Welcome to the BrevFeed daily tech briefing for Tuesday, September 29. We've got 12 stories for you today across A.I., security, software, cloud, hardware and startups. Let's get into it.

In hardware, Nvidia has reported a record 96.2 billion dollars in revenue for the past quarter, with profits more than doubling to 59.7 billion dollars. The company projects an even higher 108 billion dollars in revenue for the next quarter, a milestone previously achieved by only a few tech giants.

This significant growth is primarily fueled by Nvidia's data center business, which generated 89 billion dollars last quarter, marking a 117% increase year-over-year. This performance highlights Nvidia's critical role in the AI industry, as major tech companies rely on its chips for their AI infrastructure.

While the data center segment is soaring, Nvidia's consumer gaming business, part of its edge computing category, also saw a 27% increase year-over-year, contributing 7.2 billion dollars. However, this segment faced challenges from slower PC sales and higher component prices.

In a major move for artificial intelligence, AMD has announced its acquisition of World Labs, an AI firm specializing in deep learning models for understanding physical reality. The deal is valued at $8.2 billion in an all-stock transaction.

As part of the acquisition, World Labs founder Dr. Fei-Fei Li will join AMD as Executive Vice President and Chief Scientist. She will work directly with AMD CEO Dr. Lisa Su to integrate AI development closer to hardware and expand AMD's AI ecosystem.

World Labs, which launched in 2024 and was valued at a billion dollars within months, develops models that can create interactive 3D worlds from prompts, like their product Marble. This acquisition aims to align AMD's chip development with future AI model requirements, particularly for applications in robotics and autonomous systems.

In intellectual property news, a federal jury in San Diego has ordered Apple to pay Taction Technology over 5.7 billion dollars in damages. The verdict, delivered on Friday, found that Apple infringed on two of Taction's haptics patents related to vibration technology.

The lawsuit, filed in 2021, alleged that Apple's Taptic Engine, used in iPhones and Apple Watches, incorporated Taction's patented vibration technology without a license. The Taptic Engine debuted in the Apple Watch in 2014 and came to the iPhone 6s and 6s Plus in 2015, replacing older vibration motors.

Apple had initially seen the lawsuit dismissed in 2023, but the US Court of Appeals for the Federal Circuit reinstated the case, leading to this jury trial. Apple denies using Taction's technology and plans to appeal the verdict, though the jury did not find the infringement to be willful.

In AI, Anthropic is accusing several Chinese AI labs of illegally distilling its Claude models to train competing AI technology. Jacob Klein, Anthropic's head of threat intelligence, stated that this practice is not legal competition but rather theft of intellectual property. The alleged distillation allows these labs to create cheaper, copycat versions of Anthropic's advanced models.

Anthropic specifically named Moonshot AI, DeepSeek, MiniMax, and Alibaba as engaging in these activities. For example, Moonshot AI's Kimi K3 model is alleged to have been illegally trained using Claude, and Alibaba is accused of a massive distillation attack to capture capabilities from Anthropic's frontier AI models. Anthropic observed nearly 200 million exchanges linked to these distillation attacks, with Alibaba's campaign alone involving over 151 million exchanges between May and July.

This issue comes as Anthropic approaches a potential IPO, and the U.S. government has expressed concerns, with an April memo calling such distillation "unacceptable." While some, like Y Combinator CEO Garry Tan, advocate for an open market, others, including Treasury Secretary Scott Bessent, describe distillation as "theft" and have threatened sanctions against overseas companies involved. However, Aidan Gomez, CEO of Cohere, disputes that all of China's AI progress is due to distillation, stating that recent Chinese models show independent, world-class capabilities.

In space news, Skyroot Aerospace's Vikram-1 rocket has successfully reached orbit on its inaugural launch from Sriharikota. This marks India's entry into the private satellite launcher market, with the Vikram-1 becoming the country's first privately-developed orbital rocket.

The rocket, which stands about 72 feet tall, entered a 280-mile-high orbit after its three solid-fueled motors fired in succession. It can carry 350 kilograms into low-Earth orbit, positioning India to compete in the commercial space sector.

The launch, which was delayed by over half an hour due to a technical problem, ultimately proceeded successfully. This achievement is a significant milestone for India's space capabilities and its growing private space industry.

In security news, Kiteworks, a company specializing in large file transfers, recently advised its customers to shut down their systems due to an imminent cyberattack threat. This recommendation came after Kiteworks received credible intelligence from law enforcement about a potential attack.

The company's Chief Information Security Officer, Frank Balonis, clarified that the advisory was a precautionary measure, not a response to a confirmed breach. The concern was that unknown zero-day vulnerabilities could be exploited before patches were available. Kiteworks recommended a shutdown window, with specific times for different regions, and stated that all known vulnerabilities are addressed in their latest software release, version 9.5.1.

While the vulnerability primarily affected the Advanced Forms product and fewer than 1% of customers, Kiteworks operates a Private Content Network with over 100 million end-users. The shutdown recommendation was lifted on Sunday, and by Monday, all hosted customer systems were back online. Kiteworks reported finding no evidence of compromise or suspicious activity.

In security news, cryptocurrency exchange Bitget has reported a major breach, with hackers stealing 351.6 million dollars from its hot and warm wallets. The attack, which occurred on September 24th, is believed to be the largest crypto heist of the year. Bitget has temporarily suspended withdrawals while it investigates the incident, but assures users that its 464 million dollar User Protection Fund will cover all losses.

The company stated that its cold wallets and most platform assets remain secure, and customer account balances are accurate. Deposits and trading are also operating normally. Bitget has enlisted cybersecurity firms Mandiant and SlowMist to assist with the investigation, and has reported the incident to relevant authorities.

Bitget CEO Gracy Chen noted that the attack methods, which exploited a backend wallet system by forging transfer approvals and using suspicious IP addresses, are highly consistent with patterns seen from North Korean hacker organizations. The breach impacted multiple blockchain networks, including Ethereum, XRP Ledger, and Avalanche, with XRP accounting for the largest single-chain loss.

In a significant move for enterprise AI, Meta Platforms has hired CJ Desai, the former CEO of MongoDB, to lead its new Meta Enterprise Platform. Desai will serve as Chief Enterprise Platform Officer, reporting directly to Mark Zuckerberg, and will focus on developing AI tools and services for businesses.

The new platform aims to commercialize Meta's AI stack, offering products like the Muse agent, Meta Business Agent, Muse API, and Muse Code. This initiative is part of Meta's strategy to leverage its advanced AI models and infrastructure to compete in the enterprise AI market.

The announcement had an immediate impact on the stock market. MongoDB shares dropped 20% following Desai's departure, with Dev Ittycheria named interim CEO. Meta's stock also saw a decline of about 4% after the news.

In developer tools, Worker Previews has officially launched, offering isolated preview environments for every change made within a Git branch. This new feature aims to provide a testing environment that closely mirrors production, allowing developers to validate changes before they are deployed live.

Each Git branch now receives a dedicated space to run, complete with its own code, configuration, URL, observability tools, and state. This setup enables developers to deploy isolated previews using a simple command, ensuring variables, secrets, and bindings are separate from production.

A stable preview URL is provided for each branch, which updates with every push, facilitating continuous testing of UI and runtime responses. Developers can also inspect logs, errors, metrics, and traces within each preview to confirm functionality and identify issues, creating a pre-production feedback loop for every branch.

In cloud development, Cloudflare has announced the general availability of Python Workers, making Python a fully supported language on its Developer Platform. This means developers can now run Python applications directly within the Cloudflare Workers runtime, two years after the initial preview.

This update allows for seamless integration with various Cloudflare services, including Workers AI, R2, and D1. It also supports popular Python web frameworks like FastAPI, Django, and Flask, enabling developers to deploy their existing Python code and libraries within this environment.

The development of Python Workers involved leveraging WebAssembly and Pyodide, allowing a WebAssembly-compiled Python interpreter to run efficiently within the Workers environment. This multi-year effort aimed to provide a scalable platform for Python applications with strong performance.

In AI news, Florida Attorney General James Uthmeier has requested a court order against OpenAI, aiming to prevent ChatGPT from exhibiting human-like attributes. Uthmeier argues that this practice deceives users and boosts engagement, which in turn aids data collection.

The motion also seeks to block OpenAI from developing new AI models without independent, third-party approved safety guardrails, and to restrict minors' access to ChatGPT. This action is part of an ongoing lawsuit against OpenAI, which began in June.

The lawsuit itself stems from a criminal investigation initiated after a 2025 mass shooting, where the suspect allegedly used ChatGPT. Uthmeier stated that there should be no new model development without independent safety guardrails, and no more harvesting children's data.

In autonomous vehicles, Aurora is significantly expanding its driverless semi-truck fleet, aiming for 200 trucks by the end of the year. These vehicles currently operate on key Texas routes, using lasers, radar, and cameras for navigation.

The company is targeting profitability with a "driver as a service" model, where customers pay per mile. Analysts project this could lead to substantial cost reductions, with Aurora's service costing around 85 cents per mile compared to $1.30 for human-driven trucks.

This model addresses the ongoing driver shortage and allows trucks to operate for up to 20 hours a day, unlike human drivers. While the cost savings are clear, some industry experts note that human drivers offer more than just transportation, including judgment and customer service, which autonomous systems don't yet replicate.

That's the BrevFeed daily briefing. We'll be back tomorrow with the stories that matter in tech. Thanks for listening.