Should Samsung build its own AI models for Galaxy AI or keep relying on Google Gemini and Perplexity?

Introduction

Galaxy AI is Samsung's most important software story since One UI itself, but much of its intelligence comes from partners. Google's Gemini powers Circle to Search, many generative features and the default assistant; on the Galaxy S26, Perplexity joined as a second assistant and reportedly powers parts of the new Bixby. Samsung also builds its own models (Samsung Gauss) and runs many features on-device. The strategic question: should Samsung invest heavily in its own frontier AI models and assistant, or deepen its partner-first approach with Google and Perplexity?

Key business objectives for this strategy include:

1. Differentiating Galaxy against iPhone and Chinese Android brands
2. Protecting Samsung's direct relationship with over a billion device users
3. Controlling AI costs as usage scales
4. Opening future revenue from AI services

In addressing this challenge, I'll follow these steps:

1. Clarify our strategic goals and context
2. Analyse the market and competitive landscape
3. Define product vision and roadmap
4. Prioritise initiatives
5. Plan execution and resources
6. Set metrics
7. Manage risks

Step 1

Clarify the Strategic Goals (3-4 minutes)

• Is the goal to win the AI phone narrative, or to build a long-term AI platform across all Samsung devices (TVs, appliances, SmartThings)?

Why it matters: A phone-only goal favours partners; a cross-device platform favours more ownership. Expected answer: Long-term platform across devices, with phones as the lead. Impact on approach: We'll consider an orchestration layer that Samsung owns even if models are partly external.

• What time horizon and investment envelope are we discussing?

Why it matters: Frontier model training costs billions per generation. Expected answer: A five-year horizon; Samsung will invest significantly but not at the scale of Google or OpenAI. Impact on approach: Building a frontier general model is unlikely to be the best use of capital.

• How constrained are we by the Google relationship (Android, Play, Gemini default agreements)?

Why it matters: Galaxy depends on Android and Google services; switching defaults has contractual and regulatory implications. Expected answer: Google remains a core partner; Samsung wants leverage, not a breakup. Impact on approach: Multi-partner strategy is a lever for negotiation.

Based on these insights, our key strategic assumptions are:

1. Samsung wants AI as a cross-device platform, not just a phone feature.
2. Samsung will not out-spend Google on frontier models.
3. Keeping multiple partners gives Samsung leverage and resilience.