Asro
Taiwan stands at a critical juncture as the global artificial intelligence computing revolution reshapes the geopolitical and economic landscape, prompting a specialized industry committee to issue a comprehensive set of policy recommendations aimed at securing the island’s semiconductor dominance. While Taiwan remains the undisputed leader in advanced chip manufacturing, the Committee warns that structural constraints—ranging from energy volatility to talent shortages—could undermine national resilience and technological competitiveness if left unaddressed. As the industry migrates toward sub-2-nanometer nodes and massive AI deployment, the call for coordinated public-private action has intensified, focusing on four pillars: energy security, talent retention, tax reform, and a transition toward a distributed "cloud-to-edge" AI model.
The Geopolitical and Economic Context of Taiwan’s Silicon Shield
For decades, Taiwan’s semiconductor industry has served as a "Silicon Shield," providing a layer of geopolitical protection due to the world’s reliance on its high-end chips. However, the rise of generative AI has fundamentally altered the demand profile for computing power. By 2028, electricity demand driven by AI workloads is projected to reach historical highs, placing unprecedented strain on a power grid that is already operating under thin margins. The Committee’s report highlights that Taiwan’s leadership is no longer just a matter of fabrication precision but is now tied to the stability of the entire value chain, including energy inputs and the ability to attract global experts.
The global semiconductor race has seen the United States, Japan, and the European Union pour billions into domestic chip production via initiatives like the U.S. CHIPS and Science Act. In response, Taiwan must refine its internal policies to ensure it remains the most attractive platform for the global digital economy. The Committee’s findings suggest that while manufacturing strength is high, the "soft" infrastructure—policy, energy, and human capital—requires urgent modernization.
Prioritizing Energy Resilience and Predictability
The most pressing concern identified is the sustainability of Taiwan’s power supply. The semiconductor industry is notoriously energy-intensive; a single extreme ultraviolet (EUV) lithography machine consumes significantly more power than previous generations of equipment. With AI-driven demand expected to surge, the Committee emphasizes that energy security has become a strategic resource equivalent to the chips themselves.
Strengthening LNG Infrastructure and Legal Frameworks
Taiwan’s current energy mix is heavily reliant on imported Liquefied Natural Gas (LNG), which has overtaken coal as the primary source of power generation. This creates a significant vulnerability: nearly 98% of Taiwan’s energy is imported. Geopolitical tensions in the South China Sea and the Taiwan Strait pose a direct threat to the maritime supply lines required for LNG delivery.
To mitigate this, the Committee recommends a two-pronged approach. First, the government must accelerate the construction of LNG receiving terminals, such as the controversial yet vital Third LNG Terminal. Capacity constraints at existing terminals currently limit Taiwan’s ability to build a sufficient buffer. Second, the Committee calls for an amendment to Article 31 of the Natural Gas Industry Act. Currently, LNG stockholding requirements are managed through administrative measures that provide limited buffer capacity. By comparison, neighboring economies like Japan and South Korea maintain statutory reserves that are significantly higher, allowing them to weather supply disruptions for longer periods. Establishing clear, enforceable, and statutory minimum stockholding requirements is seen as essential for industrial stability.
Predictability in Pricing
Beyond supply, the Committee notes that electricity price volatility creates significant hurdles for long-term corporate budgeting. As the government adjusts rates to reflect global fuel costs and the transition to greener energy, the Committee urges for a "no-surprises" approach. Providing advance notice of rate changes would allow manufacturers to adjust their operational plans without suffering sudden shocks to their bottom lines.
The Global War for Talent: Enhancing Tax Competitiveness
The second priority focuses on the human element of the semiconductor ecosystem. While Taiwan has successfully launched the Employment Gold Card program to streamline visas for foreign professionals, the Committee argues that the current tax incentives are insufficient for long-term retention. High-skilled professionals in the semiconductor and AI fields are part of a global mobile workforce that evaluates destinations based on post-tax income and long-term stability.
The 5+5 Incentive Model
Currently, tax benefits for foreign professionals in Taiwan are limited to a five-year window. The Committee suggests adopting a "5+5" framework, similar to models used in Italy. This would allow an initial five-year tax benefit to be extended for another five years if the individual meets specific criteria, such as continued employment or relocating their family to Taiwan.
Furthermore, the Committee points out that after the incentive period ends, Taiwan’s standard tax rates can be less competitive than those in the Netherlands or Spain, which offer flat rates or partial exemptions for qualifying researchers. To prevent a "brain drain" after the initial five-year period, a review of the top-tier tax brackets for specialized tech talent is recommended.
Equity-Based Compensation Reform
The report also targets Article 19-1 of the Industrial Innovation Act, which governs equity-based compensation. In the tech world, stock options are a primary tool for alignment between employees and company success. The Committee suggests that current caps and taxation timing—often occurring at the moment of exercise rather than sale—can create liquidity issues for employees and reduce the attractiveness of these incentives.
Amending the Industrial Innovation Act for High-Growth Firms
A critical technical recommendation involves Article 10-2 of the Industrial Innovation Act, often referred to as Taiwan’s "Chips Act." While the law was intended to incentivize R&D, the implementation regulations have created an "intensity ratio" trap. Currently, companies must meet a minimum R&D expenditure threshold and a specific R&D-to-revenue percentage.
The Committee notes that during periods of high revenue growth—which are common in a booming AI market—a company’s R&D intensity ratio may actually drop even if their absolute R&D spending increases by millions of dollars. This creates a paradoxical situation where the most successful, scaling companies are disqualified from tax incentives precisely when they are investing the most in next-generation technology. The Committee recommends decoupling these thresholds or adjusting the intensity requirements to ensure that long-term R&D planning is not penalized by short-term revenue spikes.
The Strategic Shift: From Cloud AI to Edge AI
Perhaps the most forward-looking aspect of the report is the call for a "distributed cloud-to-edge" AI strategy. Currently, most AI processing occurs in massive, centralized data centers. While this is efficient for training Large Language Models (LLMs), it places immense pressure on the power grid and network bandwidth.
The Benefits of Edge Computing
Edge AI involves performing data processing directly on the device—whether it is a smartphone, an autonomous vehicle, or a factory robot—rather than sending it to a remote server. The Committee argues that promoting Edge AI will:
- Reduce Energy Strain: By distributing the workload, the massive power spikes associated with centralized data centers can be mitigated.
- Enhance Security: Data processed on-site is less vulnerable to interception during transmission, a key concern for sensitive industrial applications.
- Improve Latency: Real-time applications, such as autonomous driving or robotic surgery, require the instantaneous response times that only Edge AI can provide.
The Committee urges the government to integrate Edge AI into national flagship programs like the "Chip-based Industrial Innovation Program." This includes recognizing AI-capable devices (PCs and workstations) as part of the national AI infrastructure and providing grants or tax vouchers for SMEs to adopt edge solutions.
International Cooperation and Cross-Ministerial Synergy
Finally, the Committee emphasizes that AI and semiconductor policy cannot be siloed within a single ministry. The development of an AI ecosystem involves land use, environmental review, grid development, and national security. The report calls for an Executive Yuan-level coordinating authority to oversee AI policy across ministries, ensuring that the Ministry of Economic Affairs, the Ministry of Digital Affairs, and the National Science and Technology Council are aligned.
On the international front, the Committee highlights the importance of the U.S.-Taiwan Economic Prosperity Partnership Dialogue (EPPD). Deepening cooperation with the U.S. on supply chain security, drones, and robotics is seen as essential for creating a "trusted AI technology ecosystem." The report specifically mentions the need for high-quality traditional Chinese-language datasets to ensure that AI models developed in the West are culturally and linguistically accurate for the Taiwan market, further cementing Taiwan’s role as a critical partner in the global democratic supply chain.
Analysis of Implications
The Committee’s recommendations represent a roadmap for Taiwan to evolve from a "foundry for the world" into a comprehensive "AI island." If the government adopts these measures, particularly the statutory LNG reserves and the "5+5" tax model, it could significantly de-risk the semiconductor sector in the eyes of global investors.
However, the transition is not without challenges. Accelerating LNG infrastructure often meets local environmental resistance, and adjusting tax codes for foreign professionals can raise questions of domestic equity. Nevertheless, the fact-based analysis provided by the Committee suggests that the cost of inaction—potential power outages, a thinning talent pool, and missed opportunities in the Edge AI market—is far higher. By addressing these structural constraints now, Taiwan can reinforce its position not just as a manufacturer, but as the indispensable backbone of the 21st-century digital economy.
Lina Irawan
Pevita Pearce
