Microsoft has demonstrated its seriousness in the artificial intelligence arena by announcing a US$1.7 billion investment in Indonesia. This funding is allocated to build its first data center in the country, strengthening cloud and AI infrastructure, and supporting Indonesia’s digital transformation ambition toward “Indonesia Emas 2045”.
Beyond relying on OpenAI, Microsoft is now developing its own internal AI models, known as Microsoft AI (MAI). This move aims to reduce dependency on external technologies, while Microsoft is also testing other AI models such as DeepSeek and Meta AI.
In various 2025 benchmarks, Microsoft Copilot still ranks below top AI players such as ChatGPT-4o, Grok 3, DeepSeek R1, and Gemini 2.0. Copilot excels in Office productivity integration, but in terms of deep reasoning and creative writing, it still trails behind standalone AIs.
AI Model | Strengths | Weaknesses |
ChatGPT-4o | Superior reasoning, strong coding skills | Limited in source citation |
Grok 3 | Real-time reasoning, Big Brain mode | Limited public access |
DeepSeek R1/V3 | Highly efficient mathematical reasoning | Weaker visual reasoning |
Gemini 2.0 | Multimodal text-image-code capabilities | Issues with time and context understanding |
Microsoft Copilot | Maximum productivity integration in Office | Lagging in complex reasoning capabilities |
Recent data shows that ChatGPT dominates 59.5% of the AI chatbot market, while Microsoft Copilot holds 14.3%. Despite trailing in open-ended reasoning, Copilot retains a loyal user base thanks to its strong performance in business and productivity sectors.
Through massive investments and the development of proprietary AI models, Microsoft demonstrates a serious commitment to AI leadership. However, Copilot’s position in the global AI race still trails behind ChatGPT-4o, Grok 3, and DeepSeek in terms of deep reasoning and freeform creativity. Productivity integration remains Copilot’s key strength, but innovation in complex reasoning must be accelerated to close the gap.