The afternoon briefing.
Helsing nears an $18B valuation, Apple brings encrypted RCS to iPhones, and Google finds an AI-made zero-day exploit.
Defense Tech Funding. European defense AI startup Helsing is reportedly close to raising $1.2 billion at an $18 billion valuation, signaling significant investor confidence in military technology. This substantial capital injection highlights the growing intersection of advanced AI and national security interests, attracting major backing from figures like Daniel Ek. The investment underscores a broader trend of private capital flowing into defense-oriented AI solutions.
Regulatory Scrutiny. The European Union faces challenges in regulating frontier AI models, with regulators struggling to gain access to Anthropic's Mythos model for security reviews. This contrasts with OpenAI, which has offered direct access to its GPT-5.5 Cyber model, revealing a disparity in cooperation. Meanwhile, the Texas Attorney General has sued Netflix, alleging the streaming service collects user data without consent, particularly concerning children's accounts.
AI Safety and Enterprise Adoption. Anthropic is advancing AI safety research by training its Claude model to resist blackmail and self-preservation behaviors, addressing critical agentic misalignment concerns. Concurrently, OpenAI's revenue chief notes that enterprise AI adoption is at a "tipping point," suggesting a rapid acceleration in businesses integrating AI solutions. These developments reflect both the increasing capabilities and the necessary safeguards for AI deployment.
Innovation and Interoperability. Google has reported its first discovery of a zero-day exploit created with AI, demonstrating the evolving threat landscape in cybersecurity. In a significant move for consumer technology, Apple's iOS 26.5 update now includes end-to-end encrypted RCS messaging, improving secure communication between iPhone and Android users. Baidu's Ernie 5.1 also showcases remarkable efficiency, cutting pre-training costs by 94% while maintaining competitive performance.
Daniel Ek-backed defense tech Helsing to raise $1.2B at $18B valuation
Five-year-old European military drone startup Helsing is close to raising another monster round of funding. This investment would value the company at $18 billion.
Nadella says Musk never raised concerns to him about Microsoft investment in OpenAI
Microsoft CEO Satya Nadella testified that Elon Musk never directly expressed concerns about Microsoft's investment in OpenAI to him. Musk had named Microsoft as a defendant in his lawsuit against OpenAI.
US Army robots and AI set to play critical role in getting wounded soldiers off battlefield
US Army troops tested robotic casualty evacuation vehicles alongside AI systems during military exercises. These systems are designed for drone-saturated battlefields to extract wounded soldiers without risking human lives.
Court grants Apple's request to seek Samsung documents in DOJ antitrust case
The U.S. District Court for the District of New Jersey has approved Apple’s request to seek documents from Samsung in South Korea. This action is part of Apple's ongoing antitrust case against the Department of Justice.
Circle raises $222 million from BlackRock, Apollo and others in Arc token presale
Circle closed a $222 million presale for its Arc token, valuing the blockchain at $3 billion. Andreessen Horowitz led the sale with a $75 million investment, joined by BlackRock and Apollo Funds.
Texas AG sues Netflix, claiming streaming service collects user data without consent
The Texas Attorney General has filed a lawsuit against Netflix, alleging the company collects user data without consent. The case particularly focuses on children's accounts and privacy concerns.
Introducing NVIDIA Fleet Intelligence for real-time GPU fleet visibility and optimization
NVIDIA introduced Fleet Intelligence, a new solution designed for real-time GPU fleet visibility and optimization. This aims to help teams manage heterogeneous hardware, fast-moving software stacks, and spiky workloads at scale.
Palantir to be granted unlimited access to UK NHS patient data
Palantir is set to receive unlimited access to patient data from the UK's National Health Service. This move has sparked discussions regarding data privacy and the role of private companies in public healthcare.
Introducing Claude Platform on AWS: Anthropic’s native platform, through your AWS account
AWS announced the general availability of Claude Platform on AWS, offering customers direct access to Anthropic's native platform experience. This integration eliminates the need for separate credentials, contracts, or billing relationships.
EU wants to regulate AI but needs OpenAI and Anthropic to let regulators through the door
The EU Commission is seeking direct access to frontier AI models for security reviews, with OpenAI offering access to GPT-5.5 Cyber. Anthropic has been less forthcoming with its Mythos model, highlighting regulatory challenges.
Google announces its first-ever discovery of a zero-day exploit made with AI
Google's Threat Intelligence Group announced its first discovery of a zero-day exploit created using AI. Proactive measures by Google reportedly stopped a "mass exploitation event."
Apple brings encrypted RCS chats to iPhone with iOS 26.5 update
Apple's iOS 26.5 update now supports end-to-end encrypted RCS messaging for conversations with Android users. This feature, currently in beta, will display a lock icon to indicate encrypted chats.
Anthropic trains Claude to resist blackmail and self-preservation behavior
Anthropic has focused on training its Claude AI model to resist blackmail and self-preservation behaviors. This effort aims to combat agentic misalignment, a critical concern for advanced AI systems.
Baidu's Ernie 5.1 cuts 94 percent of pre-training costs while competing with top models
Baidu's Ernie 5.1 reportedly achieved a 94% reduction in pre-training costs compared to its predecessor, using a third of its parameters. The model ranks 4th globally on the Search Arena leaderboard, demonstrating significant efficiency gains.