AI News Archive: May 27, 2026 — Part 4
Sourced from 500+ daily AI sources, scored by relevance.
- PlasmidGPT: A generative framework for plasmid analysis and generation
Science Advances, Volume 12, Issue 22, May 2026.
- Teachers Union Urges Schools to Curb A.I. Chatbots and Screen Time
The American Federation of Teachers recommended “no screens” at all for those in second grade or younger, and no A.I. chatbots for students in elementary school.
Score: 51🌐 MovesMay 27, 2026https://www.nytimes.com/2026/05/27/technology/ai-screens-schools-weingarten.html - Automated parking clears regulatory hurdle, but automakers show limited momentum
Automated parking clears regulatory hurdle, but automakers show limited momentum Automotive News
- AI bots ignore evidence. Can we trust them with science?
Scientists rethink their ideas after experiments. AI agents struggle to learn from evidence and recognize when an idea is obviously incorrect.
- What the Pope Said About A.I.
Leo XIV’s new encyclical, “Magnifica Humanitas,” presents a remarkable case for placing moral concerns, and not profit, or competitive advantage, or efficiency, at the center of any discussion of artificial intelligence.
- Travelport, Cognizant and Anthropic Collaborate to Power Travel Technology for the AI Era
Travelport, Cognizant and Anthropic Collaborate to Power Travel Technology for the AI Era The Straits Times
- DuckDuckGo Sees Surge in Installs After Google Goes All-In on AI Search at I/O
DuckDuckGo Sees Surge in Installs After Google Goes All-In on AI Search at I/O PCMag
Score: 51🌐 MovesMay 27, 2026https://www.pcmag.com/news/duckduckgo-sees-surge-in-installs-after-google-goes-all-in-on-ai-search - US law enforcement warns of "anti-tech extremism" as AI hatred grows
The feds are raising the alarm about a new category of threat.
Score: 51🌐 MovesMay 27, 2026https://arstechnica.com/ai/2026/05/us-law-enforcement-warns-of-anti-tech-extremism-as-ai-hatred-grows/ - Claude Code's creator on the end of the software engineer
Anthropic's Boris Cherny tells me major job loss due to automation really is coming — but job creation is, too. PLUS: the Pope's AI encyclical, and Trump abandons an AI executive order
- Scoop: Trump appoints Bondi to White House AI panel
President Trump has appointed former Attorney General Pam Bondi to an advisory committee focused on AI policy, Axios has learned. Driving the news: Bondi, whom Trump ousted as AG last month, will be on the Presidential Council of Advisors on Science and Technology (PCAST). The panel is chaired by former White House AI adviser David Sacks and White House science adviser Michael Kratsios. It also includes more than a dozen tech executives, including Nvidia co-founder Jensen Huang, Meta CEO Mark Zuckerberg and Oracle co-founder Larry Ellison. What we're hearing: Bondi will be charged with facilitating coordination between the government and the tech titans on the panel. What they're saying: "Pam has been an enormously valuable asset to the president's team, and I'm thrilled for her and for all of us that she's going to remain involved in confronting some of the most important issues the administration faces," Vice President JD Vance said in a statement. Bondi will also serve in a newly established advisory role on national infrastructure. Between the lines: Bondi was diagnosed with thyroid cancer shortly after departing the Justice Department, according to a source. She underwent treatment and is recovering.
- Mapping the Glymphatic System with AI
Can we prevent neurodegenerative decline by auditing the brain’s internal cleansing system during deep sleep? A collaborative study leverages physics-informed artificial intelligence to track the fluid flow velocity of the glymphatic system from MRI data.
- The future of AI is not conversation, it is action
For the past few years, the public image of artificial intelligence has been shaped almost entirely by the chatbot. People type questions; AI answers. People ask for summaries, emails, scripts, ideas; AI produces words. This is genuinely powerful. But it is prologue. The real future of AI is not conversation. It is action, and the […] The post The future of AI is not conversation, it is action appeared first on e27 .
Score: 50🌐 MovesMay 27, 2026https://e27.co/the-future-of-ai-is-not-conversation-it-is-action-20260523/ - FuriosaAI partners with Broadcom on AI inference platform
The companies are targeting AI inference, which has become a growing focus as enterprises push AI products into broader commercial use.
Score: 50🌐 MovesMay 27, 2026https://www.bizjournals.com/sanjose/news/2026/05/27/furiosaai-broadcom-ai-inference-platform.html?ana=brss_6150 - AI models have a religion favoritism problem, and new research exposes it
Researchers tested 14 major AI models on religious bias and found a consistent pattern: models subtly favor some faiths over others, with Grok showing the strongest bias and Anthropic and Meta performing the best.
- Broadcom's custom ASIC biz adds South Korea's FuriosaAI to its empire
Third-gen chips to use Broadzilla's advanced packaging, networking tech
Score: 50🌐 MovesMay 27, 2026https://www.theregister.com/ai-ml/2026/05/27/broadcom-lands-furiosaai-as-latest-custom-ai-chip-partner/5246536 - Man (48) convicted and fined for possession of child pornographic material using AI
App had been used to digitally undress image of teenage girl, court hears
- OneQode partners with AMD to roll out global AI infrastructure
OneQode partners with AMD to roll out global AI infrastructure verdict.co.uk
- Canadians may ask AI for medical advice but don't want it replacing humans, poll suggests
Canadians may ask AI for medical advice but don't want it replacing humans, poll suggests CBC
- Stop AI-driven character assassination on YouTube
Kim Se-ui, head of the YouTube channel Garosero Research Institute, who is accused of spreading false information, arrives at the Seoul Central District Court in Seocho District, southern Seoul, on May 26 to attend a pretrial detention hearing. [YONHAP] Kim Se-ui, head of the YouTube channel HoverLab, was detained on Tuesday on charges of spreading false claims that actor Kim Soo-hyun pressured the late actor Kim Sae-ron over debt repayment, allegedly contributing to her death in February of last year. The Seoul Central District Court approved the arrest warrant. Before appearing in court, the YouTube channel runner insisted that the allegations were “filled with obvious falsehoods.” The court nevertheless appears to have determined that a substantial portion of the accusations had merit. According to the Gangnam Police Station in Seoul, which has investigated the case for more than a year, Kim Se-ui spread claims through YouTube and other platforms that Kim Soo-hyun had dated the actor while she was a minor and that financial pressure from the actor’s side contributed to her death. Investigators also allege that AI technology was used to manipulate the actress’s voice and fabricate KakaoTalk conversations. The possibility that AI tools were used to create material difficult for the public to distinguish from reality is especially alarming. If the allegations outlined by police are accurate, the case represents not merely malicious misinformation but what many would describe as “character assassination.” As demonstrated by the damage allegedly inflicted on Kim Soo-hyun, sensational accusations generate explosive public attention and enormous online traffic. YouTube channels can reap substantial advertising revenue in the process. Yet for those targeted, the consequences can be devastating, undermining their mental well-being, public reputation and professional livelihood. Related Article Prosecutors indict Hoverlab operator Kim Se-ui on charges to stalking, threatening Tzuyang Kim Soo-hyun files additional charges against Kim Sae-ron's family, Hoverlab YouTuber Cops intervene in Hoverlab livestream about Kim Soo-hyun Kim Soo-hyun adds to suits against Hoverlab, Kim Sae-ron's family Such conduct is hardly new. So-called cyber wreckers, a Korean term describing online channels that profit from scandal-driven content, have repeatedly manipulated stories about celebrities or spread unverified rumors to maximize views and revenue. In some cases, operators have secretly filmed celebrities’ private lives and attempted extortion. The practice of repackaging tragedy and gossip into monetized online entertainment has long crossed ethical boundaries. Despite YouTube’s growing influence, often rivaling that of traditional media, mechanisms for verifying truth remain dangerously weak. The platform’s recommendation algorithms reward provocative content because outrage and controversy drive clicks and advertising income. At the same time, regulatory oversight remains insufficient. The case also highlights broader concerns surrounding generative AI technology. Deepfake audio, fabricated messages and manipulated video can now spread rapidly before facts are verified. Once false information circulates online, the reputational damage is often irreversible even if later disproven. That is why stronger safeguards against the harms associated with YouTube and AI-driven misinformation are urgently needed. Korea has repeatedly debated measures to address the social damage caused by online rumor channels, but meaningful regulation has lagged behind the speed of technological change. This article was originally written in Korean and translated by a bilingual reporter with the help of generative AI tools. It was then edited by a native English-speaking editor. All AI-assisted translations are reviewed and refined by our newsroom.
- Snapdragon AI Lab
Snapdragon AI Lab Qualcomm
- Within five years we may have AI that does science
EPFL professor discusses AI transforming scientific knowledge
Score: 50🌐 MovesMay 27, 2026https://actu.epfl.ch/news/within-five-years-we-may-have-ai-that-does-science/ - France’s startup scene is falling behind the US and Europe, and AI is both the cause and the cure
A new report on the French tech ecosystem by Alexandre Dewez, a partner at venture firm 20VC, paints a picture of a startup scene that is growing more dependent on a handful of AI companies while the rest of the market stalls. French startups raised €6.7 billion across 411 funding rounds in 2025, a 5% decline in […] This story continues at The Next Web
- 칼럼 | 임원 사칭부터 결제 사기까지…딥페이크, 기업 리스크로 번지다
업무 환경에서의 신뢰는 오랫동안 자연스럽게 내재된 요소였다. 통화에서 익숙한 목소리가 들리거나, 화면에 잘 아는 얼굴이 등장하거나, 임원 명의의 메시지가 전달되면 대부분의 직원은 이를 의심할 이유가 없었다. 하지만 이러한 전제는 이제 점점 설득력을 잃고 있다. 최근에는 딥페이크의 활용 방식과 적용 영역이 빠르게 변화하고 있다. 합성 미디어는 결제 승인부터 임원 커뮤니케이션, 고객 지원 요청에 이르기까지 일상적인 비즈니스 프로세스 전반으로 확산되고 있다. 이러한 상호작용이 디지털 채널로 이동할수록 모방은 쉬워지고, 진위 확인은 더욱 어려워진다. 과거에는 공공 사기나 허위정보, 소셜미디어 문제로 여겨졌던 딥페이크가 이제는 기업 보안의 핵심 이슈로 떠오르고 있다. 가트너에 따르면 전체 조직의 62%가 이미 딥페이크 기반 사회공학 공격을 경험한 것으로 나타났다. 이는 CISO와 리스크 관리 책임자, 경영진과의 논의에서 체감되는 우려 수준을 그대로 반영한다. 대부분의 기업은 피싱 대응 교육과 이메일 보안에 상당한 투자를 해왔다. 이러한 통제는 여전히 중요하지만, 기존의 사기 방식에 맞춰 설계된 것이다. 딥페이크 공격이 특히 효과적인 이유는 노골적으로 의심스러운 요소를 드러내지 않기 때문이다. 오히려 정상적인 요청처럼 보이도록 만들어 사용자를 속인다. 왜 점점 더 강력해지고 있는가 이러한 현상은 조직이 더 분산되고 디지털화되며, 빠른 커뮤니케이션에 의존하게 된 흐름과 맞물려 있다. 화상회의, 채팅 플랫폼, 모바일 기기, 협업 소프트웨어는 이제 의사결정 방식 자체를 바꾸고 있다. 이들은 업무 속도를 높이지만 동시에 판단 시간을 압축한다. 사람들은 제한된 정보 속에서 신속한 대응을 요구받으며, 이는 조작에 유리한 환경을 만든다. 많은 기업에서 회의 중 지시나 메시징 플랫폼, 모바일 대화로 전달된 요청에 즉시 대응하는 것은 매우 일반적인 일이다. 이는 현대 업무 방식의 자연스러운 모습이지 결함이라고 보기는 어렵다. 문제는 이러한 패턴이 사람의 얼굴, 목소리, 커뮤니케이션 방식과 같은 신호를 신뢰할 수 있다는 전제 위에 구축돼 있다는 점이다. 최근 사례는 이 문제가 얼마나 광범위해졌는지를 보여준다. 엔지니어링 기업 아룹에서는 한 직원이 CFO를 포함한 고위 임원으로 보이는 인물들과 화상회의를 진행한 뒤 약 2,500만 달러(약 375억 원)를 송금하는 사건이 발생했다 . 또 다른 사례에서는 봄베이증권거래소 CEO 순다라라만 라마무르티가 딥페이크 영상으로 사칭돼 투자자들에게 잘못된 주식 정보를 전달하기도 했다 . 하나는 내부 승인 절차를, 다른 하나는 공신력 있는 리더에 대한 대중의 신뢰를 악용한 사례다. 이는 합성 미디어가 다양한 환경에서 의사결정을 조작할 수 있음을 보여준다. 이러한 공격을 가능하게 하는 기술 역시 접근성이 높아지고 있다. 과거에는 전문 지식과 상당한 투자가 필요했지만, 이제는 저렴한 AI 모델과 공개 도구, 소량의 데이터만으로도 구현이 가능하다. 짧은 음성 샘플만으로도 목소리를 모방할 수 있고, 제한된 이미지로도 설득력 있는 영상 사칭이 가능하다. AI 기술 발전은 방어 기술에도 영향을 미치고 있지만, 핵심 문제는 여전히 남아 있다. 딥페이크가 빠르게 확산되는 이유는 기존 업무 방식과 자연스럽게 맞아떨어지기 때문이다. 결국 과제는 조작된 콘텐츠를 탐지하는 데 그치지 않는다. 신뢰가 악용될 수 있는 구조 자체를 줄이는 것이 더 중요한 문제로 떠오르고 있다. 사고 대응과 거버넌스, 현실에 맞게 고도화해야 현재 대부분의 기업은 피싱, 랜섬웨어, 데이터 유출에 대응하는 성숙한 대응 매뉴얼을 갖추고 있다. 그러나 조작된 미디어를 활용해 임원을 사칭하거나 부정 승인 절차를 유도하는 상황까지 구체적으로 대비한 곳은 많지 않다. 이러한 공백은 실제로 큰 문제로 이어질 수 있다. 많은 조직이 딥페이크 사고 역시 기존의 사기 또는 사이버 대응 절차로 처리할 수 있다고 가정하지만, 이를 면밀히 점검해보면 허점이 빠르게 드러난다. 특히 모의훈련은 이러한 취약점을 확인하는 데 효과적이다. 책임 소재가 불명확한 지점이나, 압박 상황에서 프로세스가 제대로 작동하지 않는 구간을 명확히 드러내기 때문이다. 예를 들어 가짜 임원의 지시를 가정한 시나리오는 적절한 검증 절차가 실제로 작동하는지를 빠르게 확인할 수 있다. 기업이 간과해서는 안 될 더 큰 문제는 거버넌스다. 금전적 손실이나 데이터 유출이 발생할 경우, 검증은 보안 조직을 넘어 이사회와 규제 당국으로 확대된다. 해당 리스크를 사전에 인지했는지, 그리고 합리적인 통제 체계를 갖추고 있었는지가 핵심 쟁점이 된다. 유럽연합(EU)의 디지털 운영 복원력법(DORA) 과 같은 규제 프레임워크는 이러한 변화 흐름을 반영한다. 딥페이크는 재무 통제, 정보 관리, 브랜드 신뢰, 운영 연속성에 직접적인 영향을 미친다. 이에 따라 법무, 인사, 커뮤니케이션, 경영진 등 다양한 조직이 함께 대응해야 하는 문제로 확대되고 있다. 이는 단순한 보안 이슈를 넘어, 기업 전반의 리스크 관리 과제로 자리잡고 있다. 신뢰는 더 이상 ‘가정’이 아니라 ‘설계’의 영역 CISO에게 요구되는 다음 단계는 신뢰를 정책과 프로세스 기반으로 재정립하는 것이다. 단 한 번의 상호작용만으로 민감한 업무가 처리되지 않도록 구조를 바꿔야 한다. 금전, 계정 정보, 기밀 데이터, 평판 리스크가 포함된 요청이라면 반드시 기존 커뮤니케이션과 분리된 별도의 검증 절차를 거쳐야 한다. 이와 관련해 반복적으로 강조되는 핵심 원칙은 다음과 같다. 위협 모델을 확대 합성 미디어는 이메일, 내부 메시징, 공개 플랫폼, 외부 노출 콘텐츠 전반에서 주요 공격 경로로 간주해야 한다. 직원과 고객, 파트너가 이러한 환경에서 신뢰 기반 의사결정을 내리고 있다면, 해당 채널 역시 반드시 리스크 관리 범위에 포함돼야 한다. 콘텐츠 자체에 제로 트러스트 적용 많은 기업이 접근 제어와 신원 관리 영역에서 제로 트러스트 원칙을 적용하는 데 일정 부분 성과를 거두고 있다. 이제는 사람이 보고 듣고 받는 콘텐츠에도 동일한 수준의 기준을 적용해야 한다. 설득력 있는 영상이나 음성 메시지, 문서만으로 민감한 업무를 승인해서는 안 된다. 머신 속도의 자동 탐지 체계 구축 인간의 판단은 여전히 중요하지만, 조작된 미디어가 빠르게 생성·확산되는 환경에서는 그것만으로 충분하지 않다. 탐지 체계는 전사 환경 전반에서 실시간으로 작동해야 하며, 공격 수준에 상응하는 정교함을 갖춰야 한다. 딥페이크 특화 대응 체계 마련 사고 대응 계획은 단순한 사기나 피싱 시나리오에 머물러서는 안 된다. 임원 사칭, 결제 사기, 브랜드 악용, 내부 또는 외부로 확산되는 조작 콘텐츠 등 딥페이크 특화 상황에 대응할 수 있는 별도의 매뉴얼이 필요하다. 이러한 원칙은 실제 통제 방안으로 이어져야 한다. 특히 다음과 같은 조치는 기본적이면서도 중요하다. • 고위험 요청은 반드시 다른 채널을 통해 재확인해야 한다 • 커뮤니케이션과 승인 절차는 분리해야 한다 • 긴급하거나 비정상적인 요청에는 명확한 에스컬레이션 경로를 설정해야 한다 • 논의용 채널과 승인용 채널을 명확히 구분해야 한다 궁극적으로 효과적으로 대응하는 조직은 심각한 사고가 발생하기 전에 선제적으로 변화에 적응한 기업이다. 딥페이크는 단순한 사기 수법이나 기술적 예외 상황이 아니라, 운영 복원력과 거버넌스, 그리고 압박 상황에서의 의사결정 역량을 시험하는 요소로 자리잡고 있다. 비즈니스에서 신뢰는 앞으로도 중요한 요소로 남는다. 다만 그 기반은 달라지고 있다. 현대 업무 환경에서 신뢰는 더 이상 ‘보이는 것’이나 ‘들리는 것’에 의존할 수 없다. 반드시 검증되어야 하며, 체계적인 프로세스와 함께 다른 모든 비즈니스 통제 요소와 동일한 수준의 엄격함으로 관리돼야 한다. dl-ciokorea@foundryco.com
- Researchers say Google AI Mode changes recommendations based on your emails — and it risks creating a giant confirmation-bias machine
Researchers claim Google AI Mode may prioritize brands found in your Gmail inbox, raising fears that AI-powered search could a confirmation-bias machine.
- 'Threat actors are adapting social engineering and monetization strategies to modern user behavior': Microsoft warns AI chatbots may be sending victims to malicious websites — so be on your guard when clicking
What happens when an AI recommends a tool, and that tool ends up being malicious?
- Datacentre dive: Do AI datacentre physics make on-premise unviable?
The ravenous power and cooling requirements of graphics processing units (GPUs) in artificial intelligence (AI) processing are set to make direct-to-chip liquid cooling mandatory. This is the key factor in the shift away from traditional datacentres and towards AI factories. It means a significant change in the datacentre landscape that could spell the end of the on-premise datacentre, as both cost and complexity spiral away from the ability of enterprises to build their own. These are the key takeaways from an event held by datacentre equipment provider Schneider Electric last week, where industry figures discussed the imminent future of the datacentre scene and visited TeraWulf’s under-construction 750MW site on the shores of Lake Ontario . In this four-part set of articles, we look at the rapid pace of construction at the TeraWulf site, how giant leaps in GPU power dictate datacentre design changes, their effects on the power grid and water use, and colour in the picture as rust belt gives way to AI factory . Read more about datacentres and TeraWulf’s AI factory From rust belt to megawatt AI factory : We visited Terawulf’s Lake Ontario 750MW datacentre development. Photos and recordings weren’t permitted, so we took notes and wrote them up in more traditional ways. ‘We’re at Chinese levels’ at TeraWulf 750MW AI factory : We see the latest in AI factory technology and construction at TeraWulf’s Lake Ontario datacentre, where a former coal-fired power station is site of a rapid transformation. GPU power draw will require grid partnerships : But water use will likely decrease. We look at energy as the key driver – and bottleneck – in development, and why water use is less of an issue now datacentres aren’t like a VW Beetle. The enormous increase in energy consumption driven by AI has brought a step change in datacentre design. Core to this is the requirement to power and cool GPUs to an extent that was not necessary in “traditional” air-cooled datacentres. Hence the advent of the AI factory. Datacentre cooling has been a predictable exercise in industrial heating, ventilation and air-conditioning (HVAC) design, where one slotted servers into racks and blew chilled air across the chassis. AI has rewritten the story. The hardware that powers the AI revolution – GPU especially – operates at thermal and electrical densities that render traditional air-cooling methods obsolete. The silicon demands of large language model training and inferencing cannot be sustained by more or faster fans. Instead, the industry faces an inflection point that mandates direct-to-chip liquid cooling and a transformation of rack-level power delivery to 800-volt direct current (VDC). Liquid cooling mandatory “Liquid cooling isn’t an option, it’s mandatory,” said Rich Whitmore, CEO of Motivair by Schneider Electric, a thermal management firm recently acquired by the latter in 2024 ( assembly workers at Motivair pictured above ). “It is the baseline for all of these high-voltage processors. The changeover point was at about 700W processors [GPUs] like the H100. That was the crossover point between bending the rules of the laws of physics for air cooling and reality. People simply do not have a choice anymore.” The physics underpinning the shift are that when a single processor crosses the 700W threshold, air can no longer move fast enough or hold enough thermal energy to prevent the silicon from throttling or melting. While historical enterprise racks averaged 10kW to 50kW, modern AI training environments routinely deploy 140kW and 150kW clusters. Systems that hit 200kW are set for roll-out, and reference architectures for megawatt-level racks are in place for the end of the decade. That level of energy concentration converts 100% of electrical input into heat in a footprint the size of a fridge. Paradoxically, this transition unlocks thermodynamic efficiencies. Traditional datacentres require energy-intensive refrigeration to supply highly chilled air. Liquid cooling systems operate with far warmer fluid temperatures and allow operators to use high-temperature chillers or fluid-to-air dry coolers. “Air-cooled datacentres are like the old Volkswagen engines where the heat from the load rejects directly into space,” said Tuan Hoang, head of cooling technology and product development at Schneider Electric. “Liquid cooling is like modern automobiles. It is the radiator that removes the heat from the engine. Zero water consumption is actually needed to cool an AI factory when you transition to these closed-loop radiators.” 800V DC the new standard While thermal limits are bringing fluid dynamics into datacentre white space – the revenue-generating area where IT hardware lives – the current required to drive 200kW to 400kW server configurations would overwhelm existing low-voltage distribution frameworks. Until now, cloud facilities have relied on Open Compute Project (OCP) standards that deliver alternating current (AC) to the rack and internal power supplies convert it to 48V or 54VDC to supply individual servers. But, as rack densities climb past 200kW, things become mechanically and structurally impossible. “As you look at trying to use that architecture, you start to run out of headroom,” said Steven Carlini, chief advocate for AI and datacentre at Schneider Electric. “It’s really a mechanical and electrical issue. Right now, you have eight power cables coming into these high-density racks. As you get up towards a megawatt, you would need 32 even larger cables coming into this thing, which is impractical.” To circumvent this bottleneck, the datacentre design is pivoting decisively towards 800VDC power delivery. More volts equals fewer amps, equals smaller cables. By upgrading the distribution architecture to high-voltage DC, datacentre operators can cut the thickness, weight and complexity of copper feeds entering the cabinet. This electrical transformation necessitates new designs for power delivery, which can come from a so-called “sidecar architecture” designed for hybrid environments and brownfield retrofits, and takes the power conversion infrastructure out of the primary IT rack and positions it adjacent to compute hardware, or consolidated centralised distribution targeted at greenfield sites where AC-to-DC conversion is upstream at the facility level, distribution bay or end-of-row. The knock-on effects of changes at silicon level Re-engineering the datacentre down to the silicon level fundamentally changes how infrastructure is designed and maintained. When compute clusters scale at their current rate, minor electrical anomalies or thermal drops carry catastrophic commercial consequences. “Datacentres are fundamentally changing,” said Manish Kumar, executive vice-president for secure power and datacenters at Schneider Electric. “We believe datacentres are becoming AI factories of massive scale and complexity. You have to reimagine how you design, build or bring a datacentre to market and think about the datacentre holistically across the full lifecycle.” This industrial complexity begins with digital twin modelling before physical deployment begins. Because AI developers face large financial penalties for every day GPUs sit idle waiting for power, simulating thermal loads and electrical selectivity in advance derisks capital expenditure and compresses deployment timelines. Meanwhile, transitioning to an 800VDC framework introduces system protection issues. Unlike AC systems, high-voltage DC circuits lack zero points at which it is easier to break a circuit. This necessitates the development of specialised solid-state circuit breakers so that if a single fault occurs at blade level, only that specific breaker trips and doesn’t take down an entire multimillion-dollar training cluster. Datacentres are at a crossroads. Operators and enterprise infrastructure face a strategic fork in the road: abandon legacy air and low-voltage electrical power delivery, or potentially face obsolescence as the physical realities of the AI age leave existing infrastructure behind. Does AI direct-to-chip cooling put paid to on-premise datacentres? CIOs have existed in a comfortable equilibrium where the corporate data model evolved into a hybrid form. In this, non-critical, elastic workloads migrated to the public cloud, while sensitive core business systems, proprietary datasets and predictable processing loads remained inside corporate walls in traditional air-cooled on-premise server rooms. AI potentially shatters this model. With the shift from standard central processing unit computing to accelerated GPU clusters, the physical requirements of modern AI hardware cannot work with legacy on-premise designs. With next-generation silicon demanding mandatory direct-to-chip liquid cooling and unprecedented power densities, is this the end of the on-premise corporate datacentre? Liquid cooling unviable for the majority? As we have seen, the root of the infrastructure inflection point lies in the thermal intensity of AI hardware. For some in the industry, the complexity and capital expenditure required to deploy liquid cooling frameworks means on-premise AI is unviable for the vast majority of enterprises. In the past, an enterprise could construct a high-quality datacentre building, install the electrical and cooling infrastructure, and reliably run three, four, or even five successive generations of IT hardware refreshes over 15 years without altering the underlying facility. AI hardware has broken that model. The acceleration of chip design means each consecutive generation of AI processors brings new physical dimensions, power profiles and fluid-flow requirements that are fundamentally incompatible with infrastructure built just a year before. “In the old days of datacentres, you would build the building and the facility, the power and the cooling systems, and you could do three, four and five IT refreshes,” said Chris Burnett, account executive at Cloudflare. “[With] today’s datacentre … very few people are going to build double the size of the power and the cooling for the next generation. You’re building it for today; it’s extremely challenging.” For an enterprise CIO, the commercial implications are that constructing an on-premise datacentre capable of handling 200kW racks requires millions of pounds in specialised upfront capital expenditure. If that bespoke facility design becomes obsolete in a single IT lifecycle because the next iteration of silicon requires entirely different fluid dynamics or higher voltages, the financial return on investment evaporates. Therefore, the argument for outsourcing to massive public cloud hyperscalers or specialised multi-tenant colocation providers becomes compelling. Or democratic deployment for all? Others suggest that declaring the death of the corporate datacentre is premature. From this perspective, the long-term future of enterprise AI will not consist solely of monolithic foundational model training – which undeniably belongs in specialised hyperscale environments. Instead, the real commercial value for the average enterprise lies in fine-tuning smaller, highly secure, domain-specific models on proprietary corporate data. “Will enterprises deploy direct liquid cooling or is that going to stay out of their reach? I think they definitely will,” said Schneider’s Carlini. “They definitely will move to direct-to-chip liquid cooling.” He said that as direct-to-chip liquid cooling technologies mature, the market will undergo a process of industrial standardisation with infrastructure providers delivering modular, self-contained “plug-and-play” liquid-cooled enclosures designed specifically to fit into existing corporate footprints. Carlini highlighted that once the initial mechanical barrier is crossed, the inherent thermodynamic efficiencies of liquid systems work in favour of the enterprise. “With the efficiency of liquid cooling and the temperatures you can run at, the water use is much less,” he said. By operating at significantly warmer fluid temperatures, these systems eliminate the need for massive, complex external refrigeration units, potentially making localised high-density compute more operationally efficient than legacy air systems. Hybrid probably the key Meanwhile, there is also the possibility of a hybrid approach structured around the lifecycle phases of AI. For the resource-intensive training phase – where thousands of GPUs must be tightly clustered together to ingest petabytes of data over weeks or months – the corporate datacentre is definitively unviable. This work will be outsourced to specialised hyperscale or colocation environments that possess the native 800VDC electrical distribution and high-capacity liquid cooling loops. But once a model is trained, the operational focus shifts entirely to inferencing that requires significantly lower computational density per query and must be located physically close to the company’s operational data stores to minimise network latency and comply with data protection legislation. This is where the on-premise liquid-cooling services described by Carlini might find their home. In this scenario, enterprise datacentres will be retrofitted to support compact, highly efficient, liquid-cooled inferencing zones. CIOs should audit their requirements The advent of direct-to-chip liquid cooling has dissolved the traditional datacentre playbook. The legacy corporate server room cannot adapt to the physics of modern accelerated silicon. CIOs who try to force AI workloads into traditional air-cooled configurations potentially face thermal throttling, energy waste and ballooning costs. But also, those who attempt to build on-premise replicas of hyperscale datacentres risk capital lock-in on infrastructure that could be obsolete by the next chip generation. The path forward requires a rigorous, application-driven approach to infrastructure. CIOs should audit their AI application pipelines separately from high-density training needs and localised inferencing. A hybrid model can leverage the scale of specialised colocation providers for heavy lifting, while preparing their internal teams to adopt standardised, closed-loop liquid systems for secure inferencing.
Score: 50🌐 MovesMay 27, 2026https://www.computerweekly.com/news/366643672/Datacentre-dive-Do-AI-datacentre-physics-make-on-premise-unviable - Gartner Says CFOs Risk Falling Behind Without a Scalable AI Strategy
Gartner Says CFOs Risk Falling Behind Without a Scalable AI Strategy Gartner
- ITBench-AA: Frontier Models Score Below 50% on the First Benchmark for Agentic Enterprise IT Tasks — by Artificial Analysis and IBM
ITBench-AA: Frontier Models Score Below 50% on the First Benchmark for Agentic Enterprise IT Tasks — by Artificial Analysis and IBM
- Salesforce's beat fails to convince the market that software can survive AI
Salesforce is doing everything it can to signal to the market that this year's sell-off is unwarranted.
Score: 50🌐 MovesMay 27, 2026https://www.cnbc.com/2026/05/27/salesforces-beat-fails-to-convince-market-that-software-can-survive-ai.html - AI reveals how the brain clears harmful waste
AI reveals how the brain clears harmful waste EurekAlert!
- GPUs And RAM Are In Short Supply, But The Real Bottleneck For AI Is Electricians
GPUs And RAM Are In Short Supply, But The Real Bottleneck For AI Is Electricians
- New technology allows AI agents to read and respond to people's facial expressions
Technology now allows the creation of increasingly realistic AI agents. Human-like AI agents—such as the digital characters that appear as virtual assistants, game characters, and the increasingly lifelike "metahumans" used in film and interactive media—can look remarkably realistic, but their facial behavior during conversation remains stiff and generic. A doctoral thesis defended at Tallinn University has addressed this gap by building a system that enables an AI agent to both interpret and generate facial expressions.
Score: 49🌐 MovesMay 27, 2026https://techxplore.com/news/2026-05-technology-ai-agents-people-facial.html - OpenAI’s GPT-5.5-Cyber: Advanced cyber security or potential threat?
The AI model is being made available through OpenAI’s Trusted Access for Cyber – a controlled programme for verified defenders and approved organisations.
- AI training data is becoming a seller's market. Here's what it's worth
From Reddit's $60 million API deal to personal data worth "a few dollars," the new economics of AI training data are taking shape
- Roku rolling out home screen redesign with AI-powered app organization, ‘Top Picks’ [Video]
Roku has just unveiled its biggest home screen redesign in years with a new “Top Picks” row on top, and AI to organize your apps, among other updates. more…
- Inbound Call Center: definition, benefits, and how AI is transforming customer service
Inbound Call Center: definition, benefits, and how AI is transforming customer service
- Decision Intelligence: Getting Value From AI Agents and Multiagent Systems
Decision Intelligence: Getting Value From AI Agents and Multiagent Systems Gartner
- How AI is Transforming Scientific Discovery While Keeping Humans at the Center
From designing new antibodies to simulating 1,000 years of climate in a day, AI is transforming what's possible—but humans remain the ones deciding what matters.
Score: 48🌐 MovesMay 27, 2026https://hai.stanford.edu/news/how-ai-is-transforming-scientific-discovery-while-keeping-humans-at-the-center - Why AI shopping can finally move from recommendations to checkout
Alibaba’s integration of Qwen and Taobao shows why data may become the moat for shopping in the AI era.
Score: 48🌐 MovesMay 27, 2026https://kr-asia.com/why-ai-shopping-can-finally-move-from-recommendations-to-checkout - American AI firms want it both ways in limiting, profiting off China
OpenAI does not officially offer its services in mainland China, Hong Kong or Macau, yet it is hiring Mandarin-speaking developer experience engineers in Singapore. On May 14, the day US President Donald Trump met President Xi Jinping in Beijing, Anthropic published a policy paper arguing the United States and its allies must lock in a 12-to-24-month lead in frontier artificial intelligence (AI) by 2028 to avoid “authoritarian AI leadership”. In the same quarter, Anthropic disclosed US$1.56...
- Oklahoma Launches AI Chatbot to Answer Medicaid Questions
The Oklahoma Health Care Authority’s new AI tool aims to help answer eligibility questions from the state’s Medicaid program members to reduce wait times and expand access to information.
Score: 48🌐 MovesMay 27, 2026https://www.govtech.com/artificial-intelligence/oklahoma-launches-ai-chatbot-to-answer-medicaid-questions - Figma Brings AI Into The Design Canvas With New Creative Agent
Figma Brings AI Into The Design Canvas With New Creative Agent YourStory.com
- Extreme dynamic symmetry enables omnidirectional and multifunctional robots
Science Robotics, Volume 11, Issue 114, May 2026.
- Marvell sees custom chip revenue topping $10 billion by 2029 as AI demand grows
Marvell sees custom chip revenue topping $10 billion by 2029 as AI demand grows Reuters
Score: 48🌐 MovesMay 27, 2026https://www.reuters.com/technology/marvell-technology-forecasts-quarterly-revenue-above-estimates-2026-05-27/ - AI model of early child development also explains how language evolves
AI model of early child development also explains how language evolves EurekAlert!
- I was intrigued by Google's new video-cloning Omni AI - then I considered the implications
Google's Gemini Omni combines realism, avatars, style control, and natural-language editing in one AI video tool.
- AI to turbocharge patent creation at India tech hubs, executives say
INDIA-SUMMIT/PATENTS (PIX):AI to turbocharge patent creation at India tech hubs, executives say
- Artificial Intelligence (AI) for Manufacturing Workshop
U.S. manufacturing expects to gain transformative productivity and resilience improvements through AI integration in product development and production processes. Despite AI's potential to transform manufacturing operations and workflows, significant
Score: 48🌐 MovesMay 27, 2026https://www.nist.gov/news-events/events/2026/05/artificial-intelligence-ai-manufacturing-workshop - First Take: Subversive AI Threat Demands Action From Heads of EA
First Take: Subversive AI Threat Demands Action From Heads of EA Gartner
- What’s New for Game Developers in NVIDIA RTX: DLSS 4.5 for UE5 and Multilingual AI Characters
NVIDIA RTX provides game developers with direct paths to AI-driven characters, frame generation, and ray-traced rendering. This post walks through a meaningful...