Early specs for AMD’s next-gen mobile platform, codenamed Medusa Point, have surfaced. The Zen 6-based APUs will come in 28W and 45W configurations, aiming for a broad range of notebooks with a focus on efficiency and AI.
Samsung has developed new GDDR7 memory modules running at 40 Gbps per pin. With 3GB per chip, this tech could enable GPUs with massive VRAM without complex board designs.
A new analyst note suggests Apple might use Intel’s 18A-P process for non-Pro iPhones in 2028. This follows Ming-Chi Kuo’s report of an NDA and sample procurement. But shifting core silicon away from TSMC is a monumental risk.
The tech world saw two massive money moves. Meta is reportedly reallocating its massive Reality Labs budget from long-term metaverse bets to AI-powered smart glasses. Meanwhile, in China, AI chipmaker Moore Threads just raised $1.13 billion in a wildly popular IPO.
Google’s stock is up 66% in 2025, fueled by its AI ambitions. A reported multi-billion dollar deal to supply Meta with its custom Tensor Processing Units (TPUs) could signal a major shift in the high-stakes AI hardware race, directly challenging Nvidia.
Google’s in-house AI chips, called TPUs, are a major reason its stock surged 31% last quarter. Investors are betting the company could start selling them, creating a potential $900 billion revenue stream. The internal strength is clear, but the external market is a whole different game.
AMD is reportedly implementing price hikes across its CPU and GPU product lines. The increases, which could be around 10%, are said to affect the new Ryzen 9000 series and older chips, and take effect at midnight tonight.
A new report claims NVIDIA has secured exclusive early access to TSMC’s next-generation A16 manufacturing process. This 1.6nm-class technology is slated for NVIDIA’s future Feynman GPUs, expected after the Rubin family in 2026/2027.
Intel’s advanced packaging technology is gaining significant attention from major AI players. Google is reportedly considering EMIB for its TPU v9 expected in 2027, while Meta may use it for MTIA accelerators. This comes as TSMC’s CoWoS capacity remains heavily allocated to NVIDIA and AMD.
Meta is reportedly planning to use Google’s tensor processing units in its data centers starting in 2027. The move could signal growing competition for Nvidia in the AI chip market and sent Nvidia shares down 3% on the news.
