A new bipartisan U.S. bill aims to make large data centers pay for the grid upgrades needed to power them instead of passing those costs to households.

The Ratepayer Protection Act would require state utility regulators to consider rules making large-load data centers cover the full incremental cost of new generation, transmission, and distribution infrastructure needed to serve them. The House Energy and Commerce Committee advanced the bill unanimously in July 2026. Meanwhile, BloombergNEF projects U.S. data centers could consume about 20% of the nation’s electricity by 2035, up from roughly 5.9% today.

#TechNews #EnergyGrid #photography

References:
U.S. House Energy and Commerce Committee: Ratepayer Protection Act Advances From Committee To Protect Consumers From Data Center Energy Costs.
Rep. Kathy Castor: Bipartisan Ratepayer Protection Act Would Prevent Families From Paying For Data Center Grid Upgrades.
Data Center Dynamics: U.S. Data Centers Could Consume 20 Percent Of National Electricity By 2035, BloombergNEF Forecasts.

Rapid expansion of AI infrastructure is facing new challenges as communities and companies navigate the demands of building large-scale data centers.

Research from Sightline Climate indicates that many announced data center projects planned for 2026 have faced delays, highlighting obstacles such as power availability, grid limitations, equipment shortages and permitting challenges.

The slowdown has sparked debate over how quickly AI infrastructure should grow. Supporters argue that data centers bring investment, jobs and technological progress, while critics point to concerns about electricity demand, water use and the impact on local communities.

As AI continues to expand, the challenge is not only developing more powerful systems but also creating the infrastructure needed to support them responsibly.

The future of AI growth will likely depend on balancing innovation with energy planning, environmental considerations and the needs of communities where these facilities are built.

#fblifestyle #Polarbear #AI #Technology #Environment

2. Unprecedented Hardware Depreciation Cycles
​Traditional Infrastructure: A power plant, railroad, or cloud server building shell depreciates over 15 to 30 years.
​AI Silicon Reality: Advanced GPUs and AI accelerators (e.g., Nvidia architectures) become technologically outdated or less energy-efficient within 3 to 5 years due to rapid microchip advances. Unlike traditional real estate, up to 70%+ of an AI data center’s capital cost goes into silicon that loses its competitive edge in a few short years. To achieve ROI, the facility must recoup its massive capital cost before the hardware becomes obsolete.

​3. Infrastructure Built for Training vs. Inference Uncertainty
​Front-Loaded Spend: Massive “training clusters” require huge upfront capital investment before a model is even launched or monetized.
​Speculative Demand: If enterprise adoption of generative AI or autonomous agents scales slower than anticipated, hyperscalers risk holding massive, high-maintenance compute assets that operate below capacity, severely depressing returns on capital.

Tech executives and infrastructure bulls argue the ROI critique mistakes a timing delay for a structural failure:
Pre-Sold Capacity: Hyperscalers report that major data center expansions are frequently pre-committed or pre-sold to enterprise cloud clients and AI firms before construction even finishes.
Defensive CapEx: For major tech firms, building compute capacity is load-bearing insurance. Failing to build the infrastructure risks losing market share in search, enterprise cloud, and productivity software to competitors who have the compute capacity to run next-generation workloads.

Things are different on the other side of the Atlantic ocean. This happens for example in Finland, Sweden, Denmark, Norway

AI experts sound alarm about ‘terrifying’ rumoured change to how ChatGPT works
Reported update could make it harder to understand how systems are behaving, critics warn
https://www.independent.co.uk/tech/chatgpt-ai-openai-update-chain-of-thought-b3044099.html?fbclid=IwdGRjcAUI5ipjbGNrBQjmEnBkb2YFZXh0bgNhZW0CMTEAc3J0YwZhcHBfaWQMMzUwNjg1NTMxNzI4AAEenVXjws-V0w_9WGr22Ka-vHC5zBr-SsnXnnlQBe7Kn9tNPFl7MolhA0PbOGw_aem_6Qj8G8EJ5cCsZjqlipFwlA

Artificial intelligence experts have sounded alarm about a “terrifying” rumoured update to ChatGPT that could make it harder to monitor what AI systems are doing.

This week, reports suggested that OpenAI’s upcoming Astra model will use a technique called “opaque recurrence”, though staff at the company have dismissed those reports. Tech publication The Information, which first reported the rumoured change at OpenAI, said in a separate report that both Anthropic and Google DeepMind were considering similar updates.

So Not Right Now
Data Center Backlash Has Officially Rattled Sam Altman
“People are pretty negative on AI.”
https://futurism.com/artificial-intelligence/data-center-backlash-rattled-sam-altman?fbclid=IwdGRjcAUImZljbGNrBQiZhnBkb2YFZXh0bgNhZW0CMTEAc3J0YwZhcHBfaWQMMzUwNjg1NTMxNzI4AAEerT_OfuXDCBB1izLS2S1KBCoWQwzFc3r_PYwPRs8rrC-JbTQARilAsNjsZcQ_aem_6jp4EMRaL6QOOd-gET4evw

Slingin’ Slop
Employees Absolutely Loath AI in the Workplace, Survey Finds
“It’s not just about the fear being replaced by AI.”
https://futurism.com/artificial-intelligence/employees-loath-ai-workplace-bosses-glassdoor?fbclid=IwdGRjcAUImWFjbGNrBQiZRXBkb2YFZXh0bgNhZW0CMTEAc3J0YwZhcHBfaWQMMzUwNjg1NTMxNzI4AAEeT5qQADdp0GJqwefhV9t7_CYDCeq487_WVH4noBeygXt6VdxKkUVwKzeFr5o_aem_UXUjQYuGOJI7RTlchjDpPQ

It’s no secret that AI is becoming increasingly unpopular around the world, as the tech’s garbled output has become impossible for the average person to avoid. For a rising cohort of workers, AI is also becoming inescapable on the job. According to one piece of research, staffers aren’t exactly thrilled with that reality.

The new survey, conducted by the Indeed-owned workplace review site Glassdoor, found that mentions of “AI” in employer reviews has skyrocketed over the last few years — though sentiments, tellingly, are becoming overwhelmingly negative as time goes on.