NerdNews

Anthropic, the maker of the Claude AI assistant, is considering raising a new $50B round at a valuation of $850B to $900B, according to sources. The company's rapid growth, with annual revenue run rate surpassing $30B, has attracted significant investor interest. A decision on the round is expected at a board meeting in May.

Microsoft's Copilot has 20 million paid enterprise seats, with users engaging as much as they do with email. The company has quadrupled the number of companies paying for over 50,000 seats, with notable clients including Bayer, Johnson & Johnson, and Accenture.

Google Cloud's revenue has surpassed $20 billion, driven by strong demand for its AI solutions and infrastructure. However, the company's growth is capacity-constrained, with a backlog of $462 billion. Google is working to invest in its infrastructure to meet demand, with plans to work through 50% of the backlog over the next 24 months.

Meta plans to spend $145 billion this year, primarily due to its AI investments. The company is betting on AI to drive growth, with CEO Mark Zuckerberg announcing the release of two AI agents. Meta's increased spending is driven by the high cost of memory chips, which has led to a global memory crisis. The company is also laying off 10% of its workforce and offering voluntary buyouts to 7% of its US staff.

Google's revenue grew 20% in the first quarter of 2026, thanks to its cloud computing unit's $20 billion in sales. This growth is largely attributed to Gemini, Google's AI initiative, which has eaten into ChatGPT's market share. OpenAI, the creator of ChatGPT, has missed its revenue and user targets, sparking concerns about the AI industry's health.

Tracy Bannon discusses the risks of unbridled AI autonomy, highlighting the importance of governance and discipline in managing AI agents. She introduces the concept of 'architectural amnesia' and provides a framework for 'minimum viable governance' to mitigate debt and ensure trust in AI systems.

Sauce Labs has launched an AI-driven agent to automate test creation, aiming to close the DevOps 'velocity gap' by enabling teams to describe expected behavior in natural language and generate self-improving tests. The platform targets a growing bottleneck in modern DevOps, where testing struggles to keep pace with accelerated code production.

A security engineer created a fake Wikipedia entry and website to convince several AI chatbots that he was the reigning world champion of a popular German card game, highlighting the vulnerability of large language models to poisoning attacks. The experiment demonstrated how easily false information can be spread through AI systems, and the need for improved data provenance and filtering to prevent such attacks.

Wiz researchers discovered a high-severity flaw in GitHub's git infrastructure using AI-augmented tooling, allowing remote attackers full read/write access to private GitHub repositories. GitHub responded quickly, issuing fixes within six hours.

A report by KELA found 2.9 billion compromised credentials in 2025, including usernames, passwords, and session tokens. The surge in compromised credentials is attributed to infostealers, with at least 347 million obtained by infostealers found on 3.9 million infected machines. The report also notes a growing use of AI to power various stages of attacks, with cybercriminals using AI to enhance and escalate attacks.

Mistral AI has introduced Workflows, an orchestration layer for enterprise AI, to manage multi-step AI processes with durability, observability, and fault tolerance. It tackles common AI deployment issues and supports human-in-the-loop steps, stateful execution, and integration with Temporal.

Meta's superintelligence team is developing AI agents for personal and business use, aiming to make them more approachable and easier to use than existing products. The agents will build on the newly-released Muse Spark model and focus on helping users achieve their goals.

Uber has expanded its services to include hotel bookings, with over 700,000 hotels available through a partnership with Expedia Group. The company also announced a new AI-powered voice bookings feature and a search tool. These new features were developed quickly using agentic AI tools, which have reduced the development time for new features.

BMW i Ventures has launched a new $300 million fund with a focus on AI, agentic AI, physical AI, industrial software, and advanced materials. The firm believes AI will reshape the automotive industry and wants to invest in early-stage startups working on these technologies.

SenseTime, a Chinese AI company, has released a new open-source model called SenseNova U1, which can generate and interpret images faster than top US models. The model is optimized to run on Chinese-made chips, helping the company bypass US export controls. SenseTime hopes this release will help it catch up with domestic and Western AI players.

Amazon and Meta are lobbying India's payments body to address the dominance of Google Pay and PhonePe in the country's instant payments network. They will meet with the National Payments Corporation of India to discuss concerns over user acquisition practices, product design, and monetization.

Amazon Web Services (AWS) saw a 28% year-over-year increase in net sales, reaching $37.6 billion. This growth is attributed to its role in fueling the AI boom. Amazon's capital expenditures are also increasing to support the growth of its cloud business, with a focus on building out infrastructure such as data centers and servers.

ChatGPT's user growth is slowing down, with a 132% increase in uninstalls year-over-year in April. The app's monthly active users grew by 168% in January, but only 78% in April. This slowdown may cause problems for OpenAI's planned IPO, as the company has missed its internal targets for new users and revenue.

Eka's robotic claw can perform tasks like sorting chicken nuggets and screwing in lightbulbs with human-like dexterity, potentially revolutionizing how robots are used in various industries. The company's approach involves training robots to learn from simulations and invent their own solutions, rather than relying on human-provided training data.

Apple's next CEO, John Ternus, needs to launch a killer AI product to stay competitive. The company's current AI efforts are underwhelming, and Ternus must leverage Apple's custom silicon strategy to create a game-changing product. This could involve developing more powerful neural engines, potentially in partnership with companies like Broadcom, to deliver AI capabilities to devices and protect user privacy.