According to The Information, Nvidia, Microsoft, and Amazon are in discussions to jointly invest up to $60 billion in OpenAI, with term sheets reportedly close to being finalized. Nvidia plans to invest up to $30 billion; Microsoft intends to invest less than $10 billion; Amazon may contribute over $20 billion. This round is described as part of OpenAI's plan to raise up to $100 billion, potentially valuing the company at $730 billion pre-funding; the involved companies have not yet publicly confirmed the deal.

Apple has announced the acquisition of Israeli audio AI startup Q.ai, with reports indicating a transaction value approaching $2 billion, marking one of Apple's largest acquisitions in recent years. Q.ai specializes in audio machine learning technologies such as speech enhancement in noisy environments and whisper/low-voice recognition, with potential applications in devices like AirPods and Vision Pro. Apple confirmed that Q.ai CEO Aviad Maizels and founding team will join the company, though specific financial terms and integration plans were not disclosed.

Google has released the Project Genie experimental prototype, now accessible to Google AI Ultra subscribers in the U.S. Based on Genie 3, the feature allows users to generate and interactively explore virtual worlds using text or image prompts, supporting path generation, physics simulation, multi-view control, and enhanced controllability via Gemini and related technologies. Current limitations include limited realism, delayed character control, and a 60-second cap per generation session; Google stated it will gradually expand access.

Tesla disclosed a $2 billion investment in xAI, Elon Musk's AI venture, during its earnings call, and announced the discontinuation of Model S and Model X production. The company plans to repurpose manufacturing capacity and facilities for the Optimus humanoid robot, targeting an annual production capacity of 1 million units, supported by approximately $20 billion in investments for new production lines and AI infrastructure. Tesla noted Grok has been integrated into its vehicles; FSD subscription users rose from 800,000 in 2024 to 1.1 million in 2025, and paid Robotaxi trials without safety drivers have launched in Austin.

Toronto-based autonomous driving company Waabi has raised $1 billion in funding, described in a newsletter as one of the largest single financings in Canadian tech history. Uber participated with a $250 million investment and committed to deploying at least 25,000 driverless taxis powered by Waabi's system on its platform. Waabi focuses on end-to-end AI models for autonomous driving, aiming to use a unified system for both autonomous trucks and robotaxis, targeting commercial freight operations by 2027. Transaction and deployment details await further public disclosures.

AI customer service startup Decagon has secured $250 million in funding, according to a newsletter, led by Coatue and Index Ventures, raising its valuation to $4.5 billion—nearly triple its previous round. Founded in 2023, the company already serves over 100 large enterprise clients, positioning its product as an AI agent capable of continuous 'active learning' in customer support scenarios to improve issue resolution efficiency and consistency. Funds will be used to expand the product and team and accelerate enterprise adoption, though revenue figures and profitability were not disclosed.

Reuters reports that Chinese AI chipmaker Axera Semiconductor plans to raise HK$2.96 billion (approximately $379.2 million) through a Hong Kong IPO. The proceeds will support further development of AI chip technology and related business operations. Axera targets high-growth markets including smart security, autonomous driving, and edge computing. Key terms such as valuation, underwriters, and timeline have not been disclosed; further developments will depend on official prospectus filings and regulatory disclosures.

arXiv has announced that starting February 11, 2026, all submissions must be in English or accompanied by a complete English translation (previously only abstracts required translation). The platform states this aims to aid moderator review and broaden readership, but has raised concerns among non-native English researchers about increased burden, potentially diverting submissions to alternative platforms like HAL. arXiv permits AI tools for translation but recommends human proofreading; the policy has also sparked broader discussion on AI translation quality and accountability in academic writing and peer review.

The UK government has issued a 'Secure AI Infrastructure: Call for Information', jointly launched by DSIT, the AI Safety Institute (AISI), and the National Cyber Security Centre (NCSC), soliciting insights from the AI and cybersecurity sectors and academia on infrastructure protection challenges and viable solutions. The consultation focuses on risks including model weight theft, data leakage, and system behavior tampering, and outlines technical directions such as cross-domain solutions, model digital rights management, confidential computing, advanced cryptography, inter-system defense, observability, and adversarial machine learning defenses. Feedback is due by February 28, 2026.

Cisco has released Foundation-sec-8B-Reasoning, an open-weight cybersecurity reasoning model designed for threat modeling, attack path analysis, and risk assessment workflows, supporting local/offline or air-gapped deployments. It claims superior performance over models like Llama 3.1 8B on certain security benchmarks; when paired with LlamaGuard, harmful content detection reaches 98.25%. Additionally, Cisco AI Defense has open-sourced Skill Scanner, which detects prompt injection, data leakage, and malicious code patterns in agent 'skill' configurations, supporting CI/CD integration with SARIF report output.