© Loren Elliott for The New York Times
© The New York Times
© Nathan Howard for The New York Times
Europe’s efforts to establish greater control over artificial intelligence technology are entering a new phase as the EUROPA consortium, led by Italian company Domyn, advances plans to develop one of the continent’s largest open-source AI models.
The initiative, supported by the European Commission and European high-performance computing infrastructure, aims to create an advanced artificial intelligence system capable of competing with models developed by leading American and Chinese companies.
Domyn chief executive Uljan Sharka told the Financial Times on 29 October that Europe possesses the technological resources and industrial expertise required to compete in advanced AI development.
However, he also identified two obstacles: the need for greater investment and the difficulty of persuading European businesses to adopt domestically developed systems.
These challenges extend beyond the development of a single language model.
They concern whether Europe can establish a commercially sustainable AI industry while reducing dependence on technology providers headquartered outside the EU.
The European Commission selected EUROPA in June 2026 as the winner of its Frontier AI Grand Challenge.
The competition, launched in February, was designed to support the development of a European general-purpose AI model with more than 400 billion parameters.
EUROPA’s planned model is expected to contain approximately 500 billion parameters and support all 24 official EU languages.
The consortium brings together Domyn and Germany’s Fraunhofer research organisation, including the Fraunhofer IAIS and IIS institutes.
Its work will rely on the EuroHPC network of European supercomputers.
Under the competition arrangements, the winning consortium can obtain access to as much as 2.5 per cent of EuroHPC’s total computing capacity for one year.
This allocation is intended to address one of the largest barriers to advanced AI development: the enormous computing resources required to train models at this scale.
The final system is intended to be released under an open-source licence, allowing organisations to inspect, adapt and deploy the technology.
Nevertheless, parameter count alone does not demonstrate that a model is competitive.
Its actual performance will depend on training data, architecture, computational efficiency, evaluation results and the ability to perform reliably across different tasks.
One of EUROPA’s distinguishing objectives is its coverage of all official EU languages.
The commercial AI market has historically been dominated by models developed primarily for English-language applications, although leading international systems increasingly support multiple languages.
For Europe, linguistic capability has particular relevance to public administration, legal services, education and cross-border business.
A model designed specifically for the European market could potentially support specialised applications involving national legislation, administrative procedures and technical terminology.
However, multilingual availability will need to be assessed against measurable performance standards.
A system capable of producing fluent responses in a particular language is not necessarily equally reliable when interpreting specialised legal or scientific information.
The consortium will therefore need to demonstrate that linguistic coverage translates into practical capabilities rather than remaining principally a technical specification.
The EUROPA project forms part of a wider effort to use European publicly supported computing infrastructure for commercial and scientific AI development.
The EuroHPC Joint Undertaking has invested in systems intended to provide advanced computing resources to researchers, businesses and public institutions.
Additional AI-focused infrastructure is being developed through the EU’s network of AI factories.
These facilities are intended to make specialised computing capacity available to organisations that would otherwise face considerable costs when developing or adapting AI systems.
However, infrastructure ownership does not necessarily provide complete technological independence.
European supercomputing projects continue to rely on internationally sourced semiconductor technology and specialised software.
Domyn itself has described a partnership with the American technology company Nvidia in its previous model-development work.
Consequently, European AI sovereignty is better understood as a question of control over important technological capabilities than complete separation from international supply chains.
On 20 October, Domyn announced the initial release of Domyn Edge, a separate AI model containing four billion parameters.
According to the company’s technical announcement, the system was trained from scratch using approximately 13 trillion tokens on Europe’s Leonardo supercomputer.
Domyn has committed to making the model’s weights, training code, intermediate checkpoints and datasets available under open licences.
This approach is intended to improve reproducibility and allow external developers to examine how the model was constructed.
Domyn Edge is substantially smaller than the planned EUROPA system and should not be confused with it.
Nevertheless, the project provides an example of the company’s approach to developing AI models using European computing infrastructure.
The EU has already established an extensive regulatory framework for artificial intelligence through the AI Act.
However, legislation governing the development and deployment of AI does not, by itself, create a competitive domestic technology industry.
Europe faces competition from American companies with substantial private investment, established cloud platforms and access to large commercial markets.
Chinese developers have also expanded their capabilities, including in open-weight AI systems.
EUROPA’s commercial prospects will therefore depend on more than technical performance.
European businesses must be willing to integrate the resulting technology into their operations, while developers will require dependable access to computing infrastructure and support services.
There is also the question of whether European institutions themselves will adopt domestically developed AI systems where they meet relevant security and performance requirements.
Without sufficient demand, even a technically successful model could struggle to establish a sustainable ecosystem.
Domyn has indicated that it aims to release EUROPA during 2027, although public statements have differed over the precise timetable.
The company’s leadership has discussed a June target, while an earlier technical announcement referred to the second half of 2027.
The immediate challenge is completing development and demonstrating that the model can compete on performance, reliability and cost.
The longer-term challenge concerns adoption.
If European businesses and public institutions continue relying predominantly on foreign technology, access to an advanced European alternative may have only a limited effect on the market.
Conversely, a competitive and widely adopted open-source model could offer organisations greater control over deployment, customisation and sensitive information.
EUROPA is therefore more than a publicly supported technology experiment.
It is a test of whether Europe can convert research institutions, computing infrastructure and industrial expertise into AI products capable of competing in the international market.
The outcome will depend not simply on whether Europe can build an advanced model, but on whether organisations consider it sufficiently capable and economical to use.
U.S. Immigration and Customs Enforcement (ICE) has a liaison at the military's Special Operations Command (SOCOM) with access to advanced surveillance technology, leaked files show. The files also reveal ICE liaisons at Africa Command (AFRICOM) and Southern Command (SOUTHCOM) with access to the same tech, and an ICE contractor associated with U.S. Central Command (CENTCOM) in the Middle East.
On Wednesday, a hacker group calling itself the Department of Peace released a massive collection of U.S. Department of Homeland Security (DHS) personnel files, including ICE and other domestic security agencies. "They have our information, so it's only fair that we have theirs," the hackers wrote in a note attached to the database, which was posted online to Distributed Denial of Secrets, a platform for data leaks.
DHS told The Washington Post that it was "aware of a potential breach" and "working to address the situation swiftly and effectively." A website hosting a searchable version of the leak, ICEpick, was taken down after the DHS sent a letter demanding that the domain name host "stop serving the personal contact and location data of individuals who did not consent."
The files include an ICE Homeland Security Investigations (HSI) unit chief who appears to oversee the agency's military liaisons. His subordinates include a "Combatant Command Program Manager" at SOCOM headquarters in Tampa, Florida; an "HSI Liaison to USAFRICOM" in Stuttgart, Germany; and a "Senior Liaison to U.S. Southern Command" in Doral, Florida.
All of the officials named in the files are listed as users of "Penlink," "Falcon," and "RavenGo." The contractor PenLink offers officials the ability to track individual cell phone users or collect data on everyone in a certain location, according to documents obtained by 404media. FALCON, created by the contractor Palantir, allows officials to combine records from governmental and commercial databases, The Guardian reports. The DHS eventually phased out FALCON in favor of an in-house system, RAVEN.
The unit chief is additionally listed in a group called "LCR_Users." The acronym may refer to the Office of the Director of National Intelligence at Liberty Crossing, McLean, Virginia. Two employees are listed in the DHS files with "ODNI" in their title and "LCR - Liberty Crossing" as their location. The first Trump administration reportedly looked into merging ICE into the federal intelligence community, but never officially did so.
The files also include an HSI contractor listed as "CENTCOM analyst" at the National Targeting Center in Sterling, Virginia. Unlike the other liaisons, this contractor does not have access to ICE's own surveillance software. The National Targeting Center is a Customs and Border Protection (CBP) unit that surveils international travel and selects passengers for questioning. CENTCOM oversees all U.S. military forces in the Middle East.
Although it has been mentioned in public, ICE's relationship with the military is shrouded in mystery. Of course, the best-documented relationship is with SOUTHCOM, responsible for U.S. forces in Latin America. Official press releases also mentioned HSI's participation in an AFRICOM maritime security training exercise in 2021. The connections to SOCOM and CENTCOM have not been publicly reported.
ICE did not respond to Reason's request for comment.
The DHS has an in-house intelligence office that is separate from HSI and publicly cooperates with the U.S. military, as well as the rest of the intelligence community. But HSI's "global footprint" and ability to straddle the line between law enforcement and foreign policy "make it uniquely capable of disrupting and dismantling the sophisticated groups and networks that we now face," HSI Executive Assistant Director Katrina W. Berger testified to Congress in 2024.
HSI is supposed to handle serious organized crimes involving the border rather than day-to-day deportations. However, the Trump administration has turned the directorate into an all-purpose domestic spying and enforcement arm. HSI is now running a massive surveillance operation against left-wing protesters, WIRED recently revealed.
Agents have reportedly been pulled away from investigating human trafficking and drug cartels to deal with immigration enforcement. HSI was involved in arresting former CNN journalist Don Lemon for showing up at an anti-ICE protest at a church, and led the raid against Rep. Ilhan Omar (D–Minn.) reportedly aimed at showing that her U.S. citizenship is invalid.
Last month, ICE announced upcoming contracts for a $100 million, fast-tracked expansion of its domestic surveillance capabilities, with contractors asked "to identify and assess relationships among individuals, organizations, locations, events, activities, data points, and other entities." Former military intelligence analyst Michael Wriston called the intensified ICE surveillance a case of "imperial boomerang" in an interview with The American Prospect. "These are the kinds of things that happened in Afghanistan, Iraq, Yemen, Syria, for years. Now that methodology is being brought into the ICE space."
The post Leaked Files Reveal ICE Liaisons to U.S. Military Commands With Access to Surveillance Tech appeared first on Reason Magazine.
Shafkat Anowar/The Dallas Morning News via Getty Images
Healthcare has long used complex contracts that can obscure who pays for what. Mark Cuban says data center developers are using a similar playbook.
"A lot of communities have been really dumb with the deals they've negotiated for data centers," Cuban said on an episode of "The A16z Show" released on Wednesday. "They've fallen for the contractual terrorism in a lot of cases."
Cuban, who co-founded Cost Plus Drugs, a company that sells generic medicines at a transparent markup, previously coined the phrase "contractual terrorism" while discussing healthcare.
On the A16z podcast, he said massive healthcare insurers and their subsidiaries use sprawling, opaque agreements to obscure costs and make it difficult for employers to know where their money is going.
Cuban said those contracts are "so big and so long" that insurers "get to be contractual terrorists" — a formulation he also applied to communities' data center deals.
His comments come as the data center boom has become a hugely contentious issue in communities across the US.
Residents living near facilities have reported incessant low-frequency hums, diesel smells, intrusive lights, construction disruption, pressure on water and electricity supplies, and the loss of trees and open land.
In Northern Virginia, the country's data center capital, local opposition has helped reshape elections, prompted lawsuits, and pushed residents to demand tighter noise and zoning rules.
A University of Massachusetts Amherst poll of 1,000 Americans from September found that 65% opposed building an AI data center in their local community, including 49% who said they strongly opposed it.
The race to build AI infrastructure has also produced far more announcements than operating facilities.
Aterio, which tracks projects through satellite imagery, permits, and other records, counted 8,268 announced US facilities as of October 20. But only 2,126 were active, and 861 were under construction.
Thousands more had just been announced, leaving local governments to negotiate consequential long-term agreements for projects that may never progress beyond a press release.
That is a risk Cuban has warned about.
In a series of X posts in September, he laid out a 10-point playbook for communities considering their first data center, arguing that every promise from a developer should come with a measurable limit, independent oversight, financial backing, and a remedy if it is broken.
Cuban's advice is to settle those questions before construction begins.
"Resolve it before granting the approval that spends the last of the town's leverage," Cuban wrote at the time.
© The New York Times
© Julia Demaree Nikhinson/Associated Press
Microsoft
Microsoft's head of the Copilot platform, AI agents, and now Office is making good on his promise to reshape how his teams work.
"The most important thing my team will do this year won't be any single product or feature we ship: instead, it will be changing how we work," Charles Lamanna, the new boss of this key Microsoft unit, wrote in August.
Lamanna has started requiring senior staff to work in offices four days a week, and he's offering relocation to major hubs, according to people familiar with the changes. Teams within his organization have also reorganized to a flatter structure removing middle managers in line with Lamanna's views about how agentic AI is changing engineering.
"We're rethinking where managerial layers are critical and where flatter structures can enable faster decision-making, clearer ownership and greater autonomy," a Microsoft spokesperson said.
The moves offer an early look at how Microsoft is reworking its software empire for the AI age. Lamanna, who already oversees the company's Copilot and AI agent platforms, now controls the Office productivity software and Teams workplace chat service that are used by millions of workers.
Lamanna took over those groups earlier in October when the previous boss, Ryan Roslansky, announced plans to leave at the end of the year. Roslansky was seen as a successor to Microsoft veteran Rajesh Jha, who retired in July. Roslansky and Lamanna were two of the three executives who made up the new Copilot leadership team.
The changes come as CEO Satya Nadella reshapes the company around AI. Microsoft has been consolidating responsibility for its AI products, and a wave of senior executive departures has made way for a new generation of leaders.
Lamanna's expanding responsibilities make him a key player in this leadership transition, giving him oversight of Microsoft's flagship workplace software and the AI tools that could change how customers use its products.
Recently, the Teams unit now under Lamanna reorganized by increasing the "span of control" for managers in engineering, according to two of the people familiar with the changes.
This refers to the number of employees who report to each manager, and it's become common for tech companies to flatten engineering organizations like this in the AI era, as coding tools and agents write more software.
The reorganization included converting managers to individual contributors, the people said. The idea is that agentic coding capabilities mean Microsoft needs fewer managers, according to one of the people. The team will now have a ratio of one manager per 12 individual contributors.
"Agentic coding capabilities have changed every level of the engineering function," Microsoft's spokesperson said. "Software engineers are no longer hand-assembling every line of code. Instead, they are now designing and directing the 'agentic coding machine' that helps build the product."
As AI makes it easier to build software, skills like making sound technical decisions, designing systems, ensuring quality and security, and understanding customer needs become more valuable.
"At the same time, some tasks that traditionally required managerial coordination, such as managing handoffs and orchestrating work across teams, become less central," the spokesperson said.
Other teams have implemented similar restructuring, the spokesperson confirmed.
Lamanna has also started reshaping where employees work. Last week, Lamanna's Copilot, Agents, and Platform (CAP) organization offered select US employees the option to relocate to offices at Microsoft's Redmond, Washington headquarters, San Francisco, or New York, according to three people familiar with the plans.
The people who spoke with Business Insider for this story asked not to be identified discussing internal matters.
Lamanna has also asked employees at the "partner" level and above — some of Microsoft's most senior staff — to start working in physical offices at least four days a week. The companywide standard is three days a week.
Have a tip? Contact this reporter via email at astewart@businessinsider.com or Signal at +1-425-344-8242. Use a personal email address and a nonwork device; here's our guide to sharing information securely.
Although artificial intelligence is drawing criticism for its energy use, it's also being used by firms such as ExxonMobil and BP to unleash new energy production. The research firm Wood Mackenzie even estimates that AI could unlock 1 trillion extra barrels of oil globally by helping companies "wring substantially more oil out of producing reservoirs."
These developments are good news for consumers, but they could have a significant impact on the environment, according to an August paper published in npj Climate Action. The authors find that AI could prolong oil and gas usage, leading to a 0.47–1.8 gigaton increase in annual global carbon dioxide emissions. The lower end of that estimate is comparable to Mexico's annual carbon dioxide emissions, study co-author Holly Alpine told The Wall Street Journal. The higher end is equivalent to about 5 percent of the world's energy emissions from 2024.
The researchers say net emissions reductions cannot occur unless renewables grow four to five times faster than oil and gas. (They do note there are other ways AI could reduce the greenhouse gas footprint of the fossil fuel industry—namely, "methane leak detection or efficiency gains in power generation.")
This is an ambitious task, considering that fossil fuels account for over 80 percent of the world's energy needs and their peak usage is years away. But there are reasons to be optimistic about AI's role in accelerating the use of clean technologies.
In the same way that oil majors are using AI to find their next well, startups are utilizing the technology to unleash geothermal energy. One such company is Zanskar, which uses a combination of AI, advanced drilling methods, and geoscience to "accurately map the Earth's subsurface and develop geothermal resources that have been overlooked or considered too risky."
At the Lightning Dock project in New Mexico—which Zanskar purchased in 2024—the company's technology has reduced surface infrastructure costs by 75 percent and sped up drilling by 35 percent. Zanskar says Lightning Dock is now "home to the most productive pumped geothermal well in the U.S. and quite possibly the world." Late last year, the company discovered "Big Blind," a naturally occurring geothermal system in Nevada with the potential to power 80,000–100,000 homes with reliable carbon-free energy. In January, it closed a $115 million funding round.
Nuclear power is also benefiting from advances in AI. In August, Atomic Canyon released its Nuclear Industry Virtual Assistant with backing from Nvidia. The tool is designed to help nuclear professionals navigate technical and regulatory landscapes, which could lead to more nuclear power being developed and deployed faster. The industry has also seen a wave of investment from tech companies looking to fuel their power-hungry data centers with clean energy. Big Tech firms such as Google, meanwhile, have poured significant resources into renewable and battery storage projects to support data centers and bring more energy to the grid.
AI is here to stay. We should embrace what it is already starting to give us: a cleaner and more abundant energy future.
The post AI Isn't Just Extracting More Fossil Fuels—It's Unlocking Renewable Energy Too appeared first on Reason Magazine.
Diego Herrera Carcedo/Anadolu via Getty Images
A new NATO artificial intelligence model gives Ukraine's forces a frontline map of where Russian drone activity is expected the next day, per the alliance's innovation center.
The system, called Initiative 13 or Model 13, "turns an array of disparate data into a clear map of expected activity" for commanders to plan their next day of operations, a spokesperson for NATO's Joint Analysis, Training, and Education Center, or JATEC, told Business Insider.
Business Insider reported on Ukraine's announcement of the program earlier this month, but JATEC is now providing new details on how it works.
Initiative 13 covers areas 5 to 50 km, or roughly 3 to 31 miles, of Ukrainian territory behind the line of contact, and divides that space into geographical cells, per JATEC.
"For each cell, the system determines the expected intensity of strike and reconnaissance activity of UAVs, the level of threat, the trend of its change, and the range of possible values," its spokesperson said.
The map shows a forecast for the next day, helping commanders to highlight risk areas so they can better deploy their forces and equipment, they said.
They added that the AI model is trained on battlefield data and considers previous Russian drone activity, what types of drones were used, when and how they were deployed, weather conditions, terrain, and distance to the line of contact.
"Initiative 13 does not replace data from the current situation and commander's decisions," the JATEC spokesperson added. JATEC is NATO's joint organization with Ukraine, based in Poland, that aims to glean lessons and innovations from the war.
Ukraine's defense ministry said in mid-October that Initiative 13, which it called Model 13, was "ready for integration into the Armed Forces of Ukraine's situational awareness systems."
The AI model, it said, could predict Russia's drone strikes with "high probability."
Kyiv's forces use a battlespace management system called Delta, which feeds data from the front lines onto an interactive map that commanders can access via laptop, phone, or tablet.
The drone-predicting AI was developed by Ukraine's forces and NATO's Allied Command Transformation, with support from JATEC.
Ukraine's defense ministry said in its initial announcement that another AI model, dubbed No. 32, for planning artillery deployments was also in the works but is still in prototype stages.
Artificial intelligence has increasingly been making its way into Ukraine's war, from helping mid-range attack drones identify and strike targets to sending uncrewed ground vehicles to autonomously complete supply runs.
Login