Reports
AI-generated structured vendor updates
Samsung GAIA AI PC Chip Samples with Memory-Centric NPU, Targeting 50 TOPS
Samsung launches GAIA AI PC processor with 4nm process and memory-centric NPU, integrating LPDDR5X controller with NPU for near-memory computing, achieving 40% energy efficiency improvement and 50 TOPS. Certified for Microsoft Copilot+ PC, Lenovo to adopt in Q4 2026.
AMD's Experimental Topological Ghost Protocol Boosts MI300X Inference 10x
AMD introduces experimental Topological Ghost Protocol (TGP) on MI300X GPUs, achieving 431 tokens/sec with 100% success in high-concurrency inference, 10x improvement over standard vLLM. TGP uses KV-cache recycling and segmented state management, still experimental but potentially redefining AI inference benchmarks.
SambaNova完成11亿美元融资估值110亿美元：推理芯片新格局确立
...
MediaTek and Alibaba Cloud Deploy Tongyi Qianwen LLM on Dimensity Chips
MediaTek partners with Alibaba Cloud to deploy a small version of the Tongyi Qianwen LLM on Dimensity 9300/8300 mobile platforms, enabling offline multi-turn conversations. This move aims to capture edge AI inference control via NPU optimization and SDK integration, directly challenging Qualcomm.
Huawei Unveils Tao's Law V2: Kirin 2026 Boosts AI Inference 40% on Same Node
Huawei's He Tingbo releases Tao's Law V2, detailing Kirin 2026 metrics: 238 MTr/mm² transistor density (+55%), 41% power reduction at iso-performance, and 40% SRAM frequency increase. Without EUV lithography, co-optimization of architecture, circuit, and process delivers equivalent performance gains, proving system-level optimization as a viable alternative to Moore's Law scaling.
Google Cloud Launches Blackwell GPU Confidential VM & Open-Source Prompt Encryption SDK, Redefining AI Security
Google Cloud upgrades its confidential computing portfolio with Blackwell GPU-based confidential VMs (Confidential G4 VMs preview), open-source Prompt Encryption SDK, and enhanced Confidential Space featuring Intel Trust Authority and Hopper GPU support, addressing TEE vulnerability CVE-2026-33697 to bolster AI inference and cross-organization training security.
OpenAI Slashes Inference Costs 50%, Runs ChatGPT on Hundreds of GPUs via System-Level Optimization
OpenAI reduces AI inference costs by over 50% through system-level optimizations: model quantization (FP16 to INT4/INT8), KV-Cache optimization, dynamic batching, and speculative decoding. Using only hundreds of NVIDIA GPUs to serve ChatGPT's unlogged-in traffic, inference gross margin jumps from 38% to 65%, nearing breakeven.
Qualcomm Enters AI Inference with Dragonfly C1000 CPU and HBC Near-Memory Compute
Qualcomm unveils Dragonfly roadmap with Oryon-based C1000 CPU and AI300 inference accelerator featuring HBC near-memory compute. Meta and Microsoft are early adopters. The strategy targets AI inference TCO reduction and memory wall breakthrough, bypassing Nvidia's training dominance.
Etched Unveils Sohu Transformer ASIC: Claims 20x H100 Inference Throughput, Challenging NVIDIA's Grip
AI chip startup Etched emerges from stealth with Sohu, a Transformer-specific ASIC on TSMC N4P with 144GB HBM3E. By hardwiring attention mechanisms, it claims 20x throughput and 140x price-performance vs. H100 on Llama 70B. With $800M total funding and first racks shipping this summer, it directly challenges NVIDIA's inference dominance.
OpenAI and Broadcom launch Jalapeño inference ASIC: 9-month tapeout, 2027 mass production, targets GPU replacement
OpenAI and Broadcom unveil Jalapeño, a custom inference ASIC designed in 9 months using OpenAI's own LLMs. Early benchmarks show superior performance-per-watt vs. current GPUs. Mass production slated for 2027, signaling a major vertical integration move by the leading AI model company.
Qualcomm Acquires Modular for $3.9B, Open-Sources Mojo to Break CUDA Lock-In
Qualcomm acquires Modular for $3.9B in stock and open-sources Mojo, a Python-compatible systems language. Mojo targets CUDA dependency, aiming to provide a high-performance alternative for AI developers. This move strengthens Qualcomm's AI inference chip software stack and edge AI competitiveness.
OpenAI and Broadcom unveil Jalapeño inference ASIC to bypass NVIDIA GPU dependency
OpenAI and Broadcom launch Jalapeño, a custom ASIC for LLM inference, achieving tape-out in 9 months. OpenAI designs architecture, Broadcom provides networking, Celestica handles integration. Planned for large-scale deployment by end-2026 with gigawatt-scale datacenters, aiming to cut inference costs and reduce NVIDIA dependency.
Huawei and Hubei Mobile Validate AI Inference Acceleration: External KV Cache Boosts Throughput 372%
Huawei and Hubei Mobile completed the first operator AI inference acceleration trial, using OceanStor A800 storage and Ascend A3 supernode with UCM to externalize KV Cache to PB-level storage, achieving up to 372% TPS improvement for long-context inference on GLM-5.1 and MiniMax M2.5 models.
Intel at Computex 2026: CPU as Agentic AI Orchestrator, x86 Reclaims Inference Control
At Computex 2026, Intel unveiled the 288-core Xeon 6+ (Intel 18A) and 3rd-gen Core Ultra, claiming Agentic AI shifts CPU:GPU ratio from 1:8 to 1:1. Partnering with SambaNova and Foxconn for rack-scale inference systems, Intel repositions the CPU as the orchestrator for multi-step AI reasoning, aiming to reclaim control from GPU-centric architectures.
Cloudflare AI Gateway 2.0: Edge Control Plane Captures AI Inference Routing and Security
Cloudflare launches AI Gateway 2.0 with smart routing across 50+ model providers claiming 30% cost reduction, Workers AI edge inference (<10ms latency), NVIDIA GPU acceleration partnership, and expanded AI firewall. This shifts the AI traffic control plane from centralized clouds to the edge network.
Arm's Self-Designed AGI CPU with Meta: Ecosystem Shift from Licensor to Silicon Vendor
Arm unveils its first self-designed data center CPU, the AGI CPU, with 136 cores on 3nm, purpose-built for agentic AI inference. Co-developed with Meta, which will deploy it across its data centers. Claims 2x rack performance over x86, reducing AI capex by $100B per gigawatt. Signals Arm's shift from IP licensing to direct silicon sales, reshaping ecosystem dynamics.
Arm AGI CPU Demand Doubles, Targets AI Inference Control, Threatens x86 Dominance
Arm doubled its demand forecast for its first in-house datacenter CPU, the AGI CPU, projecting over $2B revenue in FY2027-2028. The 136-core, 3nm Neoverse V3-based chip targets agentic AI inference, claiming 2x rack-level performance over x86. Meta is a key partner; OpenAI, Cloudflare also onboard. This marks Arm's strategic pivot from IP licensor to direct silicon vendor.
Qualcomm AI200 on AWS: Inference Chip Ecosystem Shifts from Nvidia Singularity to Multi-Alliance
Qualcomm's AI200 inference chip (768GB memory) is slated for broad AWS deployment by 2026, aiming to reduce cloud AI inference costs. This marks Qualcomm's strategic pivot from mobile to cloud, leveraging AWS's custom silicon initiative to challenge Nvidia's inference monopoly and restructure the cloud inference chip ecosystem.
AMD Zen 6 Venice 256-Core EPYC Claims 3.3x Rack Performance Over NVIDIA Vera, But Estimates Raise Questions
AMD unveils first estimated performance of Zen 6 Venice EPYC (2nm, 256 cores), claiming 3.3x rack-level integer throughput over NVIDIA Vera at 100kW total power. A direct counter to NVIDIA's Arm push, but based on projected estimates, not silicon.
Microsoft Maia 200 Mass-Produced, Cobalt 200 Previewed: AI Inference Control Shifts to Azure
At Build 2026, Microsoft announced mass production of Maia 200 AI inference chips, preview of Cobalt 200 ARM processors, and the MAI-Thinking-1 reasoning model (35B params). This signals a full-stack vertical integration to reduce NVIDIA dependency and lock Azure AI workloads.