Computing Power as New Infrastructure: Trends, Challenges, and Future Outlook

In the next 5‑10 years, computing power will become a ubiquitous "new energy" similar to water, electricity, gas, and gasoline, turning into an essential civil utility. The systems and facilities that provide and maintain this capability are referred to as the new type of infrastructure.

The term "computing fee" (算费) is introduced to describe the cost of processing 1 GB or 10 GB of data, analogous to data traffic fees today. Computing power, largely delivered through cloud computing platforms such as Alibaba Cloud, Tencent Cloud, Kingsoft Cloud, and Tianyi Cloud, is considered part of the new generation of information infrastructure.

China’s "new infrastructure" (新基建) plan emphasizes upgrading traditional physical facilities to digital, intelligent ones and developing entirely new digital infrastructures. While Western countries have been advancing similar initiatives earlier, China’s investment in infrastructure has plateaued, making the development of new infrastructure a strategic necessity.

Three pathways create new infrastructure:

Upgrading existing facilities through technological evolution (e.g., 5G, fiber‑optic networks).

Applying ICT to retrofit legacy facilities for smart operation.

Introducing entirely new technologies such as AI and blockchain infrastructure.

Computing power will increasingly be delivered via cloud‑based SaaS and PaaS models. The evolution of cloud computing over the past two decades—from VMware virtualization (1998) to SaaS, IaaS, Docker, CI/CD, cloud‑native, and distributed cloud—demonstrates a mature service ecosystem that is now moving toward full infrastructure‑level provisioning.

Future market forecasts suggest a shift from public cloud dominance to stronger demand for private and hybrid clouds, as well as intelligent cloud services (e.g., SASE, AutoML, democratized AI).

Key challenges identified include:

Material bottlenecks: Moore’s Law slowdown and the approaching physical limits of silicon (2 nm) raise concerns about continued performance gains. Alternative materials such as carbon‑based transistors or DNA computing are being explored.

Data‑center cooling: Massive server farms face escalating energy consumption for cooling. While wind‑cooling is common, liquid‑cooling solutions can achieve PUE values as low as 1.03‑1.05, dramatically reducing power usage compared with traditional wind‑cooling (PUE 1.3‑1.5). However, liquid‑cooling adoption is hindered by lack of standards, high retrofitting costs, and corrosion risks.

Network‑security computing: The rapid expansion of IPv6, IoT, 5G, and big data intensifies security demands. China’s cloud environment experiences high rates of DDoS attacks, backdoor insertions, and malware infections, requiring substantial security‑computing resources. Effective protection now depends on advanced security‑computing platforms capable of handling massive, real‑time data streams.

Security‑computing is likened to “food for the army” – a prerequisite for reliable infrastructure. Integrating security into the computing‑power stack (PaaS security, SaaS security, data security, endpoint security) is essential for a resilient future.

In conclusion, the convergence of cloud computing, AI, big data, and open‑source ecosystems will usher in a democratized era of computing power, where the infrastructure becomes universally accessible, secure, and energy‑efficient.