ARM Holdings Plc (ARM.US), a leader in chip design and provider of widely used processor architectures, delivered revenue forecasts surpassing Wall Street expectations. This optimism stems from unprecedented global demand for AI data centers, fueled by massive AI training/inference workloads. A surge in ARM-based infrastructure adoption—driven by NVIDIA’s Grace CPU and cloud giants like Amazon, Microsoft, and Google deploying custom ARM server CPUs—has significantly expanded ARM’s business, with management issuing robust guidance.

Amid recent market turmoil over "AI bubble" fears, ARM’s stellar performance has emerged as a pillar of confidence for AI-driven markets, even outperforming the bullish impact of rival AMD’s (AMD.US) earnings. ARM’s results reinforce the "long-term AI bull thesis," signaling sustained momentum for the AI compute sector led by NVIDIA, TSMC, Broadcom, Micron, and ARM itself. Post-earnings, ARM shares surged over 9% in after-hours trading, lifting peers across the AI supply chain. Year-to-date, ARM’s stock has soared 110%, pushing its market cap above $160 billion, with SoftBank retaining nearly 90% ownership.

For fiscal Q3 2026 (calendar Q4 2025), ARM projected midpoint revenue of $1.23 billion (range: $1.175B–$1.275B) and non-GAAP EPS of $0.41 ($0.37–$0.45), beating consensus estimates of $1.1 billion and $0.35, respectively. ARM’s revenue streams include licensing fees for its designs/architectures and royalties per chip shipped. Analysts note that strong licensing sales today will translate to future royalty growth.

In fiscal Q2 (ended September), revenue jumped 34% YoY to $1.14 billion, with non-GAAP EPS of $0.39, both exceeding forecasts. Operating profit skyrocketed 155% to $163 million (14.4% margin vs. 7.6% YoY), while net income rose 122% to $238 million. Licensing revenue surged 56% to $515 million (above $472M estimates), and royalties grew 21% to $620 million ($586M expected).

CEO Rene Haas highlighted booming demand for ARM’s Neoverse data center architecture, with related revenue doubling. Originally dominant in mobile (e.g., Apple’s iPhone chips), ARM is now penetrating power-constrained data centers, competing with x86 in high-efficiency server CPUs. Collaborations with NVIDIA, Amazon (Graviton), Google (Axion), and Microsoft (Azure Cobalt) underscore ARM’s pivot to AI infrastructure.

ARM’s RISC architecture offers 3.5x better performance-per-watt than x86 in cloud workloads, making it ideal for cost-sensitive, energy-efficient AI data centers. Projects like SoftBank/OpenAI’s "Stargate" and Japan’s AI initiatives further cement ARM’s role in global AI expansion.

Meanwhile, soaring DRAM/NAND prices, OpenAI’s $1 trillion infrastructure deals, and upbeat forecasts from TSMC, Samsung, and SK Hynix validate the AI compute sector’s resilience. NVIDIA CEO Jensen Huang recently dismissed bubble fears, citing tangible ROI from AI infrastructure investments. UBS analysts argue the AI-driven market remains in early stages, far from dot-com-level excesses, with valuations and tech/GDP ratios still moderate.

ARM’s results and the broader AI ecosystem’s strength counter "bubble" narratives, underscoring sustained growth in GPU, ASIC, HBM, and data center infrastructure.