Broadcom CEO Warns: Capital Is Replacing Chips and Power as the Biggest Bottleneck for AI Development

Beating AI News (from Dongcha) – AMD today announced the launch of the AMD Helion Agent AI Rack Solution. Built on an open infrastructure design compliant with Open Compute Project (OCP) specifications, the solution is engineered to meet the computing power demands of large-scale AI agent tasks, supporting flexible combinations of CPUs and GPUs. It offers two primary configurations: The first is a CPU-only rack powered by 6th Gen AMD EPYC processors, supporting up to 32,768 CPU cores, designed for hyperscale deployments, providing dedicated computing power for agent tasks and connecting to separate model inference infrastructure. The second is an integrated rack combining AMD EPYC CPUs, AMD Instinct MI350P GPUs, and general-purpose compute nodes, targeted at enterprise-grade agent AI applications. AMD states that Helion supports both liquid and air cooling configurations, allowing customers to adjust CPU and GPU resources based on their workload, performance, power consumption, and cost needs. The solution also integrates open-source software and hardware plus a built-in token router, enabling the distribution of agent model calls to both frontier and open-source models. AMD plans to submit the Helion specifications to the OCP for further community development and adoption. Currently, the solution is being developed in collaboration with ecosystem partners, with a launch scheduled for November 2026.

Beating AI News: Nicolas Bustamante, head of Microsoft’s proactive AI Agent team, predicts AI streams will become the mainstream interaction method by 2027. Chat-based models leave too much work to users—requiring them to figure out AI’s capabilities, write prompts, provide context, and repeatedly direct agents. For billions to use agents daily, not everyone can be expected to master these tasks. Bustamante’s envisioned AI stream works like an auto-updated to-do list: AI drafts emails needing responses; when sales data refreshes, team reports are prepped. Users browse it like social media, approving, editing, or ignoring items, only turning to chat for complex issues. Meta’s Muse already proactively suggests and executes tasks. The model also has commercial logic: as users open the stream daily, AI recommends restaurants, goods or services based on their plans, with clearly labeled ads covering the agent’s operational costs.

According to monitoring by AI Aunt, a crypto whale address swapped its UBTC holdings for 16,704.47 ETH over the past 22 hours, worth approximately $42.46 million, at an average purchase price of around $2,541.89.

According to BIT (Bit.com) market data, Microsoft’s stock price hit a near one-year high, rising 0.8% intraday. On the news front, Microsoft plans to gradually increase the use of its self-developed AI chips in data centers to reduce reliance on NVIDIA GPUs. Microsoft stated that it will mainly rely on its self-developed Maia chip architecture to handle AI model requests in the future, and plans to deploy several gigawatts of computing power on the Maia series chips. Microsoft unveiled its third-generation AI accelerator chip, the Maia 300, on Monday, which is expected to enter production in 2027. Each data center rack can hold 96 Maia 300 chips, double the number of the previous-generation Maia 200; compared with GPUs, its AI token processing capacity per rack can reach up to 7 times higher, and the number of tokens processed per megawatt increases by approximately 40%. Microsoft said the Maia 300 is expected to boost AI service performance and lower data center operating costs. Evercore ISI analysts forecast that by the first quarter of 2027, Maia chips will account for around 10% of Microsoft’s AI chip spending.

According to HTX market data, Gitcoin (GTC) has surged an additional 57% in the past 24 hours, currently trading at $0.2871, with its market capitalization climbing to $28.9 million. The token has rallied 136% over the past seven days and 272% over the past 30 days. Upbit officially announced today that it has listed Gitcoin (GTC), supporting trading pairs for the South Korean won (KRW) and Tether (USDT).

Goldman Sachs has upgraded its global stock rating from "neutral" to "overweight" for the next three months, while retaining its bullish stance over the coming 12 months. The bank noted that sustained corporate earnings growth, stock valuations pulling back from summer highs, and reduced investor positioning have improved stocks’ risk-reward ratio. FactSet data shows the S&P 500’s 12-month forward price-to-earnings ratio has fallen from 21.1x in early June to 19.1x, compared to roughly 22x at the start of the year. Goldman Sachs projects S&P 500 constituents will post 27% year-over-year earnings growth in the third quarter, marking the third consecutive quarter of earnings gains exceeding 25%, with AI infrastructure investment remaining a key driver. On regional allocation, the bank maintained "overweight" ratings for U.S. and Asian stocks, and an "underweight" rating for European equities. Meanwhile, Goldman Sachs upgraded its 12-month government bond rating to "overweight", recommending a "barbell strategy": allocate to tech stocks on one end, and stocks with bond-like characteristics on the other, while favoring shorter-duration bonds and cash assets. The bank also warned that rising U.S. Treasury yields, higher energy prices, and uncertainty around the U.S. midterm elections could still exacerbate market volatility. Its bullish outlook remains contingent on sustained delivery of corporate earnings and no further deterioration in financial conditions.