As the global AI race expands from model performance to data center infrastructure, liquid cooling is fast becoming the next technological high ground. National Central University (NCU), led by President Shaw-San Hsiao, sent a delegation to Silicon Valley to visit Cooler Master and Supermicro, focusing on cooperation in AI data center liquid cooling, green computing and talent development. This is more than an exchange between a university and companies. It shows Taiwan actively working to secure a strategic position in the next wave of AI hardware and high-performance computing upgrades.

Key points: The core of NCU's US trip is building international industry-academia partnerships for next-generation AI infrastructure. From liquid cooling and green computing to talent links with a research center in Taoyuan, the partnership shows Taiwan moving beyond its strength in AI hardware manufacturing toward higher-value system technology and a role as a regional talent hub.

Technical Breakdown: Why Do AI Data Centers Now Need Liquid Cooling?

Key Takeaway: AI compute density is rising fast, and traditional air cooling can no longer balance cooling efficiency, rack power density and energy costs. NCU's focus on liquid cooling with Cooler Master and Supermicro reflects how competition in AI infrastructure has expanded from chip performance to thermal management and green computing capability.

With generative AI, high-performance computing and large-scale model training continuing to heat up, data centers face a very practical problem: chips keep getting more powerful, and they keep getting hotter. As the power draw of a single AI server keeps rising, traditional air-cooled designs built on fans and air conditioning are approaching their limits in efficiency and space. That is the key reason National Central University President Shaw-San Hsiao led the delegation to the US with AI data center liquid cooling at the center of the discussions.

Liquid cooling matters not just because it "cools faster," but because it directly affects three core metrics of AI infrastructure: compute stability, energy efficiency and data center scalability. For large AI model training, insufficient cooling can cause servers to throttle when they overheat, which in turn affects training time, inference latency and overall operating costs. In other words, cooling is no longer just an equipment maintenance issue. It is part of AI competitiveness itself.

  • Higher cooling efficiency: Liquids conduct heat better than air and can remove heat from GPUs and high-power components more effectively.
  • Higher rack density: Data centers can deploy more high-end AI servers in limited space.
  • Lower power pressure: Reducing the load on air conditioning and fans helps improve PUE.
  • Support for green computing: It aligns with energy-saving goals and carbon management needs, and better fits future regulations and corporate ESG direction.

The two companies NCU visited also represent two key roles in the AI hardware supply chain. Cooler Master has long specialized in cooling solutions and can contribute heat exchange, cold plates, mechanical integration and system-level thermal optimization. Supermicro is a major player in AI servers and data center platforms and can bring cooling technology into large-scale commercial deployment. This combination of "academic research institution + cooling technology vendor + server platform vendor" means the partnership is no longer just an exchange of ideas. It is moving toward a complete chain that can be validated, deployed and used to train talent.

More notably, NCU did not focus on a single technology on this trip. It looked at liquid cooling within the overall architecture of next-generation AI systems. From advanced cooling and green computing to talent development, this shows Taiwan's academic community has begun to understand that competition in the AI era does not depend only on models or chips. It requires cross-disciplinary capability that integrates materials, mechanical design, thermal fluids, power, system design and operations management. For Taiwan, this is a very practical and very timely signal.

Upgrading Industry-Academia Collaboration: Why Did NCU Open International Partnerships in Silicon Valley?

Key Takeaway: This US trip was not just a visit. It centered on international industry-academia collaboration and targeted joint research and development of next-generation AI technology. For the university, it can shorten the distance between research and industry needs. For the companies, it allows earlier planning for technology validation and talent supply.

Silicon Valley's significance to the global tech industry has never been just as a landmark of innovation. It is where technology standards, supply chain partnerships and business model validation meet. With the university president personally leading the delegation to visit companies there, NCU sent a clear signal: Taiwan's universities are shifting from traditional academic cooperation toward a more industry-oriented model of international alliances.

According to the information provided, the trip focused on international industry-academia collaboration to develop next-generation AI technology, and discussed joint research directions with the companies, including green computing, advanced cooling technology and talent development. This agenda is crucial, because what the AI industry needs most in its next stage is not only algorithm talent, but multi-skilled engineers who understand system implementation, energy management and infrastructure bottlenecks. If universities get involved early, they can strengthen their research capacity and also align their courses and labs with industry needs.

From an industry perspective, companies willing to deepen cooperation with universities usually signal three needs emerging at the same time:

  • Early-stage R&D: Companies want to explore new architectures and new materials with academia before products are commercialized.
  • Validation and testing grounds: Universities can provide research teams, lab resources and theoretical models to help optimize technology.
  • Talent supply chain: With demand for AI infrastructure surging, companies need a more stable source of talent.

This also shows that the importance of NCU's trip lies not in a one-off news moment, but in the possibility that it marks the start of a partnership framework for years to come. In AI data centers especially, cooling technology spans a wide range of fields, from fluid mechanics, heat conduction and mechanical design to electrical control and system integration, all of which require cross-disciplinary collaboration. If Taiwan's higher education system can step in now, it can not only keep up with the AI wave but may also find a more differentiated position in the global infrastructure upgrade.

In addition, having the university president lead the delegation personally carries significance for governance and resource integration. The real difficulty in many cross-border industry-academia partnerships is not the technology itself, but whether budgets can be sustained, departments integrated, long-term contact points established, and research results turned into prototypes and frameworks companies can use. From this angle, NCU's move looks more like building a model case for the internationalization of AI in Taiwan's higher education.

Opportunities for Taiwan: How Can a Taoyuan R&D Hub Expand the Asia-Pacific AI Talent Pool?

Key Takeaway: Supermicro's Taiwan R&D center in Taoyuan means this partnership does not have to stay at the level of exchanges in the US. It has a chance to connect directly with local R&D, the supply chain and talent development in Taiwan. For the Asia-Pacific market, this would help establish an AI technology training base closer to where manufacturing and deployment happen.

The point in this news most worth the Taiwan market's attention is that the partnership specifically mentions using the geographic advantage of Supermicro's Taiwan R&D center in Taoyuan to strengthen the AI technology talent pool across the Asia-Pacific region. This means the model is not a one-way import of technology. It could form a regional network of "cutting-edge technology exchange in the US, R&D implementation in Taiwan, and talent spreading across Asia-Pacific."

Taoyuan itself has strong geographic and industrial advantages. It links an international airport, industrial clusters, the electronics supply chain and northern Taiwan's academic and research resources, making it an important node connecting manufacturing with international business. If R&D, validation and talent training for AI servers, liquid cooling systems and related components can cluster there, Taiwan would see three direct benefits.

  • First, a bigger voice for Taiwan in the AI hardware value chain. Taiwan has traditionally been strong in contract manufacturing and components, but if it can also master system cooling and data center efficiency technology, its added value will be higher.
  • Second, stronger hands-on skills for local talent. If students and research teams are exposed to real company needs, it becomes easier to develop engineering capabilities aimed at international markets.
  • Third, a stronger role as an Asia-Pacific showcase. Taiwan could become a testbed for deploying liquid-cooled AI servers, green data centers and high-density computing.

For Taiwan's tech industry, this could also bring more opportunities along the supply chain. Liquid cooling is not a single piece of equipment but an entire industry chain, including cooling modules, pumps, piping design, heat exchange materials, rack systems, monitoring software, energy management platforms and operations services. If these links can form standards and talent development mechanisms early through industry-academia collaboration, Taiwan will not only benefit from growing AI server shipments but can also move into higher-margin system integration and solutions markets.

From an Asia-Pacific perspective, companies in the region are accelerating their adoption of generative AI, but most countries still have gaps in high-density data centers, thermal design and specialized operations talent. If Taiwan can use international partnerships like this to build replicable models for talent development and technology validation, it has a chance to become a key pillar of the Asia-Pacific AI infrastructure upgrade. This is not just news about a university working with companies. It is an early skirmish in the redivision of the regional tech map.

Industry Impact: From a Chip Race to an Infrastructure Race, Taiwan Cannot Sit Out

Key Takeaway: The center of competition in the AI industry is extending from models and chips to power, cooling, data center design and talent supply. Whoever can provide stable, energy-efficient and scalable infrastructure has the best chance of capturing the next wave of AI investment.

Over the past two years, market attention has mostly focused on GPUs, AI models and cloud services, but in reality the true barrier in the AI industry is quickly shifting to infrastructure. Even the most powerful model needs a stable server platform, a controllable cooling system and a data center architecture that can handle high power consumption. This makes liquid cooling and green computing no longer just engineering issues, but investment questions that draw attention at the board level.

For companies, building liquid cooling capabilities early could create advantages in several markets:

  • AI data center construction: Continued growth in high-density computing demand is driving adoption of new cooling architectures.
  • Enterprise private cloud and edge AI: Cooling efficiency matters even more in settings with limited space and energy.
  • ESG and regulatory pressure: Companies need to reduce energy waste and carbon emission risks, and liquid cooling can be one concrete solution.
  • Global supply chain restructuring: Regions with integrated design and rapid trial production capabilities will be favored by international customers.

Taiwan's advantages in this race are its complete electronics manufacturing base, mature ICT supply chain and highly flexible network of small and medium-sized enterprises. The challenge is that high-end system design, cross-disciplinary R&D and international talent strategy still need to be strengthened faster. By stepping in from the academic side, NCU fills the piece most easily overlooked in industrial upgrading: long-term talent and forward-looking research.

If partnerships like this can expand into joint labs, co-developed courses, corporate internships and international research projects, they will help Taiwan move from "being able to make the hardware" to "defining the next generation of AI infrastructure solutions." Rising to that level is the real key to pulling ahead internationally.

Q&A: Frequently Asked Questions

Q1: What is the most important practical significance of NCU's partnership trip to the US?

The most direct significance is bringing together three threads: AI liquid cooling, green computing and talent development. This is not just a technology exchange. It lays the groundwork for future joint R&D, course collaboration, experimental validation and local talent supply.

Q2: Why do AI data centers especially need liquid cooling?

Because the power consumption and heat density of AI servers keep rising, and traditional air cooling tends to hit efficiency and space limits in high-density deployments. Liquid cooling removes heat more effectively, improves compute stability and reduces energy waste.

Q3: What concrete opportunities does this create for Taiwan's tech industry?

Taiwan could expand from AI server manufacturing into cooling solutions, system integration, data center technical services and talent export. Combined with the Taoyuan R&D center and the local supply chain, it could also become a key showcase market for Asia-Pacific AI infrastructure.

Overall, National Central University's US trip highlights that competition in the AI industry has officially entered the "deep water of infrastructure." As the market shifts from comparing model parameters to comparing data center efficiency, cooling capability and depth of talent, Taiwan has a chance to capture a higher-value position in the next wave of AI build-out if it can combine university research, corporate implementation and the strengths of its local supply chain. Over the next few years, liquid cooling, green computing and cross-border industry-academia collaboration will no longer be supporting topics. They will become important variables in reshaping Asia's tech industry map.

News Source

(https://www.taiwanhot.net/news/1141959/中央大學拓展國際產學合作+蕭述三校長率團訪美聚焦AI液冷關鍵技術)