AI Chips

SambaNova's $1B Series F at $11B valuation, backed by JPMorgan Chase going on-prem, signals that 55% of AI compute spending has shifted to inference. The AI inference market hit $117.8B in 2026 and is growing at 13% CAGR. NVIDIA's GPU monopoly on inference is fragmenting as Cerebras IPOs at $56B, Qualcomm acquires Modular for $3.9B, and OpenAI builds custom ASICs. Enterprise vendor selection matrix and on-prem inference readiness assessment inside.

OpenAI and Broadcom unveiled Jalapeño, a custom inference ASIC designed from scratch for LLM workloads. Built in nine months with AI-assisted design, it claims 50% lower cost per token than NVIDIA GPUs. With Google, Amazon, and Microsoft all building competing custom silicon, the era of GPU-only inference is ending — and the enterprise AI cost structure is about to be rewritten.

Huawei expects AI chip revenue to reach $12B in 2026 (up 60% YoY) as ByteDance, Alibaba, and Tencent place massive orders for Ascend 950PR processors—reshaping the enterprise AI chip market and forcing CIOs to reconsider vendor strategies for China operations.

OpenAI just committed over $20 billion to Cerebras chips over three years. For enterprise leaders, this deal reveals critical lessons about vendor lock-in, infrastructure cost optimization, and the hidden risks of betting on Nvidia alternatives.

Arm Enters AI Chip Manufacturing With $15B Revenue Target as Citi Calls It 'Most Significant Shift' in Company History. For enterprise decision-makers: strat...