Intelligence for the acceleration.

‘Attempted corporate murder’: Trump’s threats against Anthropic chill AI industry

President Trump ordered a government-wide boycott of Anthropic's Claude AI and threatened prosecution against the company after CEO Dario Amodei refused to permit military use of the technology for mass surveillance or autonomous armed drones. Defense Secretary Hegseth had suggested invoking the Cold War-era Defense Production Act to force compliance, which legal experts warned would constitute effective partial nationalization of the AI industry.

The billion-dollar infrastructure deals powering the AI boom

Major AI providers and cloud hyperscalers are negotiating multi-billion dollar infrastructure partnerships as the compute demands of frontier models reshape vendor relationships. OpenAI has formally diversified beyond exclusive reliance on Microsoft Azure, securing right-of-first-refusal terms while reserving capacity to use other providers if Azure cannot meet infrastructure demands; Microsoft has reciprocated by exploring other foundation models for its own AI products. Meta, Oracle, Google, and emerging players are racing to lock in the physical capacity—data centers, power agreements, and network backbone—that will determine which entities control the next phase of AI deployment.

‘Big energy users’: how will datacentres affect Australia’s power prices, water supply and emissions?

Australia hosts 260 operational data centres concentrated in Sydney and Melbourne, with energy consumption from AI facilities projected to exceed national electric vehicle fleet usage by 2030 and approach the annual consumption of four aluminium smelters by 2035. Cooling requirements drive demand for both electricity and potable water, with industry projections showing data centre expansion will slow power sector emissions reductions after 2035 despite closed-loop cooling alternatives. Tech companies are pressuring governments to accept ratepayer protections while maintaining access to grid infrastructure built for public benefit.

Datacentre developers face calls to disclose effect on UK's net emissions

Campaign groups are demanding UK data center developers disclose environmental impacts and fund renewable energy construction proportional to their projects. The government maintains data centers will help meet environmental challenges while acknowledging to MPs that future demand from the sector "remains inherently uncertain." The initiative comes as the UK's target for a virtually carbon-free power grid by 2030 faces mounting pressure from electricity cost increases.

Jack Dorsey's 4,000 Job Cuts at Block Arouse Suspicions of AI-Washing

Block Inc. eliminated nearly half its workforce—approximately 4,000 positions—this week, with co-founder Jack Dorsey attributing the cuts to AI-driven efficiency gains. The announcement sits at the center of an emerging critique that companies are exploiting AI anxiety to rebrand traditional cost-cutting as technological modernization, while labor advocates question whether the deployed AI capabilities actually justify the scale of displacement.

Stablecoin yield rewards (likely won't be) banned under OCC proposal: State of Crypto

The Office of the Comptroller of the Currency published proposed rulemaking under the GENIUS Act to govern U.S. stablecoin issuance, with provisions addressing custody controls and capital requirements. The yield-related sections contain ambiguous language that multiple observers tracking the process describe as controversial, raising unresolved questions about how regulators will treat yield-generating stablecoin products.

Google is building a bevy of renewable energy in Minnesota—including the world's largest battery system providing power for a whopping 100 hours

Google is developing a data center complex in Minnesota powered by 1.4 gigawatts of wind, 200 megawatts of solar, and Form Energy's 300-megawatt iron-air battery system capable of 100 hours of storage—enough to power 200,000 homes. The project relies on a new green tariff agreement allowing Google to self-finance its clean energy mix without passing costs to residents. Form Energy's multiday battery technology aims to solve renewable intermittency problems that have limited grid-scale adoption.

Online Platforms Are Not Liable for What Users Post. Should That Include Gen AI?

Senator Ron Wyden, co-author of Section 230, stated that generative AI tools do not automatically qualify for the law's liability protections. Speaking at a conference hosted by the R Street Institute, Wyden argued that AI-generated content differs from passive hosting of user speech, suggesting regulations should target "harmful use" rather than specific development methods. Panelists highlighted the financial risks AI companies face if courts rule that algorithmic output constitutes platform-created content rather than third-party speech.

A dangerous playbook is being revived for the giant US housing agencies

Freddie Mac and Fannie Mae are increasing purchases of mortgage-backed securities, reviving a business model that contributed to the 2008 financial crisis. The housing giants are expanding investment portfolios, concentrating systemic risk in federally-backed entities that control trillions in US housing finance.

Could a huge data centre revitalise Ayrshire - or ruin it?

A 540MW data center proposal near Kilmarnock has sparked debate over water consumption and community benefit. The facility would require millions of liters of water daily for cooling, potentially straining local resources while developers pledge community investments including walkways and water taxis.

'Silent failure at scale': The AI risk that can tip the business world into disorder

AI systems deployed across business operations are introducing a failure mode distinct from traditional software bugs: the "silent failure at scale" where systems execute instructions literally rather than as intended, compounding minor errors over weeks or months before detection. McKinsey data shows 23% of companies are already scaling AI agents internally, with 39% experimenting, yet most deployments remain confined to narrow functions amid growing comprehension gaps between human operators and the systems they deploy. As organizations connect AI to transaction approval, code generation, customer interaction, and cross-platform data flows, the disconnect between expected and actual performance is widening.

Jack Dorsey's 4,000 Job Cuts at Block Arouse Suspicions of AI-Washing

Block Inc. eliminated nearly half its workforce—approximately 4,000 positions—this week, with co-founder Jack Dorsey attributing the cuts to AI-driven efficiency gains. The announcement sits at the center of an emerging critique that companies are exploiting AI anxiety to rebrand traditional cost-cutting as technological modernization, while labor advocates question whether the deployed AI capabilities actually justify the scale of displacement.

Stablecoin yield rewards (likely won't be) banned under OCC proposal: State of Crypto

The Office of the Comptroller of the Currency published proposed rulemaking under the GENIUS Act to govern U.S. stablecoin issuance, with provisions addressing custody controls and capital requirements. The yield-related sections contain ambiguous language that multiple observers tracking the process describe as controversial, raising unresolved questions about how regulators will treat yield-generating stablecoin products.

A dangerous playbook is being revived for the giant US housing agencies

Freddie Mac and Fannie Mae are increasing purchases of mortgage-backed securities, reviving a business model that contributed to the 2008 financial crisis. The housing giants are expanding investment portfolios, concentrating systemic risk in federally-backed entities that control trillions in US housing finance.

Why China's humanoid robot industry is winning the early market

China's humanoid robot sector, prioritized under the "Made in China 2025" industrial plan, is outpacing US competitors in shipment volume and iteration speed despite a market still in its infancy. Domestic firms combine advances in multimodal AI with state-backed manufacturing to deploy humanoids in contained industrial and warehouse environments first, aiming to address labor shortages while navigating safety risks that could trigger public backlash. Global shipments hit only 13,317 units last year but projected annual doubling could reach 2.6 million units by 2035.

Your utility bills keep going up. Here's everyone you can blame—AI data centers included

Utilities are announcing hundreds of billions in infrastructure spending driven by data center demand, and ratepayers are absorbing the cost in monthly bill increases. Duke Energy CEO Harry Sideris defended rate hikes while acknowledging affordability concerns, as the PJM Interconnection region—where data centers are heavily concentrated—sees the most severe impacts. Pennsylvania Governor Josh Shapiro has called for selectivity in data center approvals, citing community, cost, and environmental concerns raised by constituents.

US confirms first combat use of LUCAS one-way attack drone in Iran strikes

U.S. Central Command confirmed the first combat deployment of the Low-cost Unmanned Combat Attack System (LUCAS) drone during Operation Epic Fury against Iran on February 28, 2026. The autonomous kamikaze drones, reverse-engineered from Iranian Shahed-136 designs, targeted Islamic Revolutionary Guard Corps command facilities, air defense systems, and military infrastructure. The deployment follows Defense Secretary Pete Hegseth's July directive to accelerate acquisition of affordable autonomous systems and establish drone squadrons capable of saturating adversaries with inexpensive, expendable platforms.

AI panic has been erasing value all around the market. Here's where 3 investing pros see it hitting next.

Wall Street analysts identify the next sectors vulnerable to AI-driven disruption panic: stretched banking valuations facing automation exposure, industrial and transport sectors confronting physical AI (autonomous logistics, warehouse robotics), and private credit markets carrying concentrated tech risk. Citi projects warehouse automation alone will grow to $112 billion by 2029. Physical AI presents "super threat" to incumbents who fail adoption.

Bitcoin is stuck in a rut but JPMorgan says new legislation could be the ultimate spark

The Clarity Act, proposed U.S. legislation to establish federal cryptocurrency oversight, has stalled in the Senate after Coinbase withdrew support over concerns the text would restrict innovation and stablecoin rewards. JPMorgan analysis suggests comprehensive regulatory clarity would unlock sidelined institutional capital from asset managers and pension funds currently deterred by legal uncertainty, potentially allowing new projects to raise up to $75 million annually without full SEC registration.

India Built the World's Back Office. A.I. Is Starting to Shrink It.

Artificial intelligence is beginning to automate the white-collar outsourcing work that transformed India into a global technology powerhouse. Indian Prime Minister Narendra Modi framed AI as a civilizational transformation comparable to electricity, while industry workers deploy chatbots designed to eliminate the call center and back-office jobs that once lifted millions into the middle class. The country is racing to adapt its workforce before automation outpaces retraining and economic transition efforts.

Space Force opens secretive space tracking to commercial firms

The U.S. Space Force is integrating commercial data and artificial intelligence into its classified satellite tracking systems. The initiative, part of what the military calls battle management, command and control, aims to improve space domain awareness by distinguishing normal orbital maneuvers from potential hostile intent. Commercial data feeds combined with AI prediction models compress decision timelines ten- to one hundred-fold, allowing operators to assess threats and respond before an attack materializes.

Hackers hit Iranian apps, websites after US-Israeli strikes

Cyber-enabled operations accompanied joint U.S.-Israeli strikes on Iran early Saturday. State-linked Iranian hacking groups conducted data-wiping attacks on Israeli targets, while unidentified actors defaced the BadeSaba religious calendar app—downloaded by over 5 million users—with messages urging Iranian armed forces to disarm. CrowdStrike observed Iranian-aligned threat actors conducting reconnaissance and DDoS attacks alongside physical military operations.

Meta won’t let morality get in the way of a product launch

Meta has introduced "Name Tag," a facial recognition feature for Ray-Ban Meta smart glasses that identifies people in real-time using cameras and AI. The launch comes as ICE expands surveillance operations and the administration weaponizes deportation infrastructure, with the company maintaining close ties to federal authorities. Unlike previous facial recognition controversies that sparked public backlash, this deployment arrives amid normalized surveillance conditions where biometric identification is becoming ambient infrastructure embedded in consumer devices.

US court blocks landmark law limiting social media use for children

A federal judge has blocked Virginia's law restricting minors' social media access, ruling the state lacks authority to limit 'minors' access to constitutionally protected speech.' Tech lobbying group NetChoice, representing Meta, Google, X and others, successfully argued the law violated the First Amendment. The Virginia ruling follows similar injunctions in Louisiana and Ohio, while California's narrower law targeting 'addictive feeds' was upheld in January.

China Asked ChatGPT for Help Crafting Online Harassment Campaigns

OpenAI's threat intelligence report reveals Chinese government operatives used ChatGPT to refine 'cyber special operations' targeting political dissidents abroad. The operation, linked to the 'Spamouflage' network, generated fake evidence for takedown requests and created impersonation accounts targeting US-based critics.

AI just leveled up and there are no guardrails anymore

New York State Assemblyman Alex Bores authored the first major AI safety law in the US and is now running for Congress, becoming a target for deregulation advocates. The article examines how AI development is accelerating faster than governance frameworks can adapt, with the Anthropic-Pentagon conflict highlighting the tension between safety constraints and government pressure.

Major Data Brokers Tried to Hide Their Opt-Out Pages From Search Engines

Sen. Maggie Hassan's investigation found data brokers Comscore, IQVIA Digital, Telesign, and 6sense Insights used "no index" code to hide opt-out pages from search engines, preventing consumers from exercising privacy rights. The companies changed their practices after being contacted, except Findem, which failed to respond and continues blocking its opt-out form. The report estimates identity theft from four major data broker breaches cost consumers $20.9 billion.

The Case for Why Better Breach Transparency Matters

RSA Conference session led by security consultants Adam Shostack and Adrian Sanabria highlights systemic lack of feedback mechanisms in cybersecurity incident response, arguing that mandated detailed breach disclosure is essential to reduce cyber-risk. Current US requirements vary state-by-state with publicly traded companies only obligated to report material-impact incidents, while The British Library's 2023 ransomware after-action report cited as rare example of comprehensive public accountability.

Trump directs US agencies to toss Anthropic's AI as Pentagon calls startup a supply risk

The Trump administration ordered federal agencies to immediately cease using Anthropic technology after the AI company refused Pentagon demands to remove guardrails on its Claude model for autonomous weapons and mass domestic surveillance. Defense Secretary Pete Hegseth designated Anthropic a supply-chain risk to national security—a label typically reserved for firms from adversarial nations like China—blocking any military contractor from working with the company. The $200 million defense contract represented a small portion of Anthropic's $14 billion revenue, but the blacklisting threatens its planned public offering and broader business relationships. Anthropic stated it would challenge the designation in court.

AI deepfakes are a train wreck and Samsung's selling tickets

Samsung executives acknowledged that AI-generated imagery is eroding the concept of photographic evidence, yet expressed little urgency about implementing protective measures. During a product launch, Samsung's mobile chief admitted the company sees a divide between users who want AI photo features and those concerned about reality erosion, while dodging questions about whether users should be able to remove AI watermarks from generated photos.

Space Force opens secretive space tracking to commercial firms

The U.S. Space Force is integrating commercial data and artificial intelligence into its classified satellite tracking systems. The initiative, part of what the military calls battle management, command and control, aims to improve space domain awareness by distinguishing normal orbital maneuvers from potential hostile intent. Commercial data feeds combined with AI prediction models compress decision timelines ten- to one hundred-fold, allowing operators to assess threats and respond before an attack materializes.

Hackers hit Iranian apps, websites after US-Israeli strikes

Cyber-enabled operations accompanied joint U.S.-Israeli strikes on Iran early Saturday. State-linked Iranian hacking groups conducted data-wiping attacks on Israeli targets, while unidentified actors defaced the BadeSaba religious calendar app—downloaded by over 5 million users—with messages urging Iranian armed forces to disarm. CrowdStrike observed Iranian-aligned threat actors conducting reconnaissance and DDoS attacks alongside physical military operations.

Strikes on Iran will test US cyber strategy abroad, and defenses at home

The federal government's cyber defense agency faces heightened threat levels while operating with significant staffing shortages, cybersecurity experts warn. Iranian-linked groups have historically targeted U.S. financial institutions, infrastructure providers, and industrial control systems through DDoS campaigns, ransomware, and hack-and-leak operations. Former officials anticipate retaliatory operations targeting exposed operational technology and internet-facing PLC hardware.

‘Attempted corporate murder’: Trump’s threats against Anthropic chill AI industry

President Trump ordered a government-wide boycott of Anthropic's Claude AI and threatened prosecution against the company after CEO Dario Amodei refused to permit military use of the technology for mass surveillance or autonomous armed drones. Defense Secretary Hegseth had suggested invoking the Cold War-era Defense Production Act to force compliance, which legal experts warned would constitute effective partial nationalization of the AI industry.

Datacentre developers face calls to disclose effect on UK's net emissions

Campaign groups are demanding UK data center developers disclose environmental impacts and fund renewable energy construction proportional to their projects. The government maintains data centers will help meet environmental challenges while acknowledging to MPs that future demand from the sector "remains inherently uncertain." The initiative comes as the UK's target for a virtually carbon-free power grid by 2030 faces mounting pressure from electricity cost increases.

Stablecoin yield rewards (likely won't be) banned under OCC proposal: State of Crypto

The Office of the Comptroller of the Currency published proposed rulemaking under the GENIUS Act to govern U.S. stablecoin issuance, with provisions addressing custody controls and capital requirements. The yield-related sections contain ambiguous language that multiple observers tracking the process describe as controversial, raising unresolved questions about how regulators will treat yield-generating stablecoin products.

Online Platforms Are Not Liable for What Users Post. Should That Include Gen AI?

Senator Ron Wyden, co-author of Section 230, stated that generative AI tools do not automatically qualify for the law's liability protections. Speaking at a conference hosted by the R Street Institute, Wyden argued that AI-generated content differs from passive hosting of user speech, suggesting regulations should target "harmful use" rather than specific development methods. Panelists highlighted the financial risks AI companies face if courts rule that algorithmic output constitutes platform-created content rather than third-party speech.

Meta won’t let morality get in the way of a product launch

Meta has introduced "Name Tag," a facial recognition feature for Ray-Ban Meta smart glasses that identifies people in real-time using cameras and AI. The launch comes as ICE expands surveillance operations and the administration weaponizes deportation infrastructure, with the company maintaining close ties to federal authorities. Unlike previous facial recognition controversies that sparked public backlash, this deployment arrives amid normalized surveillance conditions where biometric identification is becoming ambient infrastructure embedded in consumer devices.

OpenAI's Sam Altman announces Pentagon deal with 'technical safeguards'

OpenAI has reached an agreement with the US Department of War to deploy its AI models within the Pentagon's classified network. CEO Sam Altman stated the deal includes prohibitions on domestic mass surveillance and maintains human responsibility for autonomous weapon systems. The agreement follows the collapse of negotiations between the Pentagon and rival AI company Anthropic, which refused to remove safeguards against surveillance and autonomous weapons use.

US confirms first combat use of LUCAS one-way attack drone in Iran strikes

U.S. Central Command confirmed the first combat deployment of the Low-cost Unmanned Combat Attack System (LUCAS) drone during Operation Epic Fury against Iran on February 28, 2026. The autonomous kamikaze drones, reverse-engineered from Iranian Shahed-136 designs, targeted Islamic Revolutionary Guard Corps command facilities, air defense systems, and military infrastructure. The deployment follows Defense Secretary Pete Hegseth's July directive to accelerate acquisition of affordable autonomous systems and establish drone squadrons capable of saturating adversaries with inexpensive, expendable platforms.

Online Platforms Are Not Liable for What Users Post. Should That Include Gen AI?

Senator Ron Wyden, co-author of Section 230, stated that generative AI tools do not automatically qualify for the law's liability protections. Speaking at a conference hosted by the R Street Institute, Wyden argued that AI-generated content differs from passive hosting of user speech, suggesting regulations should target "harmful use" rather than specific development methods. Panelists highlighted the financial risks AI companies face if courts rule that algorithmic output constitutes platform-created content rather than third-party speech.

Estamos entrando en la era en la que cualquiera puede ser tú. Deepfakes, IA y el colapso silencioso de la confianza en la identidad digital

Over 70% of Latin Americans lack precise knowledge of what deepfakes are, creating a vulnerable population as AI-generated impersonation attacks accelerate. Security forecasts indicate 2026 marks the shift of digital identity from a peripheral concern to a primary attack target for personalized fraud using social media data.

US court blocks landmark law limiting social media use for children

A federal judge has blocked Virginia's law restricting minors' social media access, ruling the state lacks authority to limit 'minors' access to constitutionally protected speech.' Tech lobbying group NetChoice, representing Meta, Google, X and others, successfully argued the law violated the First Amendment. The Virginia ruling follows similar injunctions in Louisiana and Ohio, while California's narrower law targeting 'addictive feeds' was upheld in January.

China Asked ChatGPT for Help Crafting Online Harassment Campaigns

OpenAI's threat intelligence report reveals Chinese government operatives used ChatGPT to refine 'cyber special operations' targeting political dissidents abroad. The operation, linked to the 'Spamouflage' network, generated fake evidence for takedown requests and created impersonation accounts targeting US-based critics.

AI deepfakes are a train wreck and Samsung's selling tickets

Samsung executives acknowledged that AI-generated imagery is eroding the concept of photographic evidence, yet expressed little urgency about implementing protective measures. During a product launch, Samsung's mobile chief admitted the company sees a divide between users who want AI photo features and those concerned about reality erosion, while dodging questions about whether users should be able to remove AI watermarks from generated photos.

How Chinese AI Chatbots Censor Themselves

Stanford and Princeton researchers found that Chinese AI models including DeepSeek and Alibaba's Qwen are significantly more likely than Western counterparts to dodge political questions or deliver inaccurate responses on sensitive topics. The study reveals systematic self-censorship mechanisms embedded in Chinese AI systems, with researchers finding techniques to force models to expose their hidden refusal instructions.

How AI is supercharging Russia's online disinformation campaigns

Security experts warn that Kremlin-aligned actors are deploying AI-generated synthetic videos at scale to shape public opinion across Europe and the US, while Western governments lack adequate tools and laws to respond. A King's College London professor's identity was hijacked via AI voice-over deepfake for a Russia-linked operation dubbed "matryoshka," which embeds false claims in layers of ambient re-posts from compromised accounts.

FTC declines to enforce a kids privacy law for data collected to verify users' ages

The Federal Trade Commission announced it will not enforce the Children's Online Privacy Protection Rule (COPPA) against companies collecting personal data solely for age verification technologies, provided strict safeguards are followed. The policy shift incentivizes age verification adoption but privacy advocates warn it creates new risks. The Electronic Frontier Foundation noted that age verification systems have already experienced data breaches, citing a Discord incident where 70,000 users had government IDs exposed through a third-party vendor.

How scammers are using AI deepfakes to steal money from taxpayers

Meta's AI sending 'junk' tips to DoJ, US child abuse investigators say

Meta's AI-powered content moderation systems are flooding US law enforcement with low-quality, unreliable reports about child sexual abuse material, according to officers from the Internet Crimes Against Children task force. The flood of useless tips is draining investigative resources and hindering actual cases. The issue emerged during a New Mexico lawsuit against Meta over child safety on its platforms, where the company has defended its cooperation with law enforcement while facing questions about whether its automated detection systems create more noise than signal.

Watch a computer powered by human brain cells play Doom

Cortical Labs has trained its CL-1 biocomputing chip, composed of 200,000 lab-grown human neurons, to play the video game Doom. Visual data from the screen is translated into electrical stimulation patterns, and the living neurons respond with their own signals that control in-game actions. The demonstration builds on the company's 2022 work showing similar cultures playing Pong, representing a functional interface between living neural tissue and digital computing systems.

Space Force opens secretive space tracking to commercial firms

The U.S. Space Force is integrating commercial data and artificial intelligence into its classified satellite tracking systems. The initiative, part of what the military calls battle management, command and control, aims to improve space domain awareness by distinguishing normal orbital maneuvers from potential hostile intent. Commercial data feeds combined with AI prediction models compress decision timelines ten- to one hundred-fold, allowing operators to assess threats and respond before an attack materializes.

Hackers hit Iranian apps, websites after US-Israeli strikes

Cyber-enabled operations accompanied joint U.S.-Israeli strikes on Iran early Saturday. State-linked Iranian hacking groups conducted data-wiping attacks on Israeli targets, while unidentified actors defaced the BadeSaba religious calendar app—downloaded by over 5 million users—with messages urging Iranian armed forces to disarm. CrowdStrike observed Iranian-aligned threat actors conducting reconnaissance and DDoS attacks alongside physical military operations.

Strikes on Iran will test US cyber strategy abroad, and defenses at home

The federal government's cyber defense agency faces heightened threat levels while operating with significant staffing shortages, cybersecurity experts warn. Iranian-linked groups have historically targeted U.S. financial institutions, infrastructure providers, and industrial control systems through DDoS campaigns, ransomware, and hack-and-leak operations. Former officials anticipate retaliatory operations targeting exposed operational technology and internet-facing PLC hardware.

‘Attempted corporate murder’: Trump’s threats against Anthropic chill AI industry

President Trump ordered a government-wide boycott of Anthropic's Claude AI and threatened prosecution against the company after CEO Dario Amodei refused to permit military use of the technology for mass surveillance or autonomous armed drones. Defense Secretary Hegseth had suggested invoking the Cold War-era Defense Production Act to force compliance, which legal experts warned would constitute effective partial nationalization of the AI industry.

'Silent failure at scale': The AI risk that can tip the business world into disorder

AI systems deployed across business operations are introducing a failure mode distinct from traditional software bugs: the "silent failure at scale" where systems execute instructions literally rather than as intended, compounding minor errors over weeks or months before detection. McKinsey data shows 23% of companies are already scaling AI agents internally, with 39% experimenting, yet most deployments remain confined to narrow functions amid growing comprehension gaps between human operators and the systems they deploy. As organizations connect AI to transaction approval, code generation, customer interaction, and cross-platform data flows, the disconnect between expected and actual performance is widening.

The billion-dollar infrastructure deals powering the AI boom

Major AI providers and cloud hyperscalers are negotiating multi-billion dollar infrastructure partnerships as the compute demands of frontier models reshape vendor relationships. OpenAI has formally diversified beyond exclusive reliance on Microsoft Azure, securing right-of-first-refusal terms while reserving capacity to use other providers if Azure cannot meet infrastructure demands; Microsoft has reciprocated by exploring other foundation models for its own AI products. Meta, Oracle, Google, and emerging players are racing to lock in the physical capacity—data centers, power agreements, and network backbone—that will determine which entities control the next phase of AI deployment.

‘Big energy users’: how will datacentres affect Australia’s power prices, water supply and emissions?

Australia hosts 260 operational data centres concentrated in Sydney and Melbourne, with energy consumption from AI facilities projected to exceed national electric vehicle fleet usage by 2030 and approach the annual consumption of four aluminium smelters by 2035. Cooling requirements drive demand for both electricity and potable water, with industry projections showing data centre expansion will slow power sector emissions reductions after 2035 despite closed-loop cooling alternatives. Tech companies are pressuring governments to accept ratepayer protections while maintaining access to grid infrastructure built for public benefit.

Datacentre developers face calls to disclose effect on UK's net emissions

Campaign groups are demanding UK data center developers disclose environmental impacts and fund renewable energy construction proportional to their projects. The government maintains data centers will help meet environmental challenges while acknowledging to MPs that future demand from the sector "remains inherently uncertain." The initiative comes as the UK's target for a virtually carbon-free power grid by 2030 faces mounting pressure from electricity cost increases.

Jack Dorsey's 4,000 Job Cuts at Block Arouse Suspicions of AI-Washing

Block Inc. eliminated nearly half its workforce—approximately 4,000 positions—this week, with co-founder Jack Dorsey attributing the cuts to AI-driven efficiency gains. The announcement sits at the center of an emerging critique that companies are exploiting AI anxiety to rebrand traditional cost-cutting as technological modernization, while labor advocates question whether the deployed AI capabilities actually justify the scale of displacement.

Watch a computer powered by human brain cells play Doom

Cortical Labs has trained its CL-1 biocomputing chip, composed of 200,000 lab-grown human neurons, to play the video game Doom. Visual data from the screen is translated into electrical stimulation patterns, and the living neurons respond with their own signals that control in-game actions. The demonstration builds on the company's 2022 work showing similar cultures playing Pong, representing a functional interface between living neural tissue and digital computing systems.

Space Force opens secretive space tracking to commercial firms

The U.S. Space Force is integrating commercial data and artificial intelligence into its classified satellite tracking systems. The initiative, part of what the military calls battle management, command and control, aims to improve space domain awareness by distinguishing normal orbital maneuvers from potential hostile intent. Commercial data feeds combined with AI prediction models compress decision timelines ten- to one hundred-fold, allowing operators to assess threats and respond before an attack materializes.

Why China's humanoid robot industry is winning the early market

China's humanoid robot sector, prioritized under the "Made in China 2025" industrial plan, is outpacing US competitors in shipment volume and iteration speed despite a market still in its infancy. Domestic firms combine advances in multimodal AI with state-backed manufacturing to deploy humanoids in contained industrial and warehouse environments first, aiming to address labor shortages while navigating safety risks that could trigger public backlash. Global shipments hit only 13,317 units last year but projected annual doubling could reach 2.6 million units by 2035.

US confirms first combat use of LUCAS one-way attack drone in Iran strikes

U.S. Central Command confirmed the first combat deployment of the Low-cost Unmanned Combat Attack System (LUCAS) drone during Operation Epic Fury against Iran on February 28, 2026. The autonomous kamikaze drones, reverse-engineered from Iranian Shahed-136 designs, targeted Islamic Revolutionary Guard Corps command facilities, air defense systems, and military infrastructure. The deployment follows Defense Secretary Pete Hegseth's July directive to accelerate acquisition of affordable autonomous systems and establish drone squadrons capable of saturating adversaries with inexpensive, expendable platforms.

How a million new satellites could turn night into day

SpaceX has proposed deploying up to one million satellites to create orbital data centers, while a separate startup seeks FCC approval to deploy reflective satellites that would beam sunlight to Earth at night. The Washington Post reports these proposals threaten to transform the night sky into a permanent artificial twilight, with satellites outnumbering visible stars and fundamentally altering humanity's relationship with the cosmos.

Could a niche 80s technology be the key to better quantum computers?

SEEQC, a quantum computing startup, is reviving superconducting computing circuits first explored in the 1980s to build more efficient quantum processors. New Scientist reports that the company, operating from a former IBM superconducting computing facility in New York, is developing digital single flux quantum technology that could dramatically reduce the energy costs and error rates plaguing current quantum systems.

Cutting-edge Chinese gene-editing technique raises prospect of new autism treatments

Chinese researchers have successfully used an advanced gene-editing tool to correct a DNA mutation responsible for cognitive and behavioral problems in laboratory mice. The South China Morning Post reports that mice engineered with the mutation showed significant behavioral changes after receiving injections with edited genes, suggesting potential pathways for treating autism spectrum disorders. The research represents a significant advance in precision genetic medicine.

Rocket Lab scrubs planned Feb. 25 launch of hypersonic scramjet vehicle for the US military

Rocket Lab scrubbed the planned February 25 launch of its HASTE suborbital rocket carrying DART AE, a scramjet-powered hypersonic test vehicle developed for the Defense Innovation Unit. The mission, dubbed "That's Not a Knife," would have been DIU's second hypersonic demonstration using Rocket Lab's HASTE platform after the November 2025 "Prometheus Run" launch. The vehicle tests technologies enabling sustained flight at Mach 5+ for missile defense and strike applications.

Breaking encryption with a quantum computer just got 10 times easier

Researchers have developed a more efficient quantum computing approach using qLDPC codes that reduces qubit requirements for breaking RSA encryption by an order of magnitude. The new method enables qubits to interact beyond nearest neighbors, increasing information density and reducing the estimated qubit count from millions to approximately 100,000.

Flagship-backed Generate Biomedicines eyes $2.2 billion valuation in US IPO

Generate Biomedicines, founded in 2018 by Flagship Pioneering (the venture firm behind Moderna), announced plans to raise up to $425 million in a Nasdaq IPO that would value the company at $2.17 billion. The company uses AI to replace traditional trial-and-error drug discovery by generating novel protein-based therapeutics computationally. Its lead candidate, GB-0895 for severe asthma, is currently in late-stage trials. Menlo Ventures, which cited Aurora Therapeutics' bespoke CRISPR work as validation of the approach, has invested $16 million in AI-driven genetic medicine startups.

‘Big energy users’: how will datacentres affect Australia’s power prices, water supply and emissions?

Australia hosts 260 operational data centres concentrated in Sydney and Melbourne, with energy consumption from AI facilities projected to exceed national electric vehicle fleet usage by 2030 and approach the annual consumption of four aluminium smelters by 2035. Cooling requirements drive demand for both electricity and potable water, with industry projections showing data centre expansion will slow power sector emissions reductions after 2035 despite closed-loop cooling alternatives. Tech companies are pressuring governments to accept ratepayer protections while maintaining access to grid infrastructure built for public benefit.

Datacentre developers face calls to disclose effect on UK's net emissions

Campaign groups are demanding UK data center developers disclose environmental impacts and fund renewable energy construction proportional to their projects. The government maintains data centers will help meet environmental challenges while acknowledging to MPs that future demand from the sector "remains inherently uncertain." The initiative comes as the UK's target for a virtually carbon-free power grid by 2030 faces mounting pressure from electricity cost increases.

Google is building a bevy of renewable energy in Minnesota—including the world's largest battery system providing power for a whopping 100 hours

Google is developing a data center complex in Minnesota powered by 1.4 gigawatts of wind, 200 megawatts of solar, and Form Energy's 300-megawatt iron-air battery system capable of 100 hours of storage—enough to power 200,000 homes. The project relies on a new green tariff agreement allowing Google to self-finance its clean energy mix without passing costs to residents. Form Energy's multiday battery technology aims to solve renewable intermittency problems that have limited grid-scale adoption.

Could a huge data centre revitalise Ayrshire - or ruin it?

A 540MW data center proposal near Kilmarnock has sparked debate over water consumption and community benefit. The facility would require millions of liters of water daily for cooling, potentially straining local resources while developers pledge community investments including walkways and water taxis.

Your utility bills keep going up. Here's everyone you can blame—AI data centers included

Utilities are announcing hundreds of billions in infrastructure spending driven by data center demand, and ratepayers are absorbing the cost in monthly bill increases. Duke Energy CEO Harry Sideris defended rate hikes while acknowledging affordability concerns, as the PJM Interconnection region—where data centers are heavily concentrated—sees the most severe impacts. Pennsylvania Governor Josh Shapiro has called for selectivity in data center approvals, citing community, cost, and environmental concerns raised by constituents.

China's Solar Power Generation Overtakes Wind for First Time

China's solar power generation exceeded wind output for the first time in 2025 as a manufacturing boom in cheap photovoltaic panels reshaped the nation's energy mix. The milestone marks a structural shift in the world's largest emitter's electricity grid, with solar capacity additions accelerating despite ongoing challenges with grid integration and storage. The transition carries global climate implications given China's position as both the largest energy consumer and the dominant manufacturer of clean energy equipment worldwide.

Dirty water, death and decline: the inside story of a privatisation scandal

A Guardian investigation reveals UK water privatization has loaded companies with debt while infrastructure crumbles. Secret 2002 government reports predicted this outcome. UN special rapporteur Pedro Arrojo-Agudo criticizes the system, while Thames Water—serving 16 million customers—struggles with £20bn debts and private equity owners demand 15-year exemptions from environmental rules.

Rising carbon dioxide levels now detected in human blood

Researchers have for the first time detected rising atmospheric carbon dioxide levels within the human body, with a key blood marker approaching its healthy limit within decades if current trends continue. The findings are especially relevant for children and adolescents, who face the longest cumulative exposure to rising atmospheric CO₂ during their developmental years.

How a million new satellites could turn night into day

SpaceX has proposed deploying up to one million satellites to create orbital data centers, while a separate startup seeks FCC approval to deploy reflective satellites that would beam sunlight to Earth at night. The Washington Post reports these proposals threaten to transform the night sky into a permanent artificial twilight, with satellites outnumbering visible stars and fundamentally altering humanity's relationship with the cosmos.

Ocean geoengineering trial finds no evidence of harm to marine life

Researchers conducted a field trial in the Gulf of Maine, pumping 65,000 litres of sodium hydroxide into coastal waters to test ocean alkalinity enhancement. The trial removed up to 10 tonnes of atmospheric carbon dioxide without measurable harm to marine wildlife, plankton, or lobster larvae. The technique increases water alkalinity, allowing oceans to absorb more CO2 and converting it into bicarbonate ions expected to remain sequestered for tens of thousands of years.

‘Big energy users’: how will datacentres affect Australia’s power prices, water supply and emissions?

Australia hosts 260 operational data centres concentrated in Sydney and Melbourne, with energy consumption from AI facilities projected to exceed national electric vehicle fleet usage by 2030 and approach the annual consumption of four aluminium smelters by 2035. Cooling requirements drive demand for both electricity and potable water, with industry projections showing data centre expansion will slow power sector emissions reductions after 2035 despite closed-loop cooling alternatives. Tech companies are pressuring governments to accept ratepayer protections while maintaining access to grid infrastructure built for public benefit.

A dangerous playbook is being revived for the giant US housing agencies

Freddie Mac and Fannie Mae are increasing purchases of mortgage-backed securities, reviving a business model that contributed to the 2008 financial crisis. The housing giants are expanding investment portfolios, concentrating systemic risk in federally-backed entities that control trillions in US housing finance.

Could a huge data centre revitalise Ayrshire - or ruin it?

A 540MW data center proposal near Kilmarnock has sparked debate over water consumption and community benefit. The facility would require millions of liters of water daily for cooling, potentially straining local resources while developers pledge community investments including walkways and water taxis.