Executive Summary

Sahabat-AI is one of the clearest company-led expressions of Indonesia's sovereign and local-language AI ambitions. When GoTo and Indosat launched the project in November 2024, they framed it as an open-source LLM ecosystem for Bahasa Indonesia and local languages, explicitly tied to digital sovereignty and Indonesia's wider 2045 ambitions.¹ By June 2025, the effort had scaled into a 70-billion-parameter model and multilingual chat service available through sahabat-ai.com and the GoPay app.²

That progression matters because it shows Indonesia's AI story moving through language relevance, locally hosted infrastructure, and mass-market distribution rather than through imported frontier branding alone.

Why Sahabat-AI Fits Indonesia

The project is strategically well matched to Indonesia's real constraints and opportunities. GoTo and Indosat built the initiative around Indonesian and regional languages, local cultural context, and open access for developers, universities, and public institutions.¹² That is the right lane for a market where linguistic diversity, mass adoption, and public-service relevance matter more than chasing a global benchmark headline.

The June 2025 update strengthened that point by expanding the model to operate across Bahasa Indonesia, Javanese, Sundanese, Balinese, and Bataknese, while also emphasizing locally accessible infrastructure and use by startups, university labs, and public-service institutions.²

Why the Distribution Layer Matters

Sahabat-AI is not only a model story. It is also a distribution story. GoTo says the chat service is available through GoPay, giving the project access to a mass consumer interface, while Indosat links the system to its GPU Merdeka sovereign AI cloud.² That combination is unusually important. It means Indonesia's local-language model effort is being tied to both infrastructure and user reach.

That is what makes the project more durable than a one-off research milestone. If local models are going to matter in Indonesia, they need to be usable by developers, reachable by citizens, and credible enough for government and enterprise deployment. Sahabat-AI is one of the few efforts visibly trying to cover all three.

Why This Matters for Indonesia's AI Position

Indonesia's AI future is unlikely to be won through a single national lab or one central ministry alone. It will depend on whether company-led systems can align with local languages, sovereign infrastructure, and public-service possibilities. Sahabat-AI matters because it gives Indonesia a plausible route into that middle layer between state ambition and end-user adoption.

The collaboration model matters too. GoTo and Indosat describe a wider ecosystem involving universities, media groups, researchers, and public bodies.¹² That makes Sahabat-AI look less like a proprietary product launch and more like an attempt to seed a national AI operating layer.

What To Watch

The most important signals will be broader public-sector use, more enterprise integrations, stronger open-source adoption, and continued progress on local-language performance beyond Bahasa Indonesia alone. If those signals keep improving, Sahabat-AI could become one of the strongest proofs that Indonesia's AI edge will come from local relevance and distribution, not just imported models.

Sources

1. Indosat Ooredoo Hutchison and GoTo Launch Sahabat-AI: Indonesia's Open-Source LLM for Empowering Digital Sovereignty
2. Sahabat-AI Gets Smarter: Indosat and GoTo Launch New 70 Billion Parameter Model with Multilingual Chat Service