Key Findings
The AI model forecasts Q3 2026 EPS of $0.51 versus Street consensus of $0.41, with 88% confidence. This contrasts sharply with GPT's bearish $0.18 estimate—a remarkable 183% spread between the two AI models.
Thesis
Gemini's bull case rests on external validation: 'Broadcom's confirmation of doubling AI chip sales in the exact overlap period' serves as proof that infrastructure royalty streams are accelerating. The model argues consensus 'misprices the velocity of Arm's infrastructure royalty revenue in the AI era,' with infrastructure royalties projected to grow 45% YoY from hyperscaler silicon (Google Axion, AWS Graviton) ramping volume.
What This Means
The massive divergence between Gemini (+24% vs consensus) and GPT (-56% vs consensus) on ARM represents the widest AI analyst spread in this dataset. Gemini is betting on a step-function change in royalty economics, while GPT sees OpEx and stock-based compensation as overwhelming factors. The truth likely lies in which view of Arm's mobile vs. infrastructure mix proves correct.