Reports
AI-generated structured vendor updates
TSMC CoWoS Capacity to Reach 200k Wafers by 2027, Diversifying from GPU to CPU and ASIC
TSMC targets 200k wpm CoWoS capacity by 2027, narrowing supply-demand gap from 20% to 10%. Customer base diversifies from NVIDIA GPU to include AI server CPUs (MediaTek, AMD) and ASICs (Broadcom). CoPoS panel-level packaging enters pilot production in 2027.
AWS Boosts Trainium3 ASIC Shipments, Accelerating Custom AI Chip Ecosystem Against NVIDIA
Amazon AWS has notified its supply chain to increase Q3 2026 shipments of Trainium3-based ASIC servers by 20-30%. This reflects growing confidence in its custom AI chips and a strategic push to reduce reliance on NVIDIA GPUs. AWS also partnered with OpenAI to develop a Stateful Runtime Environment on Bedrock.
AWS boosts Trainium 3 shipments, accelerating ASIC substitution for NVIDIA GPUs
Supply chain sources indicate Amazon AWS has instructed vendors to increase Trainium 3 shipments for Q3 2026 by 20-30%. This signals strong confidence in its custom ASIC strategy to reduce dependence on NVIDIA GPUs, leveraging superior cost and power efficiency for cloud AI training.
Anthropic Starts Custom AI Chip Development, Talks Samsung 2nm, Aims for Compute Independence
Anthropic has initiated its own AI chip development and is in talks with Samsung for 2nm foundry services. The move aims to reduce reliance on NVIDIA GPUs, optimize inference costs, and strengthen its technology moat ahead of a potential IPO. It joins OpenAI, Google, and others in the custom ASIC race, signaling a shift from software to hardware competition.
AWS Trainium 3 Shipments Surge 20-30%, Shifting AI Compute Control from NVIDIA to Custom Silicon
Supply chain sources indicate AWS has raised Q3 Trainium 3 server shipments by 20-30%, driven by Anthropic. Trainium 2 is sold out, Trainium 3 nearly fully booked, with customers already queuing for Trainium 4 and development of Trainium 5 underway. This signals AWS's aggressive push to own the AI compute stack via custom silicon.
Meta Shifts MTIA ASIC to Samsung 2nm: Ecosystem Restructuring in AI Chip Fab
Meta partners with Samsung for next-gen MTIA ASIC production, moving from TSMC to Samsung 2nm node. Targeting hundreds of thousands of units to support 5GW data center goal by 2030, with new chip every six months, restructuring the AI chip supply chain ecosystem.
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.
AWS and Google Open Custom AI Chips for External Sales, ASIC Shipment Growth Surpasses GPU, TCO Inflection Point Reached
In Q2 2026, AWS Trainium and Google TPU are commercialized externally for the first time. Custom ASIC shipment growth of 44.6% surpasses GPU's 16.1%. ASIC TCO advantage reaches 40-65% for large-scale inference; Midjourney cut monthly compute cost from $2.1M to $0.7M after migrating to TPU. This marks a structural inflection point in AI compute.
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.
OpenAI and Broadcom Tape Out First Inference ASIC Jalapeño in 9 Months, Targeting NVIDIA Dominance
OpenAI and Broadcom unveil Jalapeño, their first custom inference ASIC, fabricated on TSMC 3nm and optimized for Transformer models. Targeting a 50% inference cost reduction, it taped out in 9 months and is slated for deployment in gigawatt-scale data centers by late 2026, marking OpenAI's strategic pivot to full-stack AI infrastructure and a direct challenge to NVIDIA's inference hegemony.
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 Unveils AI-Centric Network with Token Monetization, UCM Caching Breaks Long-Context Barriers
At MWC Shanghai 2026, Huawei unveiled an AI-native network architecture integrating service, network, and compute, shifting from traffic-centric to intelligence-centric operations. The Unified Cache Manager (UCM) extends KV cache to petabyte-scale external storage, achieving 372% token throughput gains on GLM-5.1 at 128K sequence lengths. Token monetization frameworks and agentic operations enable carriers to charge for AI inference capacity and personalize services.
OpenAI and Broadcom Unveil Jalapeno Inference ASIC, Reshaping AI Hardware Landscape
OpenAI, in collaboration with Broadcom, has developed Jalapeno, a custom LLM inference accelerator. The chip uses a multi-chip module with HBM3E memory and achieved tape-out in just nine months. Designed for OpenAI's model stack, it aims to reduce inference costs and dependency on NVIDIA GPUs, with initial deployment planned for late 2026.
TSMC Hikes Advanced Node Prices 5-10%, Squeezing AI Chip Margins
TSMC informs clients of 5-10% price hikes across all advanced nodes (7nm+), affecting 74% of wafer revenue. Apple, Nvidia, AMD, and others face higher costs, potentially raising AI infrastructure prices.
Arm Server Share Hits 45%: NVIDIA's Bundling Strategy Reshapes AI Infrastructure
IDC data shows Arm-based servers now hold over 45% of the global server market, driven by NVIDIA's bundling of its Arm-based Vera CPU with GPU systems like NVL72 and Rubin. x86 share shrinks to 52%, while accelerated systems contribute over 70% of revenue. ODM direct sales account for 50.2%, with Dell revenue growing 244.1% YoY.
MediaTek Lands Exclusive Google TPU v9 Inference Upgrade Triggerfish with 2x SRAM
Google plans a TPU v9 inference upgrade, Triggerfish, exclusively fabbed by MediaTek. It features 2-3x on-chip SRAM, HBM4E DRAM, and a simulation die for local management. Production starts late 2027 with 1-2M units lifecycle, unit price ~30% higher than Humufish.
Google TPU v9 Switches to MediaTek, Breaking Broadcom's AI ASIC Monopoly
Google moves its TPU v9 Humufish design and integration contract from Broadcom to MediaTek, which handles I/O chip design and packaging. Combined with a split-foundry strategy (TSMC N2 compute, Samsung 2nm I/O), this marks a systematic effort to build a multi-vendor, multi-node supply chain, directly dismantling Broadcom's dominance in custom AI ASICs.
NVIDIA Dominates TOP500 with Full-Stack Lock-in: Grace CPU, InfiniBand, and GPU Integration
NVIDIA powers 81% of TOP500 supercomputers, with Grace CPU adoption rising to 26 systems and Quantum InfiniBand connecting 376. The full-stack strategy (GPU+CPU+networking) shifts procurement from open components to single-vendor lock-in; top 8 Green500 systems use NVIDIA GPUs.
AWS Lambda MicroVMs: Stateful Isolated Sandboxes via Firecracker Snapshots
AWS launches Lambda MicroVMs, leveraging Firecracker for VM-level isolation, near-instant launch/resume, and stateful execution. Users build images from Dockerfiles in S3, launch from pre-initialized snapshots, and suspend/resume automatically, enabling multi-tenant AI code sandboxes and interactive analytics.
Arm servers capture >45% data center revenue, x86 ecosystem under AI-driven assault
IDC reports Q1 2026 global server revenue hit a record $122.6B, with Arm-based servers capturing >45% share (x86 at 52%). Accelerated servers (GPU/ASIC/FPGA) generated >70% revenue. Nvidia's Grace CPU (NVL72) and hyperscaler custom Arm chips drive the shift; x86 still leads in unit volume but faces supply constraints.