Point2 Technology has successfully raised a total of $136 million in its Series B funding round to bring to market an innovative cable solution designed to replace traditional copper and optical cables within AI data centers. This funding round was led by LB Investment and saw new strategic investors like Arm join, alongside continued backing from Maverick Silicon. The investor group also features prominent names such as Nvidia, highlighting strong industry confidence in Point2's breakthrough technology.

Corma conducted hundreds of simulations based on Fortune 500 companies, equipped with multiple security tools, to test top AI models like GPT and Claude. In these exercises, the AI models first acted as attackers, planting persistent threats within the simulated organizations. Then, these same models switched roles to defend and detect the threats they had introduced. Despite their capabilities, the attacking models succeeded in 88% of the cases, highlighting significant gaps in current cybersecurity defenses.

China’s two leading optical module manufacturers faced significant stock declines on Wednesday. Zhongji Innolight's shares dropped up to 16%, with Bloomberg reporting a 14% intraday fall, while Eoptolink Technology also saw decreases. This was triggered by a Reuters report indicating that the US FCC plans to block imports of new models of these optical components, a move initially affecting companies like Amazon and Microsoft, who rely on this technology for their AI data centers.

Anthropic has unveiled inference hooks for Claude Enterprise, a new security feature that directs all prompts and tool calls through a client's own security server before Claude processes them. This system uses a Data Loss Prevention (DLP) server to analyze content and issue an allow or deny decision, ensuring Claude only proceeds with approved data. This verification also applies to all tool interactions, adding an extra layer of security for enterprises.

Yann LeCun, the former chief AI scientist at Meta, is setting up a new venture capital firm called 224 Ventures, which has secured over $100 million in capital. Joining him as a partner is Oriol Vinyals, previously a co-technical lead at Google DeepMind and Gemini. This marks LeCun’s second venture into the VC world, signaling a strong commitment to fostering innovation in AI and related fields. Shaun Johnson is also involved in this promising new fund.

Anthropic revealed on Wednesday its initiative to build an internal silicon team focused on designing custom chips for Claude. This marks the company's first public confirmation of the project. A spokesperson explained to Business Insider that the company aims to co-develop both hardware and models to enhance Claude's performance, enabling faster and more efficient operation to meet customer demands at scale.

Anthropic has appointed Mariano-Florentino Cuellar as its first chief global affairs officer, a strategic hire during a period of strained relations with the US government. Reporting directly to Anthropic president Daniela Amodei, Cuellar will operate out of the company's San Francisco headquarters. His role will focus on navigating the complex geopolitical landscape affecting the AI industry.

Deel, the global HR and payroll platform, has acquired Israeli AI cybersecurity firm Clarity, specializing in real-time deepfake detection and identity verification. The acquisition, valued between $45 million and $50 million in cash and shares, marks Deel’s 15th purchase and its entry into the Israeli market. This strategic move addresses the rising challenge of fake hires by enhancing trust and security in remote hiring processes.

Microsoft executive VP Jay Parikh has instructed employees via an internal email to limit their AI token consumption, stating that “tokenmaxxing is not our goal.” Reported by 404 Media, the email also announces that Microsoft has shifted its default internal AI model to a more cost-effective version from OpenAI. This move is part of a broader effort to manage AI resource usage and establish budgets at the division level within the company.

NHS England has acknowledged an error in its data-protection paperwork that failed to fully disclose access rights to patient information. Contrary to earlier statements, staff at Palantir, the US-based data analytics company, are able to view identifiable patient data within the Federated Data Platform system. This clarification was provided following a request from the National Data Guardian, highlighting concerns about transparency in handling sensitive medical records.