Modelos, agentes, investigación y negocio de la IA.
SK Hynix has completed the largest initial public offering (IPO) by a foreign company in US history, raising $26.5 billion. The success of the operation reflects investor appetite for the semiconductor sector, but has also come with an explicit request from the US administration for the company to build new manufacturing plants in the country.
According to Reuters, DeepSeek is developing its own artificial intelligence chip, focused on inference. The move aims to reduce dependence on NVIDIA and Huawei, and lower the cost of running models at scale. The project is in early stages and has not been officially confirmed.
The European Central Bank has issued a directive requiring the largest banks in the eurozone to submit a detailed plan by October to counter cyberattacks powered by frontier artificial intelligence. The move reflects growing regulatory concern over the sophistication of AI-based threats.
Three of China's largest tech companies, blacklisted by the US military, are accessing frontier AI from OpenAI and Google through subsidiaries in Singapore. This legal loophole allows models like GPT-4 and Gemini to reach sanctioned entities, bypassing US export controls.
Deutsche Telekom and OpenAI have announced a collaboration to turn the German giant into an AI-native operator. The initiative spans customer service with conversational agents to network optimization and automation of internal workflows.
Unit 42 researchers have detected an active campaign where attackers contact victims via Teams, pretending to be tech support, to install the EtherRAT malware. The trojan, written in Node.js, affects Windows, Linux, and macOS and uses the Ethereum blockchain to obtain its command-and-control server.
Illinois becomes the first US state to mandate annual external audits for frontier AI models. The bipartisan law imposes fines up to $3 million and protects whistleblowers. A precedent that could accelerate federal regulation.
NVIDIA and Hugging Face have joined forces to integrate the Isaac GR00T 1.7 foundational robot model and the Isaac Teleop framework into LeRobot, the open source robotics library. This collaboration aims to standardize and democratize robot development, reducing cost barriers and fragmentation.
At the RAISE Summit in Paris, AMD CTO Mark Papermaster stated that the rise of agentic AI will require an increasing number of CPUs, not just GPUs. This contrasts with the current trend of prioritizing GPUs for training and inference.
Texas is experiencing a boom in AI-driven data centers, but the energy comes from new fossil fuel plants exploiting a regulatory loophole. The resulting pollution threatens local communities' health and contradicts climate goals.
87% of organizations believe their identity management is ready for AI, but 46% admit deficiencies. The AI security paradox reveals that trust outpaces verification, exposing companies to critical risks.
Cybersecurity firm Wiz has discovered GhostApproval, a vulnerability pattern affecting six of the most popular AI coding assistants. The flaw exploits symbolic links to trick users into approving malicious actions without being aware.