Artificial Intelligence News Highlights, 29 July 2025: AI health startup Ambience Healthcare valued at over $1 billion after new funding

Abu Dhabi's state-backed artificial intelligence firm, G42, is nearing an agreement to secure data center capacity from German high-performance computing provider Northern Data AG, according to a report in Bloomberg. The potential deal is intended to provide the necessary AI compute infrastructure to support G42's expansion of its operations into Europe. The move is part of the company's broader strategy to build out its global infrastructure for developing and deploying artificial intelligence.

Whitefiber Inc., an artificial intelligence infrastructure subsidiary of crypto treasury firm Bit Digital Inc., is seeking to raise as much as $132.8 million in an initial public offering, according to a report in Bloomberg.

Ambience Healthcare Inc., an artificial intelligence startup backed by OpenAI that focuses on streamlining administrative work for medical professionals, has secured $243 million in a new funding round, as per a report in Bloomberg. This latest investment brings the company's total valuation to more than $1 billion. The startup's technology is designed to reduce the clerical workload for doctors and other healthcare staff.

Apple Inc. has lost its fourth artificial intelligence researcher in the last month to a specialized team at Meta Platforms Inc., according to a report in Bloomberg. The departure marks a setback for Apple's AI development efforts as the researcher is joining Meta's 'Superintelligence Team,' which focuses on long-term goals like artificial general intelligence. This recent talent migration highlights the intense competition for top AI experts among major technology companies.

A proposed Artificial Intelligence data center in Wyoming could eventually consume five times more electricity than the entire state currently uses, according to a report in Ars Technica. The projection, based on developer filings for the facility, illustrates the massive and growing energy demands of AI infrastructure. This single project's potential power draw at full deployment would be equivalent to five times the current combined residential, commercial, and industrial electricity consumption in Wyoming, highlighting the significant strain such developments place on regional energy grids.

AI chip company Groq is reportedly nearing a new fundraising round that would value the company at $6 billion, TechCrunch says in a report. If the deal closes, it would double Groq's valuation in about nine months. The company is known for developing specialized Language Processing Units (LPUs) optimized for high-speed, low-latency AI inference, which is the process of running trained AI models. This focus positions Groq as a challenger to Nvidia, particularly within the growing market for AI inference hardware.

Google has introduced a 'Video Overviews' feature for its AI research and notetaking assistant, NotebookLM, according to a report in TechCrunch. This new capability allows the tool to automatically generate short, visual summaries from the user's source materials, such as uploaded documents, notes, and web articles. The feature is designed to help users synthesize and understand complex information more effectively by presenting key points in a dynamic video format, leveraging multimodal AI to transform text into a visual and auditory experience.

OpenAI has launched a new 'study mode' for its ChatGPT platform, which aims to foster deeper understanding by guiding users through problems with interactive prompts instead of just providing answers, as per a report in The Verge. The feature, developed with input from educators and pedagogy experts, uses Socratic questioning, hints, and self-reflection prompts to create a tailored learning experience. This development places OpenAI in direct competition with rivals like Anthropic, which previously released a similar feature, and Google, which also announced new AI tools for students. A key aspect of the new mode is that it is optional and can be turned off by the user, with OpenAI's VP of education, Leah Belsky, confirming there are currently no administrative tools to enforce its use.

Meta is intensifying its AI talent acquisition by targeting staff from competitor OpenAI for its new superintelligence lab, having reportedly approached more than a dozen employees, according to a report in Wired. This move is part of a broader strategy by Meta to advance its capabilities in the race for artificial general intelligence and reflects the escalating competition for top-tier AI researchers and engineers.

YouTube will begin using AI and machine learning to estimate the age of its users in the US, automatically applying safety restrictions to accounts it identifies as belonging to minors, The Verge says in a report. The platform will start a limited rollout of the technology, which analyzes user activity and account data, over the next few weeks. Accounts flagged as underage will have age-restricted videos blocked, receive non-personalized ads, and see "take a break" notifications and limits on sensitive video recommendations. YouTube will notify users if their account is re-categorized, and they can appeal the decision by providing a government ID, credit card, or a selfie for verification. The move, which may affect creator ad revenue, aligns with a broader global push for online child safety.

Iconiq Capital is poised to lead a new $5 billion funding round for artificial intelligence start-up Anthropic, a deal that would raise the company's valuation to $170 billion, according to a report in the Financial Times. This significant capital infusion for the OpenAI competitor underscores continued strong investor confidence in the foundational AI model market. The new valuation represents a substantial increase for Anthropic, which is known for developing the Claude series of AI models.

Spotify has signaled its intention to develop a more conversational and interactive voice AI interface for its application, according to a report in TechCrunch. The company stated that artificial intelligence will enable "much more interactive" consumer experiences within its platform. This move suggests a future direction beyond simple voice commands, aiming for a more natural, chat-like interaction for music and content discovery, building on its existing use of AI, such as the AI DJ feature. The development is presented as a concept or strategic direction rather than an imminent product launch.

Google is rolling out a new suite of artificial intelligence features to enhance its services, according to a report in TechCrunch. The updates include a new 'Canvas' feature, described as a tool for AI-powered visual content creation. Additionally, a 'Search Live' function will provide real-time, interactive assistance within Google Search. The company is also integrating an advanced visual search capability that allows users to search for items they see on a webpage directly while browsing.

Data from a Stack Overflow survey indicates that developers are experiencing a "hidden productivity tax" due to the time required to debug and correct 'almost right' code generated by AI tools, according to a report in VentureBeat. This "tax" refers to the unanticipated effort needed to fix code that appears correct but contains subtle bugs, inefficiencies, or contextual errors. The findings suggest that as enterprise developers increase their use of AI coding assistants, the reality of their performance often falls short of initial expectations for time savings, as the need for extensive human verification and refinement negates some of the anticipated productivity gains.

Meta will now permit job candidates to use artificial intelligence tools during its coding tests, according to a report in Wired. This policy change shifts the focus of engineering assessments from rote coding ability to a candidate's proficiency in using AI for efficient problem-solving and improving code quality. The move aligns with the increasing integration of AI assistants in professional software development and reflects Meta's strategic emphasis on AI, including a concept described by Mark Zuckerberg as 'vibecoding,' which prioritizes a developer's focus on higher-level creative tasks.

Yelp has expanded a feature nationwide on its iOS app that uses artificial intelligence to automatically generate short, vertical videos about restaurants and other businesses, according to a report in The Verge. The system combines user-submitted photos and videos with AI-generated scripts from OpenAI's large language models, voice narration from ElevenLabs, and synchronized captions from Amazon Transcribe. According to Yelp CPO Craig Saldanha, videos are created for businesses with sufficient content and are shown to users based on personalized signals, though the video content itself is not personalized. Currently, users cannot opt out of having their content used, and businesses cannot preview the videos created about them, but both can report videos they find inaccurate or offensive.

The Linux Foundation is launching a new open-source project to enable interoperability between different artificial intelligence agents, according to a report in ZDNet. The initiative is based on the AGNTCY framework, a foundational technology donated by Cisco, which is designed to act as a common language for disparate AI agents. The project's goal is to create a standardized way for agents to discover, communicate, and collaborate, forming what is being called an "Internet of Agents." This addresses the significant challenge of integrating AI agents built on various large language models and platforms, aiming to foster a more connected and modular AI ecosystem.

Adobe is introducing new AI-powered image editing capabilities to Photoshop, according to a report in TechCrunch. A key feature of the update is an improved background removal tool that utilizes artificial intelligence. The new tool is designed to produce cleaner results by preventing the creation of visual artifacts that can sometimes remain after an object is isolated from its background.

Celestica Inc. has increased its full-year revenue forecast for the third consecutive time, citing growing demand for its cloud-computing hardware driven by the artificial intelligence sector, according to a report in Bloomberg. The company, an Electronic Manufacturing Services (EMS) provider, designs and manufactures server racks and other critical infrastructure for large-scale cloud providers to run intensive AI workloads. This marks the third consecutive upward revision to the company's financial guidance, reflecting sustained demand that has consistently outpaced earlier projections.

Microsoft Corp. is in advanced talks for a new contract that would guarantee its ongoing access to critical technology from Artificial Intelligence lab OpenAI, according to a report in Bloomberg. The potential agreement is seen as a crucial step for OpenAI's unique 'capped-profit' subsidiary, as it would provide financial stability and help its efforts to operate as a for-profit enterprise. For Microsoft, a deal would solidify its ability to integrate OpenAI's foundational models across its product ecosystem, including its Azure cloud services.

Nvidia-backed AI startup Cohere has reached an agreement to integrate its artificial intelligence models into services provided by BCE Inc., Canada's largest telecommunications firm by revenue, according to a report in Bloomberg. The partnership will see Cohere's technology used to enhance BCE's operations and customer-facing services. The deal marks a significant enterprise adoption for Cohere, which is an investment of chipmaker Nvidia and competes in the market for large language models.

Turkish power-barge company Karpowership is expanding into building floating data centers to meet the soaring demand from artificial intelligence, according to a report in Bloomberg. The company, which specializes in constructing floating power plants, is betting that waterborne data hubs can be deployed more quickly than traditional onshore facilities. This strategic shift addresses the challenges faced by land-based data centers, which are struggling to find adequate space and power to support the massive computational needs of modern AI applications.

A new poll shows that 60% of US adults use artificial intelligence, with information searches being the most common application, according to a report in AP. The data reveals a significant age gap, as 74% of adults under 30 use AI, and they are more likely to utilize it for brainstorming and work-related tasks. Fewer adults overall use AI for functions like work, drafting emails, or shopping, and while most do not use it for companionship, this use case is more common among the younger demographic.

An AI agent from OpenAI successfully bypassed an "I am not a robot" CAPTCHA verification test, a task designed to block automated bots, Ars Technica said in a report. During the demonstration, the agent reportedly articulated its understanding of the task, stating, "This step is necessary to prove I'm not a bot." This event highlights the increasing sophistication of autonomous AI agents in navigating digital environments and their ability to mimic human-like browsing patterns, which can allow them to pass advanced background checks used in bot detection systems.

Artificial intelligence data analyst startup Julius AI has raised $10 million in a seed funding round led by venture capital firm Bessemer Venture Partners, according to a report in TechCrunch. The company develops AI-powered tools designed to automate complex data analysis and business intelligence tasks.