The United States opened its criminal case against Huawei in Brooklyn federal court on Wednesday, with a prosecutor describing two decades of conduct by the world’s largest telecoms equipment maker as theft, lies and cover-up. “Theft, lies, cover-up. For 20 years, that’s how Huawei victimized American companies,” said Taylor Stout, a Justice Department trial attorney, according […]

IVEX, a Heverlee-based DeepTech startup developing technology for automotive safety testing, has raised €5 million in Series A to accelerate its engineering roadmap and expand its presence in North America and Asia. 

The round was backed by existing investors, The Faktory and Dedicated.lu, as well as private investors. Alongside the funding round, IVEX has strengthened its board of directors with Jan Leuridan, former CEO of Siemens Industry Software NV; Omar Mohout, Director at Deloitte; Simon Alexandre, Partner at The Faktory; and IVEX co-founder Quentin de Clercq, who now leads AI strategy at ASML in the United States. They join IVEX co-founder and CEO Mario Torres on the board.

Torres, co-founder and CEO of IVEX, said, “Vehicle safety testing generates enormous amounts of data, but much of its analysis still depends on repetitive manual work. AI and rigorous mathematical algorithms allows us to make the testing process significantly faster and more scalable while keeping engineers in control and ensuring that every result remains traceable to the original data. With this funding, we can deepen the capabilities of our platform and bring IVEX technology to North America and Asia.”

Founded in 2018 as a spin-out from KU Leuven, IVEX is on a mission to help companies to create the safest vehicles on earth. The company provides technology for the analysis, validation, testing, and safety scoring of vehicles, with a focus on evaluating Advanced Driver Assistance Systems (ADAS) and Autonomous Driving (AD) systems.

According to the company, assessing increasingly sophisticated automotive technologies such as driver assistance systems (ADAS) and autonomous driving systems (AD) requires testing organisations and manufacturers to review large volumes of camera, sensor and vehicle data. 

It notes that in on-road tests, engineers have traditionally had to examine footage, identify road signs and conditions, compare them with dashboard information and determine whether the vehicle responded correctly. It opines that such manual analysis cannot scale without a proportional increase in engineering work.

IVEX claims that it was founded to address this problem. Its platform uses computer vision, AI and cloud-based data processing to automate the most time-consuming parts of on-road vehicle safety testing. 

It further states that the technology can simultaneously interpret the car’s surroundings, capturing, for example, speed limits and other signs, reading the vehicle’s displays to the driver, and automatically comparing multiple signals. 

IVEX says that the resulting analysis helps determine key information, such as whether the vehicle recognised the relevant road information and responded as required by the applicable safety protocol.

The company also highlighted that it operates with a human-in-the-loop model, allowing engineers to review the platform’s output and retain control over the final assessment. It argues that IVEX’s technology can reduce the time engineers spend analysing data by more than 95%. It reported one example wherein the platform processed 400 hours of multi-sensor driving data in only 8 hours of engineering time.

The Belgian startup’s long-term goal is to build the operating system for automotive safety testing: an AI-powered platform on which various safety protocols, testing requirements, and vehicle functions can run as applications. 

It highlighted that the platform’s underlying models improve as IVEX processes data from additional vehicles and driving environments. This creates a continuously expanding foundation for current and future automotive safety-testing applications.

The company plans to use this capital to build on its position in Europe and to establish a stronger presence in other major global automotive markets, including North America and Asia. It also intends to bring its AI-powered testing and data-processing technology to regulators, testing institutions and automotive companies operating in these regions. 

IVEX will also continue expanding the range of NCAP protocols supported by its platform and strengthening the underlying infrastructure required to process larger volumes of data across additional countries, vehicle types and testing environments.

The company is a trusted supplier to Euro NCAP, providing technology used in official vehicle assessments. Its platform has also been deployed in on-road testing campaigns across more than 20 countries, including markets throughout the EU, the UK, Australia, Brazil, and Japan. 

The startup has accumulated more than 100,000 kilometres of driving data and worked with data from more than 50 vehicle models. IVEX holds six patents, including one in the United States, and is ISO 27001 certified. In February 2023, the company raised its Seed funding round backed by PMV, SPDG, The Faktory Fund, Noshaq, and Dedicated. 

The post KU Leuven spin-out IVEX lands €5 million to accelerate AI-powered automotive safety testing appeared first on EU-Startups.

The US Justice Department is examining whether Nvidia deliberately structured its licensing deal with inference chip designer Groq in a way that avoided antitrust review, according to a New York Times report. The Justice Department opened its inquiry shortly after the deal was announced in December and has since sent Nvidia a formal request for […]

Bernie Sanders will convene a private, bipartisan briefing for senators on the risks of artificial intelligence on 16 September. Geoffrey Hinton, Max Tegmark and Ajeya Cotra will address the room for an hour on Capitol Hill. The briefing was reported by Axios, byline Maria Curi. Sanders said AI has been receiving enormous attention in recent […]

Paris-based AI company Arlequin AI has raised €28 million in Series A funding to accelerate the development of its topological neural network architecture and expand the international deployment of its technology. The round was co-led by redalpine and OTB Ventures, with participation from Bpifrance’s Defence Innovation Fund. Existing investors Vsquared Ventures and 10x Founders increased their stakes, while Xavier Niel also joined the round.

Founded in 2024 by Hugo Micheron, a specialist in terrorism studies and geopolitical instability, and Antoine Jardin, a former CNRS research engineer specialising in data science and human behaviour, Arlequin develops AI systems for organisations making consequential decisions from large and fragmented datasets.

The company’s platform analyses relationships across different types of information, including documents, transactions, video and operational data. It is designed to help users identify connections within complex datasets while allowing them to trace results back to the underlying evidence.

Arlequin is now developing proprietary models based on topological neural networks, an AI architecture designed to learn from both data points and the relationships between them. The company says the approach can capture complex interactions involving multiple elements at once, enabling analysis of systems that become difficult to interpret as datasets grow in scale.

Today, another revolution is taking shape: the development of new AI systems capable of understanding highly complex dynamics hidden within millions of data points,

said Hugo Micheron, CEO and co-founder of Arlequin AI.

Arlequin sees potential applications in security and defence, criminal investigations, fraud and money-laundering detection, information integrity, cybersecurity, and AI safety and security. Its technology is already being used by governments and large organisations across Western and Eastern Europe.

Antoine Jardin, CTO and co-founder, said the company believes further AI advances will require different architectures rather than simply larger models trained on more data and computing power. Arlequin’s approach is also intended to require less compute than large-scale AI systems, reducing its reliance on energy-intensive infrastructure and advanced semiconductors.

The new capital will be used to expand the international team developing and training its proprietary models, and to support commercial deployment in Europe and internationally. The company has opened offices in London and Berlin and plans to establish an AI lab in Silicon Valley in the coming months. 

OpenAI has called for mandatory, capability-based national AI safety regulation in the United States, setting out a list of requirements it wants imposed on itself and its competitors, and urging Congress to legislate before it adjourns. The proposal was published on Wednesday by Chris Lehane, the company’s chief global affairs officer. It asks for common testing […]

Bynario, a Milan-based cybersecurity company whose researchers used frontier AI models to uncover serious flaws in Apple’s operating systems, has closed a €2.1 million pre-Seed funding round.

The round was led by 360 Capital Partners, with participation from PranaVentures.

“Attackers are fundamentally incentive-driven. As AI continues to reduce the cost of vulnerability research and exploitation, more exploitable vulnerabilities become attractive targets. The challenge for defenders is only becoming more acute,” said Alfredo Pesoli, CEO and co-founder of Bynario.

Founded in 2025 by Pesoli, Giancarlo Russo and Lorenzo Cavallaro, Bynario builds an autonomous application security platform designed to help organisations identify, validate, prioritise and remediate software vulnerabilities at scale, spanning code, cloud infrastructure and the software supply chain.

Instead of flagging every possible issue, Bynario’s security platform confirms which vulnerabilities are actually exploitable in a customer’s environment, allowing security teams to prioritise their work. The company is already working with customers across software, cloud and enterprise environments, and says it is helping them cut remediation times while giving them clearer visibility into software risk.

Bynario shot to fame after its researchers used frontier AI models as part of an offensive security research programme, uncovering serious vulnerabilities in Apple’s macOS Screen Sharing technology in August. The findings were disclosed to Apple, which subsequently issued fixes for the affected vulnerabilities.

According to Verizon’s 2026 Data Breach Investigations Report, software vulnerabilities are now the most common route into enterprise systems, accounting for 31% of incidents. As tools push developers to ship more code, the sheet volume makes it hard for security teams to separate real risk from noise.

Bynario’s Apple vulnerability findings immediately showcased how quickly AI tools can now accelerate the discovery and exploitation of previously unknown software flaws, intensifying pressure on defenders who still largely rely on manual investigation.

“Bynario’s approach of autonomous vulnerability discovery, validation, prioritisation, and remediation is what is needed in modern cybersecurity. We are proud to stand alongside them as they build this game-changing AI-based platform from the ground up,” said Cesare Maifredi, Partner at 360 Capital.

While its claim to fame is vulnerability discovery, Bynario wants to go beyond that. The Milan-headquartered startup wants to build a continuous exposure management layer that keeps pace with how fast modern software and the threats against it are evolving. With the cybersecurity skills shortage continuing to leave many organisations unable to keep pace with a growing volume of vulnerabilities, Bynario is betting that autonomous, AI-driven defence is the alternative to manual triage.

The fresh capital will go toward accelerating the development of Bynario’s AI-powered security platform, growing its engineering team, and scaling to meet enterprise demand.

The post Milan-based Bynario raises €2.1 million pre-Seed after AI research exposed serious Apple vulnerabilities appeared first on EU-Startups.

Cybersecurity company Bynario has raised €2.1 million in a pre-seed round to develop its AI-powered platform for identifying, prioritising and addressing software vulnerabilities. The round was led by 360 CapitalPartners, with participation from Prana Ventures.

Bynario is developing a suite of security products designed to help organisations manage software vulnerabilities across their systems, from code to production. Its technology combines AI with offensive security research to assess vulnerabilities and help security teams prioritise remediation.

The company is targeting a growing challenge for security teams as AI changes both software development and vulnerability research. AI-assisted development is increasing the speed at which software is created, while security researchers and attackers can use advanced models to identify vulnerabilities more quickly. At the same time, organisations are managing growing numbers of applications, dependencies and third-party components, increasing the volume of potential security issues they need to assess.

Alfredo Pesoli, co-founder and CEO of Bynario, said falling costs associated with AI-assisted vulnerability research could make a wider range of software flaws attractive targets for attackers, increasing the pressure on security teams to identify and address relevant risks.

Defense at scale needs the same depth as offense. Attackers are fundamentally incentive-driven. As AI continues to reduce the cost of vulnerability research and exploitation, more exploitable vulnerabilities become attractive targets. The challenge for defenders is only becoming more acute,

Pesoli said.

Bynario has also used advanced AI models in offensive security research. In a recent programme focused on Apple’s macOS Screen Sharing technology, its researchers uncovered previously unknown vulnerabilities, which were disclosed to Apple and subsequently addressed through software fixes.

The research reflects the company’s broader focus on helping organisations distinguish relevant security risks from a growing volume of findings. Bynario says it is already working with customers across software, cloud and enterprise environments.

The funding will be used to further develop Bynario’s security platform, expand its engineering team and support its work with enterprise customers. 

Grubel, a Munich- and Tübingen-based AI research lab developing systems that adapt to individual legal matters, has raised €3 million in pre-Seed funding to support research, product development, and team expansion.

The round was led by venture capital firm Point Nine, with participation from angel investors including Jeff Dean (Google’s former chief scientist), Chris Ré (Stanford professor and Together AI co-founder), Ion Stoica (UC Berkeley Professor and Databricks co-founder), and Harvey co-founders Gabe Pereyra and Winston Weinberg, among others.

“Complex legal matters do not come with ready-made training data, task environments, or clear tests of success. We are automating the AI specialisation loop that identifies the right information, curates it into training data and environments, adapts the system to the matter, and tests whether its work meets the required standard,” said Reinhard Heckel, co-founder of Grubel. 

Founded in 2026 by machine-learning researchers Moritz Hardt and Heckel, Grubel is automating AI specialisation: gathering and preparing relevant information, adapting the system to the legal work at hand, and testing whether its work meets the required standards.

Hardt is a director at the Max Planck Institute for Intelligent Systems and a former UC Berkeley professor and Google Brain researcher. He has co-developed test-time training and co-authored Lawma. Heckel is a TUM professor of machine learning, currently on leave and formerly a researcher at IBM Research. He focuses on data-centric machine learning and helped create DataComp-LM and OpenThoughts. 

According to the company, complex knowledge work beyond coding remains surprisingly dependent on highly skilled human labour, with AI still playing only a limited role. Grubel opines that this gap is particularly apparent in legal services.

The company notes that legal AI has lagged in part because complex legal work is both highly specialised and very difficult to evaluate. It often requires expertise specific to an individual matter, including its facts, documents, terminology, prior work product, and client standards. 

It further notes that today, teams of engineers and lawyers build specialised AI systems for specific legal workflows by curating data, adapting the systems, and evaluating their performance. Although this specialisation is powerful, it requires substantial manual effort. It states that an engineering team cannot build and maintain a separate system for every legal matter. 

Grubel claims to be automating that specialisation process. Its thesis is that every matter needs its own AI. To make that possible, the company is automating the process of specialising an AI system for each matter. Its specialisation loop has three components: a data engine that discovers and curates task-relevant data for each matter or client; a test-time adaptation layer that adapts both the model and the agent to the matter before acting; and a continual evaluation framework that checks the work against standards derived from the matter itself. The loop repeats until the system delivers what the legal task demands. 

It’s bringing data curation, adaptation, and evaluation into a specialisation loop that trains and improves on the matter itself.

“Over the past two to three years, we’ve seen many legal AI companies emerge, but few are pushing the boundaries of AI research in the legal field as ambitiously as Reinhard, Moritz and the Grubel team. We believe their approach can unlock new capabilities in legal AI and, over time, other forms of complex knowledge work. We’re excited to support them on this journey,” said Louis Coppey, Partner at Point Nine.

The company plans to use this capital to support its research, product development, and team expansion. An initial evaluation showed that Grubel’s specialisation loop can improve performance over general-purpose frontier systems on the legal tasks tested. The company aims to further develop its automated specialisation loop to improve performance on complex legal work while allowing more routine tasks to be handled by smaller, more efficient models.

The post New German AI research lab Grubel raises €3 million to build systems that adapt to individual legal cases appeared first on EU-Startups.

Google Research and NASA JPL have released a deep learning model that maps methane plumes globally at 60 metres, finding about 50% more than human analysts and 23,000 additional plumes, including at 24 of the 25 largest-emitting landfills. Europe’s Methane Regulation ordered a satellite monitoring tool and a super-emitter alert system, but it covers only […]

Google’s autumn roundup of its AI subscriptions leads on Gemini Spark handling web errands through Chrome for US subscribers, alongside voice tools, an image editor and a spreadsheet app builder that do reach Europe. Spark itself is unavailable across the European Economic Area, Switzerland and the UK, and Google has given no reason. Google has […]

Abu Dhabi Global Market reported more than $100bn in AI-focused capital in its first-half results, alongside 54% growth in assets under management. The same release describes that sum as deployed from the centre in one place and held by entities in another, which are not equivalent claims. Roughly half the total is MGX, the state-backed […]

Dynamic Creatures, a Boston Dynamics spinoff, has launched to build robots for theme parks and retail, writing software and buying hardware from Unitree, AgiBot and Boston Dynamics. Europe’s Machinery Regulation covers AI safety functions from January 2027, and the French company that created this category was sold to a Shenzhen buyer last year. Dynamic Creatures, […]

Shipt and Instacart both launched AI assistants that build a grocery cart from a prompt or a photo on 9 September. Instacart said orders placed through its assistant contain more items than a typical order and exceed its $115 average basket, while Shipt’s announcement gives no equivalent figure and frames the product as relief from […]

Apple has made a foldable iPhone. It is called the iPhone Duo, it opens to 7.6 inches, and it starts at $1,999. The price is the news. Analysts expected $2,500. Samsung, which has had this category almost to itself for seven years, sells its Z Fold8 Ultra at $2,099. Apple arrived last and priced below […]

Apple’s new watches can replay the last 15 seconds of a conversation you just had. Press the Digital Crown twice and the words appear as text on your wrist. The feature is called Live Rewind. It arrives with the Apple Watch Series 12 and the Apple Watch Ultra 4. It will not work in the […]