from tens of thousands up to ~100k XPUs.
- Background and details: The platform leverages features such as Virtual Output Queuing (VOQ), large ingress buffers, hierarchical buffering, Arista EOS with real-time load balancing, and telemetry tools (AI Analyzer, Latency Analyzer); hardware resiliency includes redundant supervisors, fabric cards, and field-replaceable components, and the product is presented as interoperable with partners like AMD, Anthropic, Arm, Broadcom, Nvidia, OpenAI, Pure Storage and Vast Data.
Recent news
-
From Data Centers to Models: White House Targets AI Risk at the Source
The White House issued an executive order directing federal agencies to establish a voluntary process for pre-release evaluation of “frontier” AI models that could pose risks to national security or critical infrastructure.
- Main action: The order creates a voluntary federal review framework for frontier models (large foundation and reasoning models) to allow government evaluation before public release when systems could affect critical infrastructure, financial networks, government operations, healthcare, emergency services, or national security; the order does not create mandatory licensing or permitting requirements and emphasizes participation is voluntary.
- Background and implementation details:NIST CAISI has agreements with Google DeepMind, Microsoft, xAI, Anthropic, and OpenAI to conduct pre-deployment evaluations that examine capabilities including cybersecurity, biosecurity, and chemical hazards; the order signals implications for data centers (demand for secure pre-release test environments, strict access controls, telemetry, audit capabilities), and federal actions include the DOE identifying 16 federal sites for data center support and the NLR’s Agora testbed for grid integration.
-
From Data Centers to Models: White House Targets AI Risk at the Source
The White House has issued an executive order directing federal agencies to establish a voluntary review process for certain “frontier” AI models before public release.
- Main announcement: The executive order directs federal agencies to create a voluntary review framework for frontier AI systems that could pose risks to national security or critical infrastructure, asking agencies to evaluate models prior to public release and to coordinate across government. The order does not create mandatory licensing, permitting, or pre-clearance requirements and emphasizes voluntary participation by developers.
- Background and implementation details: The order builds on existing testing agreements—NIST/CAISI has announced pre-deployment evaluations with Google DeepMind, Microsoft, xAI, Anthropic, and OpenAI; the Department of Energy (DOE) has identified 16 federal sites that could support data centers and energy infrastructure; agencies are focused on emergent capabilities (cybersecurity, biosecurity, chemical weapons) and on requirements for secure test environments, access controls, telemetry, and audit capabilities.
-
Alphabet’s $80B Fundraising Spotlights AI’s Soaring Capital Needs
Alphabet has announced a planned equity raise of approximately $80 billion to fund AI infrastructure and global compute capacity.
- Main announcement: Alphabet will raise ~$80 billion in equity to fund capital expenditures to scale AI infrastructure and global compute, including a $10 billion private placement with Berkshire Hathaway (Berkshire to buy $5 billion each of Class A and Class C shares).
- Background and details: The article cites analysts and industry observers noting dramatically increased data center capex (Google expected to double data center capex this year) and emphasizes that funds are intended for data centers, power delivery, networking, cooling, and AI accelerators; portions of proceeds may also cover employee equity and corporate finance rather than direct buildout.
-
FCC's Carr Sends a Warning: ‘No Broadcaster Has a Right to Use the Public Spectrum'
The FCC issued a Public Notice reminding broadcasters that their access to public spectrum depends on meeting public interest obligations.
- Main announcement: On May 28, the FCC issued a Public Notice (link included) stressing that broadcasters receive government-granted spectrum access only if they meet public interest obligations, stating “this Public Notice serves to remind broadcasters of their longstanding public interest obligations and further ensure that broadcasters are continuing to comply with the public interest obligations that underpin their licenses.” The notice also said “no broadcaster has a ‘right’ to use the public spectrum.”
- Other items and details: Headlines in the same roundup include SpaceX calling for automatic mobile phone unlocking within 180 days; GCI sending a crew to repair subsea fiber damage in the Aleutians; a coalition mocking Anthropic’s knowledge of submarine-cable security; NAB urging broadband ISPs and Big Tech to help fund the FCC’s budget; CFTC accusing a Google employee of making $1.2 million from insider trading on Polymarket; and New Mexico awarding $300,000 in planning grants.
-
Land and Expand: NVIDIA, IREN, Coatue, Microsoft, Switch, Cerebras, Core Scientific
NVIDIA announced two major partnerships to accelerate industrial-scale AI infrastructure deployment with IREN and Corning Incorporated.
- Main announcement: NVIDIA partnered with IREN to target deployment of up to 5 gigawatts of NVIDIA DSX-aligned AI infrastructure (focus on IREN’s 2-gigawatt Sweetwater campus in Texas) and separately partnered with Corning Incorporated to expand U.S. optical connectivity manufacturing (10x optical connectivity capacity increase; >50% domestic fiber production increase; construction of three new advanced manufacturing facilities in North Carolina and Texas). The IREN deal includes a five-year right for IREN to sell NVIDIA up to 30 million ordinary shares at $70 per share (potential consideration up to $2.1 billion).
- Background and details: The article details additional industry moves into powered land, gigawatt campuses, crypto-to-AI conversions, and domestic supply-chain expansion, including Coatue/Next Frontier & Fluidstack’s 430 MW Indiana campus backed by $5.7 billion in senior secured notes (first 65 MW online by July 2027), Digi Power X’s 10-year MSA with Cerebras for a 40 MW Columbiana, AL campus (initial contract ~$1.1 billion, potential $2.5 billion, Phase 1 ready-for-service targeted Dec. 15, 2026), CloudBurst’s Texas campus ($14.5 billion investment; 1.2 GW planned), and Core Scientific’s acquisitions and campus expansions (e.g., $421 million cash acquisition of Polaris DS LLC; Muskogee and Pecos expansions to ~1.5 GW gross power).
-
The Hidden Cost Of Compute
Google has begun constructing a $15 Bn AI data centre hub in Visakhapatnam.
- Main announcement/action:Google is building a $15 Bn AI data centre hub in Visakhapatnam to deliver large-scale compute and connectivity; Microsoft has committed $3 Bn to expand cloud and AI infrastructure in India, while major operators including AWS, AdaniConneX, STT GDC, and CtrlS are adding capacity across Mumbai, Chennai, and Hyderabad. The article synthesises these industry commitments and market projections rather than announcing a single new policy.
- Background and details: India’s data centre market was valued at ₹9.33 Lakh Cr (2025) and is projected to reach ₹20.53 Lakh Cr by 2030; a CEEW study reports data centres used ~0.5% of India’s electricity in 2025 and 150 Bn litres of water annually (both metrics forecast to more than double by 2030). The piece also notes state-level incentives, preferential power arrangements, and Maharashtra’s AI policy target of ₹10,000 Cr in investments (policy timelines through 2031).
-
AI Data Center Growth Is Now a Power Infrastructure Problem
Michelle Buckner argues that power infrastructure is now central to the AI buildout.
Main point: The article asserts that AI-driven data center expansion has shifted the primary gating question to power availability — developers are now designing sites around firm megawatts, transmission headroom, substation capacity, and interconnection timelines rather than solely processors or networking. It cites a concrete market milestone: Southwest Power Pool (SPP) expanded into the Western Interconnection on April 1, 2026, creating a 732,000-square-mile footprint across 17 states and serving 20 million people, and bringing in utilities including Basin Electric, Colorado Springs Utilities, Platte River Power Authority, Tri-State Generation and Transmission, and multiple Western Area Power Administration regions.
Details and background: The piece describes how developers favor regions with firm generation (natural gas, nuclear, hydropower), behind-the-meter generation, on-site storage, and microgrids; notes that generation timelines, permitting, transmission upgrades, and fuel logistics now shape deployment schedules; and highlights the need for power-aware design, converter ICs, thermal design, and workload orchestration to reduce megawatt demand. No new commercial contract values or funding amounts are announced.
-
AI Capacity Is Being Pre-Sold at Gigawatt Scale
Google is reportedly negotiating to invest up to $40 billion in Anthropic and to supply reserved compute capacity tied to future performance, though the companies have not publicly confirmed the terms.
- Main announcement: Reported deal would provide an initial $10 billion investment with up to $30 billion contingent on performance, and Google would supply as much as 5 GW of computing power over five years to Anthropic; the companies have not publicly confirmed these terms.
- Background and structure: The arrangement pairs capital with pre-negotiated capacity (including TPU commitments) and builds on existing ties with Google and a separate Broadcom partnership for custom silicon; delivering ~5 GW would require multiple hyperscale campuses, years of land acquisition, permitting, and grid interconnection, and coordinated planning and financing.
-
Google-Anthropic Deal: AI Capacity Now Pre-Sold at Gigawatt Scale
Google is understood to be in talks to invest up to $40 billion in Anthropic.
- Main terms: Reported initial $10 billion investment, with up to $30 billion contingent on performance targets (total potential $40 billion), plus up to 5 GW of computing power over five years supplied by Google; companies have not publicly confirmed these terms.
- Background and implementation details: Builds on prior Google investments and TPU commitments to Anthropic and a separate partnership with Broadcom; delivering 5 GW approaches utility scale and will require multi-year coordination on land acquisition, permitting, grid interconnection, and multiple hyperscale campuses.
-
Karnataka Drafting Responsible AI Framework To Embrace ‘I-Governance’: Priyank Kharge
The Karnataka Government announced a plan to transform citizen services using real-time data, automation and agentic AI systems, while strengthening sustainable data centre policy and expanding deeptech grants.
- Main announcement: The state will transition from traditional e-governance to “i-governance” using real-time data, automation and agentic systems for citizen services (examples: semi-autonomous chatbots for grievance management, image-manipulation detection for crop assessments, nose-print recognition for livestock). The government has set up a committee for responsible AI (experts from technology, law, academia) expected to deliver a comprehensive framework in the next couple of months.
- Background and implementation details: Karnataka has launched a Deeptech Decade, increased deeptech grants from ₹50 Lakh to ₹1 Cr (first cohort drew over 937 startup applications), and plans course correction to the existing Data Centre Policy 2022-27 to strengthen sustainability (focus on liquid cooling, recycled water). The state runs the Dialogue with Diplomats initiative and cites partnerships/events such as India-France Year of Innovation, VivaTech, GITEX Dubai, and Embedded World; the Global Innovation Alliance involves 35+ partner countries and engagement from 80+ countries overall.
-
AI Infrastructure Brief: Power, Capital, and the Feeling That Something Is Tightening
Matt Vincent (Data Center Frontier) summarized the week’s announcements showing an accelerating AI data-center buildout paired with mounting power and coordination constraints.
- Main observation: The industry is prioritizing power and speed: major deals and project announcements include Bloom Energy and Oracle planning up to 2.8 GW of deployment, Aligned Data Centers breaking ground on a 540 MW Project Caprock, an EdgeConneX affiliate proposing a 430 MW natural gas plant in New Albany, Ohio, proposals for 2 GW in New Mexico and 1.2 GW in Irwin County, Georgia, and Microsoft expanding datacenter operations in Cheyenne. The Maine legislature passed a temporary, exemption-inclusive ban on data centers, signaling emerging social-license constraints.
- Capital and implementation details: Financial moves include Switch raising $768 million via ABS, Fluidstack reported in talks for a $1 billion round at an $18 billion valuation, and Jane Street signing a $6 billion AI cloud agreement with CoreWeave; CoreWeave also expanded a multi-year relationship with Anthropic. Utilities are signing long-term power agreements (e.g., NiSource with Alphabet and expanded ties with Amazon). AWS has launched “Project Houdini” to accelerate construction timelines. All items are factual recaps of announcements and reports from the week (no speculative outcomes included).
-
OpenAI pulls out of a second Stargate data center deal
OpenAI pulled out of two European Stargate data center deals (one in the United Kingdom and one in Norway).
- Main action: OpenAI withdrew from two Stargate data center projects — a UK deal was put on hold and negotiations for a Norway facility (Narvik) with neocloud provider Nscale were exited; Microsoft subsequently leased the Norway facility in Narvik and will provide compute power to OpenAI under an unspecified agreement.
- Background and context: Observers describe the moves as financial discipline ahead of an imminent IPO (OpenAI’s potential valuation referenced at $800 billion+); analysts cite runaway power/energy costs, regulatory scrutiny, supply-chain risks (including unrest in the Middle East), and a shift from capex-heavy builds toward renting capacity from hyperscalers like Microsoft.
-
🤖 La Machine #72: France’s AI Push Pits Sovereignty Against Scale
Plume announced a €3.3 million funding round led by AENU to expand its AI platform for renewable energy site selection across Europe and the United States.
- Funding & investors: Plume raised €3.3M with participation from Y Combinator, Kima Ventures, Raise Sherpa, and Collab Fund; funds will support expansion across Europe and the US and continued product development for geospatial/regulatory AI tooling.
- Context & related announcements: The newsletter also highlights the Stanford HAI Index finding that France ranks highly for AI talent and adoption but lags the US/China on models, funding, and scale, and lists government and industry moves on digital sovereignty (DINUM migrating off Windows toward Linux) and sovereign cloud/data centre initiatives (e.g., OVHcloud defence unit).
-
Hyperscalers will own two-thirds of data center capacity by 2031
Synergy Research Group reported that hyperscalers will account for 67% of all data center capacity by 2031.
- Main announcement: Synergy Research Group says hyperscalers (Google, Microsoft, AWS) will reach 67% of global data center capacity by 2031, with enterprise on-prem data centers dropping from 56% in 2018 to 19% by 2031; the report also notes almost 60% of hyperscale capacity is in own-built facilities and non-hyperscale colocation accounts for ~20%.
- Background & details: The article cites planned > $500 billion in capex by Google/Microsoft/AWS for AI infrastructure in fiscal year 2026, cites hyperscalers operating ~1,297 large data centers in Q3 2025 (1,360 by end-2025), references commitments such as the Ratepayer Protection Pledge (Google, Oracle, xAI, Meta, Microsoft, OpenAI, Amazon) and highlights electricity demand concerns (EIA: price hikes up to 79% in areas like Texas by 2027); it references expanded compute partnerships (Anthropic–Google/Broadcom; OpenAI–AMD) with multi-gigawatt capacity starting 2027.