AMD

Microsoft has officially rolled out its AI upscaling technology Auto SR to the Asus ROG Xbox Ally X handheld and select AMD Ryzen Copilot+ PCs, enabling handheld battery-powered use and 1080p game support for the first time. A new Performance Mode raises the frame-rate cap from roughly 60-75 FPS to 120 FPS and can now run alongside Auto HDR. The rollout is staged and requires updated Xbox Game Bar, Windows 11, and the Auto Super Resolution Package from the Microsoft Store.

AMD and Spectro Cloud launched Instinct Coder 355, an enterprise AI coding platform built on eight Instinct MI355X accelerators plus dual 64-core EPYC 9575F CPUs, running GLM-5.3 locally at 98% of frontier-model performance. Each node serves up to 85 developers (50 concurrent) and defaults to on-prem compute instead of cloud APIs, targeting private enterprise coding workloads.

AMD launched the EPYC Embedded 9006 series on the new SP7 platform, built on Zen 6 with up to 256 cores and 512 threads, debuting PCIe Gen6 and CXL 3.1 in embedded markets. It delivers up to 2.7x memory bandwidth and 1.8x multi-core throughput over the prior generation, targeting ATE testers, enterprise storage and high-end firewalls, with the 128-core SKU shipping November 15.

AMD's Zen 6 core family extends Zen 5 with wider out-of-order structures and a reworked integer scheduler layout that echoes Zen 3 and Zen 4. The direction predictor, already the strongest measured in Zen 5, is carried forward and refined. Zen 6 and Zen 6c are no longer the same core at two densities but two distinct cores tuned for different operating points, signaling deeper microarchitectural specialization.

Google unveiled two new eighth-generation TPU variants, TPU 8tx and TPU 8i Max, targeting confidential inference and training plus large-memory workloads, as it moves to cut reliance on third-party silicon. TPU 8tx, co-designed with AMD, pushes security into the hardware layer to isolate sensitive data and model weights even during active use.

Anthropic's latest draft prospectus shows the company aims to raise up to $64.4 billion in its IPO, plus $26 billion via a private placement to cornerstone investors including Google, Nvidia and AMD, bringing total fundraising to as much as $90.4 billion. With an initial price range of $70-78 per share, the listing roadshow is imminent, marking one of the largest capital events in AI history.

Google expanded its eighth-generation TPU lineup with TPU 8i Max, which packages two TPU 8i chips, and TPU 8tx, which pairs TPU 8t silicon with 5th Gen AMD EPYC hosts to protect data and weights in use. Google claims 8i Max doubles compute, SRAM and HBM bandwidth to roughly 17.2 TB/s per package versus 22 TB/s on NVIDIA Rubin, while 8tx claims up to 2x lower latency; no absolute specs, prices or benchmarks were published.

AMD CEO Lisa Su cited Meta's Muse as a standout example of rapid AI adoption, noting over 10 million activations in 45 days. She said annual operating plans that were once fixed now shift month to month because of AI adoption.