AI Infrastructure Energy: Why Delta Electronics Is Expanding Beyond Power Supplies

Posted on by White Hsu - The Multinational IT Couch Potato

當 AI 開始耗電：台達電正從「節能公司」走向「能源基礎設施公司」

====

AI Infrastructure Energy: Why Delta Electronics Is Expanding Beyond Power Supplies
AI Infrastructure Energy: Why Delta Electronics Is Expanding Beyond Power Supplies

AI Infrastructure Energy: Why Delta Electronics Is Expanding Beyond Power Supplies

When discussing artificial intelligence, most of the spotlight tends to focus on semiconductors.

Companies like TSMC have become globally recognized for enabling advanced chips designed by firms such as NVIDIA and AMD.

However, as generative AI and large language models continue to scale, a new constraint is becoming increasingly visible:

AI infrastructure energy is becoming one of the most critical bottlenecks in the next phase of AI growth.

Compute performance continues to advance rapidly, but the ability to provide stable, scalable, and sustainable electricity is emerging as a strategic challenge.

This structural shift helps explain why Delta Electronics has been gaining global attention in recent years.

Delta is evolving from a traditional power supply manufacturer into something far more strategic:

an energy infrastructure company positioned at the core of AI infrastructure energy systems.

AI Infrastructure Energy Is Becoming the Next Bottleneck

For the past two decades, competition in the cloud computing industry has focused on:

  • CPU performance
  • storage scalability
  • network latency

Generative AI is changing the equation.

The rapid scaling of large language models and AI training clusters is dramatically increasing global electricity demand.

Some estimates suggest that next-generation AI data centers may require:

100MW – 500MW of electricity capacity

That is roughly equivalent to the power consumption of a mid-sized city.

As a result, hyperscalers are increasingly focused on AI infrastructure energy questions such as:

  • Can local grids support new AI facilities?
  • How can energy costs be stabilized over long investment cycles?
  • How can companies meet ESG and net-zero commitments?
  • How can electricity reliability be ensured for 24/7 workloads?

Energy infrastructure is becoming a core competitive factor in AI deployment.

From Saving Energy to Generating Energy
From Saving Energy to Generating Energy

From Saving Energy to Generating Energy

Historically, Delta Electronics built its core capabilities around improving energy efficiency:

  • switching power supplies
  • power converters
  • inverters
  • thermal management systems

In other words:

making electricity more efficient to use.

But the company is increasingly expanding upstream into energy generation technologies, including:

  • hydrogen fuel cells (SOFC)
  • hydrogen production via electrolysis (SOEC)
  • battery storage systems
  • microgrid energy architecture

This shift reflects a broader strategic transition:

from

Saving Energy

to

Generating Energy

and ultimately toward

Orchestrating Energy.

Delta appears to be positioning itself as a long-term player in AI infrastructure energy ecosystems.

Hydrogen as Long-Duration Energy Storage for AI Infrastructure Energy Systems

Hydrogen is gaining renewed attention because renewable energy sources are inherently intermittent.

Solar energy depends on sunlight.

Wind energy depends on weather patterns.

But AI data centers require stable electricity 24 hours a day.

Hydrogen can function as a form of:

long-duration energy storage.

For example:

solar energy generated during the day can be converted into hydrogen, which can later produce electricity at night using fuel cells.

Solid oxide fuel cells (SOFC) can achieve high efficiency levels, particularly when combined with heat recovery systems.

For AI infrastructure energy environments that demand continuous uptime, this combination becomes increasingly attractive.

Microgrids and the Rise of Self-Supplied Energy
Microgrids and the Rise of Self-Supplied Energy

Microgrids and the Rise of Self-Supplied Energy

Another emerging trend is that large technology companies are beginning to consider generating their own electricity.

In some regions, power grid expansion cannot keep pace with the growing energy demand of AI workloads.

As a result, hyperscalers are investing in:

off-grid microgrid systems.

Microgrids may integrate:

  • solar power generation
  • battery storage
  • hydrogen fuel systems
  • AI-driven energy management software

Delta’s long-term ambition appears to extend beyond selling individual components toward providing integrated AI infrastructure energy architecture.

This moves the company into the role of:

energy system integrator.

Internal Carbon Pricing as a Strategic Signal

Delta has implemented internal carbon pricing as part of its sustainability strategy.

This approach embeds the future cost of carbon emissions into present-day decision making.

Strategically, this creates several advantages:

  • early adaptation to potential carbon taxation policies
  • improved ESG competitiveness
  • increased differentiation in energy solutions
  • enhanced long-term cost predictability

As global enterprises evaluate supply chain emissions more closely,

energy infrastructure capabilities increasingly influence vendor selection decisions.

Delta's Strategic Evolution Across the Energy Stack
Delta’s Strategic Evolution Across the Energy Stack

Delta’s Strategic Evolution Across the Energy Stack

Delta’s transformation can be viewed across multiple layers:

Phase 1: Power Components

  • switching power supplies
  • industrial power modules
  • cooling components

Phase 2: Power Systems

  • data center power architecture
  • EV charging infrastructure
  • industrial automation power systems

Phase 3: Energy Systems

  • microgrid infrastructure
  • battery storage solutions
  • smart grid systems

Phase 4: AI Infrastructure Energy Platforms

  • hydrogen energy technologies
  • AI-optimized energy management
  • net-zero infrastructure solutions

This evolution reflects a shift from:

component supplier

to

infrastructure provider

within the AI infrastructure energy landscape.

Taiwan’s Expanding Role in AI Infrastructure Energy

From a broader industry perspective, Taiwanese companies are strengthening their presence across multiple layers of the AI infrastructure stack:

  • TSMC → compute infrastructure
  • Delta Electronics → AI infrastructure energy systems
  • Quanta Computer → AI server manufacturing
  • Wiwynn → hyperscale computing platforms
  • Asia Vital Components → cooling infrastructure

Taiwan’s role is expanding beyond semiconductor manufacturing.

It is increasingly becoming a key hub for AI infrastructure energy technologies.

Energy May Become the Next Strategic Layer of AI
Energy May Become the Next Strategic Layer of AI

Conclusion: Energy May Become the Next Strategic Layer of AI

The competitive landscape of AI is expanding beyond chips.

Electricity supply, cooling efficiency, and energy orchestration are becoming critical enablers of future AI growth.

If TSMC provides the compute foundation for AI,

Delta is positioning itself to provide the energy foundation.

Over the next decade,

companies capable of delivering stable, scalable, and sustainable AI infrastructure energy solutions may become some of the most strategically important players in the global technology ecosystem.

Delta Electronics appears determined to be one of them.

If you enjoyed this analysis

Subscribe to Taiwan Tech Intelligence on Substack for deeper research and premium insights on:

  • Taiwan’s IT world within the global ecosystems
  • Taiwan’s role in the global AI supply chain
  • Data-driven company and industry analysis
  • Platforms and Startups

👉 https://whitehsu.substack.com/

—

For ongoing shorter insights and industry updates, follow Taiwan Tech Dispatch on LinkedIn:

👉 https://www.linkedin.com/newsletters/taiwan-tech-dispatch-7355532535025586178

—

Explore more articles on

Further Readi

====

當 AI 開始耗電：台達電正從「節能公司」走向「能源基礎設施公司」
當 AI 開始耗電：台達電正從「節能公司」走向「能源基礎設施公司」

當 AI 開始耗電：台達電正從「節能公司」走向「能源基礎設施公司」

在討論 AI 時，大多數焦點都集中在晶片。

例如 TSMC 如何透過先進製程支撐 NVIDIA、AMD 等公司所設計的 AI 晶片。

然而，隨著生成式 AI 與大型模型（LLM）的快速發展，一個新的瓶頸正在浮現：

AI 的真正限制，不再只是算力，而是電力。

而這正是 Delta Electronics（台達電）近年逐漸走上全球舞台中央的原因。

台達電的角色，正在從「電源供應器製造商」，轉變為「能源基礎設施提供者」。

甚至更進一步：

成為 AI 時代的 Energy Platform company。

AI 時代的新瓶頸：Power Density 與 Energy Stability

過去 20 年，雲端產業競爭的焦點主要在：

  • CPU performance
  • storage scalability
  • network latency

但生成式 AI 帶來的變化是：

資料中心的電力需求呈現指數型成長。

新一代 AI data center 的耗電量，可能達到：

100MW – 500MW

這相當於一座中型城市的用電量。

因此 hyperscaler 開始關注的問題包括：

  • 電網容量是否足夠
  • 是否能取得穩定電力
  • 是否能降低能源成本
  • 如何達成 ESG 與淨零排放目標

也因此，電源與能源管理能力的重要性正在快速提升。

從 Saving Energy 到 Generating Energy
從 Saving Energy 到 Generating Energy

從 Saving Energy 到 Generating Energy

台達電長期以來的核心能力，在於提高電力使用效率：

  • switching power supply
  • inverter
  • power conversion
  • thermal management

也就是：

讓電更有效率地被使用。

然而近年來，台達電開始往能源供應端延伸，例如：

  • 氫燃料電池（SOFC）
  • 電解水製氫（SOEC）
  • 儲能系統
  • microgrid 微電網架構

這代表公司戰略的重大轉變：

從：

Saving Energy

走向：

Generating Energy

甚至進一步走向：

Orchestrating Energy

氫能：為 AI data center 提供長時間穩定能源

為什麼氫能在 AI 時代重新受到關注？

原因是再生能源的特性：

太陽能與風能雖然乾淨，但具有間歇性。

而 AI data center 的需求是：

24 小時穩定供電。

氫能可以作為：

長時間儲能（long-duration energy storage）

例如：

  • 白天利用太陽能製氫
  • 夜間透過燃料電池發電

固態氧化物燃料電池（SOFC）甚至可透過熱回收，提高整體能源效率。

這對於需要穩定供電的 AI infrastructure 具有吸引力。

Microgrid：企業開始建立自己的電廠
Microgrid：企業開始建立自己的電廠

Microgrid：企業開始建立自己的電廠

另一個重要趨勢是：

大型科技公司開始思考「自備電源」。

原因是：

部分地區的電網容量，已經無法快速支援 AI data center 的需求。

因此 hyperscaler 開始投資：

off-grid microgrid

也就是：

小型、自主控制的電力系統。

microgrid 可能整合：

  • 太陽能
  • 儲能電池
  • 氫能系統
  • 智慧電網管理軟體

台達電的角色，正是提供整套能源架構。

從單一 power supply module，升級為：

energy system integrator。

內部碳定價：建立未來競爭優勢

台達電長期推動內部碳定價（internal carbon pricing）。

這種策略的意義在於：

提前將碳排成本納入投資決策。

這帶來幾個優勢：

  • 提早適應未來碳稅制度
  • 提升能源解決方案的競爭力
  • 強化 ESG 客戶需求
  • 建立差異化能力

當企業開始重視 supply chain 的碳足跡時，

能源管理能力就成為競爭門檻。

台達電的轉型路徑
台達電的轉型路徑

台達電的轉型路徑

如果將台達電的發展分成幾個階段：

第一階段：Power Components

  • switching power supply
  • industrial power supply
  • cooling fan

第二階段：Power Systems

  • data center power architecture
  • EV charging infrastructure
  • industrial automation power

第三階段：Energy Systems

  • microgrid
  • battery storage
  • smart grid

第四階段：Energy Platform

  • hydrogen energy
  • AI-driven energy optimization
  • net-zero infrastructure

這代表公司正在從：

零件供應商

走向：

基礎設施供應商。

台灣正在形成 AI Infrastructure Cluster

如果從產業角度來看：

台灣企業正在 AI infrastructure 的多個關鍵層級建立優勢：

  • TSMC → compute infrastructure
  • Delta Electronics → energy infrastructure
  • Quanta Computer → AI server manufacturing
  • Wiwynn → hyperscale server platform
  • Asia Vital Components → cooling infrastructure

台灣的角色，不再只是 semiconductor island。

而是：

AI infrastructure hub。

能源，可能是下一個戰略制高點
能源，可能是下一個戰略制高點

結語：能源，可能是下一個戰略制高點

AI 的競爭，已經不只發生在晶片層。

而是延伸到：

電力、散熱與能源管理。

如果 TSMC 提供 AI 的算力基礎，

台達電正在嘗試提供 AI 的能源基礎。

在未來十年：

能夠穩定提供電力的企業，

可能會成為 AI 時代最關鍵的基礎設施供應者之一。

而台達電，顯然希望成為其中的一員。

如果你喜歡這篇分析

歡迎訂閱我的 Substack 專欄 Taiwan Tech Intelligence，獲取更深入的產業研究與觀察：

  • 台灣的IT以及高科技產業
  • 台灣在全球 AI 供應鏈中的關鍵角色
  • 以數據與案例為基礎的公司與產業分析
  • 平台以及新創生態

👉 https://whitehsu.substack.com/

—

也歡迎在 LinkedIn 上追蹤 Taiwan Tech Dispatch，掌握即時觀點與產業動態：

👉 https://www.linkedin.com/newsletters/taiwan-tech-dispatch-7355532535025586178

—

更多文章請見：
https://whitehsu.blog/

延伸閱讀

Leave a Reply