Piper Sandler’s $2.2 Trillion AI Compute Forecast: Why These Five Chips Are Non-Negotiable

Posted on by

(SeaPRwire) –

By: Ethan Gallagher, a Silicon Valley Hardware Architect and Infrastructure Strategist.

These five names are not recommendations; they are survival maps for the next hardware cycle. The analyst dismisses incremental upgrades, focusing on raw compute throughput and supply rigidity. Nvidia stands as the unavoidable core, its 80 percent AI compute dominance creating a moat supply constraints will not breach easily. Broadcom complements this with ASIC control points, while AMD, Arm, and Marvell each anchor critical infrastructure layers. Any portfolio ignoring this concentration risks obsolescence.

Nvidia receives a $300 target, implying 34 percent upside as the outright leader. O’Connor describes the AI compute market hitting $2.2 trillion by 2030, with supply likely staying tight for two to three years. The rise of agentic AI workloads this year has pushed demand higher, pricing the chipmaker at roughly 14 times fiscal 2028 earnings estimates. Broadcom enters at $460, holding 75 percent of the ASIC market for AI inference and visibility into 12 gigawatts of fiscal 2027 demand. Its custom chips are deemed the cheapest in the AI universe.

AMD earns the highest price target at $600, around 15 percent upside, framed as an Agentic AI Sweetspot. Key clients like OpenAI, Meta, and Anthropic rely on its server CPUs and Helios GPUs, driving revenue growth at 50 percent compound annual rate through fiscal 2030, with earnings per share growing at 65 percent. Arm Holdings at $320 offers 21 percent upside, leveraging CPU IP dominance with potential to double earnings by capturing 10 percent of the ASIC market. Marvell at $270 gains traction from a $120 billion Google agreement, labeled transformative, with an October analyst day as a near-term catalyst.

Piper Sandler also rates Intel and Qualcomm Neutral, with targets at $110 and $190 respectively. Intel’s gains are seen as already reflected in foundry expectations, while new Qualcomm design wins appear priced in. This selective activation of capital toward the top five creates a bifurcated landscape where execution precision determines survival. The supply chain will reward those who secure capacity and architecture control.

Author bio: Ethan Gallagher, a Silicon Valley Hardware Architect and Infrastructure Strategist, dissects semiconductor bottlenecks through hands-on design cycles and long-term node roadmaps.