Taiwan Mobile: AI and Enterprise Services Drive Growth

Taiwan Mobile has announced its goal to achieve NT$1 trillion in revenue, placing artificial intelligence (AI) and enterprise services at the core of its growth strategy. This move underscores the increasing importance of AI not only as an enabling technology but as a true economic engine for large telecommunications companies and service providers.

Taiwan Mobile's ambition reflects a global trend: enterprises are integrating AI into every aspect of their operations, from customer service to network optimization, and the development of new products. To achieve goals of this magnitude, the choice of underlying infrastructure becomes a critical factor, directly influencing the ability to innovate and maintain a competitive edge.

AI and Enterprise: Deployment Challenges

The adoption of AI, particularly Large Language Models (LLM), within the enterprise context presents significant challenges, especially concerning deployment. Companies find themselves at a crossroads: relying on cloud-based solutions or investing in self-hosted infrastructures. The decision is not trivial and involves a careful analysis of multiple factors.

For enterprise services, data sovereignty is often a top priority. Sectors such as finance, healthcare, or public administration are subject to stringent regulations that mandate data residency and specific controls. In these scenarios, an on-premise or air-gapped deployment can offer the necessary level of control and compliance, even if it entails a higher initial investment and more complex internal management. The long-term Total Cost of Ownership (TCO), which includes operational, energy, and licensing costs, becomes a fundamental parameter for evaluating the sustainability of each option.

Hardware and Data Sovereignty: The Role of Self-Hosted Solutions

To support complex AI workloads, such as inference for large LLM or fine-tuning of specific models, hardware plays a crucial role. GPUs with high amounts of VRAM and computing capabilities are essential. Choosing a self-hosted infrastructure allows companies to select the hardware best suited to their specific needs, optimizing performance for tokens/sec or p95 latency, and scaling the environment based on demand.

On-premise deployment also offers granular control over the entire AI pipeline, from data management to model orchestration. This is particularly relevant for companies that wish to maintain intellectual property and the security of their sensitive models and data within their own boundaries. For those evaluating on-premise deployment, AI-RADAR offers analytical frameworks on /llm-onpremise to assess the trade-offs between initial costs, operational flexibility, and security requirements.

Future Outlook for AI Infrastructure

Taiwan Mobile's focus on AI and enterprise services is emblematic of a broader transformation in the technological landscape. Companies are no longer just looking for "off-the-shelf" AI solutions but desire platforms that offer control, customization, and security. This drives an increasing adoption of hybrid or fully on-premise strategies for the most critical AI workloads.

The ability to manage AI infrastructure internally, or through specialized partners in self-hosted solutions, is becoming a competitive differentiator. It allows not only for compliance with data sovereignty regulations but also for optimizing resources and developing strategic internal competencies. The infrastructure decisions made today will determine companies' ability to fully capitalize on the potential of AI in the coming decade.