祝融说。

We are living in an unprecedented era.

It is an era where order and chaos coexist, where opportunity and risk are intertwined. The slightest tremor in the global supply chain can trigger an industrial tsunami; every move in geopolitical strategic interplay can redraw the map of the world economy; a wave of information on social media can topple a brand's century-old reputation, or even destabilize an entire region. The certainty we once relied upon is being diluted by unprecedented uncertainty.

To borrow a concept from physics, the "entropy" of the world we inhabit is increasing dramatically. Entropy is a measure of a system's disorder. Increasing entropy means a rise in disorder, randomness, and unpredictability. From the macro level of international relations, to the meso level of industrial ecosystems, to the micro level of individual lives, we can all feel this powerful current of increasing entropy. Information overload makes it hard to distinguish truth from falsehood; technological iteration induces a fear of obsolescence; sudden "black swans" and lurking "gray rhinos" make every decision-maker tread carefully.

In such an era of "great entropy increase," what is the core competitiveness of a nation, a city, or an enterprise?

It is not resources, not population, and not even technology itself. Core competitiveness is the ability to resist entropy increase. This ability, we call "composure."

Composure is not the stubbornness of clinging to the old ways, nor is it blind optimism in the face of storms. Composure is the ability to see the essence and grasp the mainstream amidst complexity; it is the ability to extract effective signals and make precise judgments from a vast sea of redundant information; it is the ability to simulate various possibilities and choose the optimal path when facing a variable future. Composure, in essence, is a powerful ability to "converge" -- to converge infinite uncertainty into finite certainty, to converge chaotic reality into orderly action.

It should be noted that "convergence" is not a rhetorical device borrowed for this book; it comes directly from the theoretical framework of Process Realism: Observational Convergence and the Generation of Certainty (RC): all certainty is a provisional locking of observation onto the substrate of possibility, and causality and order are the stable paths of observational convergence. This book positions the construction and governance of intelligent computing centers as a governance practice of multi-scale resource order -- at the macro scale, "Channeling Computing from East to West" is an ordering of energy and computing resources across regions; at the meso scale of scheduling governance, "available margin" and "time-shared leasing" correspond to RC's practical imperative to "keep the right to choose present" and to fault-tolerant buffering, while "computing-power hegemony" and "evaluation alienation" are taken directly from its analysis of power gradation in the theory of order; at the micro scale, model evaluation and content governance rebuild a publicly negotiated evaluation ecology at the scale of computing. In other words, the intelligent computing center can be seen as an amplifier of "observational convergence" at the level of infrastructure. Readers interested in the philosophical foundations may trace back through this entry point.

So, in the 21st century, what is the underlying tool that supports this "composure"?

The answer is: computing power.

Computing power is the "anti-entropy weapon" of the digital age. Through large-scale, high-speed computation, it transforms disorderly data into orderly knowledge, reduces chaotic phenomena to clear patterns, and projects the fog of the future into a sandbox for decision-making. Whoever masters stronger computing power can complete the process of "convergence" faster and more accurately, thereby stabilizing their position in the torrent of entropy increase and gaining precious strategic initiative.

Therefore, we put forward the core thesis of this book: Computing power is composure.

Building a national intelligent computing center is by no means merely a technical project or commercial investment. It is about forging a "pillar of stability" for our nation and our people to deal with all future uncertainties.

Chapter 1: The New "Water, Electricity, and Coal" of Our Era -- Reimagining National Infrastructure

For a long time, we have been accustomed to using "water, electricity, and coal" to describe a nation's most fundamental and core public resources. Water is the source of life; electricity is the lifeblood of industry; coal (and its derived energy) is the fuel that drives modern society. Together, they form the cornerstone of industrial civilization and support everything we are familiar with. A nation's strength is first reflected in its ability to mobilize these fundamental resources.

Thus, we have the world-renowned South-to-North Water Diversion Project. It spans thousands of miles, transferring water from the south to the arid north, solving the "lifeline" problem of northern cities and supporting the sustained development of China's economic center of gravity. We have the West-to-East Gas Pipeline, which, like a giant dragon, transports natural gas from the west to the eastern coast, injecting a continuous stream of "blood" and "fuel" into the industrially developed regions. These great projects are, in essence, grand deployments of fundamental physical-world resources, and they have shaped China's economic geography over the past few decades.

Today, as we enter the digital economy era, a new, equally crucial fundamental resource has emerged: computing power.

If data is the "oil" of the digital economy, then computing power is the "internal combustion engine" that drives data to create value. Without computing power, data is nothing but a pile of meaningless zeros and ones sitting quietly on a hard drive. Only through powerful computing for processing, analysis, and modeling can data be refined into insight, wisdom, and decisions, truly empowering all industries.

Therefore, we must understand the strategic significance of the national intelligent computing center through a new imagination that transcends traditional physical infrastructure.

The intelligent computing center is the "South-to-North Water Diversion" and "West-to-East Gas Pipeline" of the digital age.

The grand blueprint of the "Eastern Data, Western Computing" national project is a vivid embodiment of this concept. The eastern region has a developed economy, rich application scenarios, and massive data production, but faces tight land and energy resources; the western region has abundant energy, especially wind and solar green energy, a suitable climate, and natural advantages for building large data centers. The essence of "Eastern Data, Western Computing" is a nationwide "computing power South-to-North Water Diversion." What it deploys is no longer the water molecules of the physical world, but the computing resources of the digital world. What it solves is the contradiction between the growing "computing power thirst" of the eastern region and its energy constraints.

Once this "digital canal" is built, the wind and sunlight of the west will be transformed into surging computing power through intelligent computing centers, serving the eastern region's financial transactions, urban governance, autonomous driving, and industrial design in real time. This not only optimizes the overall efficiency of national energy utilization but, more importantly, builds a unified, flexibly dispatchable "computing power network" across the country, just like the power grid and water network we have already built.

However, there is a fundamental difference between intelligent computing centers and traditional "water, electricity, and coal" infrastructure. What flows through water pipes is water, and what travels through power grids is electricity. They are passive infrastructure, whose value lies in stable "transmission." But the intelligent computing center is active, generative infrastructure. It is not just a "pipeline"; it is also a "factory."

A hydroelectric plant only produces electricity. It does not think about how to make light bulbs brighter, nor does it create new ways of using electricity. But an intelligent computing center, while providing computing power, is itself an innovation incubator. The large models running inside it can generate new code, new designs, new drug molecular structures, and new business strategies. It is not simply "transmitting" value; it is continuously "creating" value.

This requires us, as managers, to completely change our mindset. We can no longer manage an intelligent computing center with the same thinking used to manage a substation or a water plant. The core indicators for managing a substation are safety, stability, and low loss. But for an intelligent computing center, in addition to these basic requirements, the more core indicator is "empowerment efficiency" -- how many new applications are spawned by each kilowatt-hour of computing power? How many tech enterprises are incubated? How many social governance problems are solved? How much GDP is created?

If we merely see the intelligent computing center as a "digital real estate developer" that rents out racks and servers, content with collecting "rent" and "electricity price differentials," then we are the "sub-landlords" of computing power in the new era. This would be not only a tremendous waste of the nation's massive investment, but also a severe missed opportunity for an era.

The first purpose of this book is to help managers break free from the traditional "water, electricity, and coal" mindset and truly understand the revolutionary significance of the intelligent computing center as the core engine of "new quality productive forces." It is not a fixed asset on a balance sheet; it is the innovative momentum flowing through a city's veins, a crucial piece on the chessboard of national competitiveness.

Chapter 2: The Core Logic -- Using Convergence to Win the Future in an Entropic World

Having understood that the intelligent computing center is the new "water, electricity, and coal" of our era, we must answer a deeper question: Why do we need it so urgently? What fundamental contradiction of our time does it actually solve?

The answer lies in the concept introduced at the beginning of this introduction -- resisting entropy increase.

Let us translate this somewhat arcane physics concept into the everyday challenges managers face:

• The entropy of information overload: A city mayor might need to process millions of characters of reports, data, public opinion, and meeting minutes every day. The great majority of this is redundant information, but it may hide a few critical signals, such as the precursor to an infectious disease outbreak in a certain area, or a systemic financial risk in a certain industry. How to quickly and accurately find the decisive needle in a vast ocean of information? This is a classic "entropy reduction" problem.
• The entropy of decision-making complexity: When formulating a city traffic optimization policy, the variables to consider are vast. Bus routes, subway frequencies, traffic light timings, road widths, tidal population flows, weather conditions, sudden accidents -- these variables interact with each other, and a single move affects the whole system. Any decision based on intuition or limited experience can lead to the awkward situation of "solving one problem while creating another." How to find the global optimal solution in a multi-variable, tightly coupled complex system? This is also an "entropy reduction" problem.
• The entropy of future unpredictability: A sudden global pandemic can instantly disrupt all plans. A leading tech company being placed on an entity list can risk breaking the entire industry chain. An extreme rainstorm can completely paralyze a city's transportation. We cannot predict when a "black swan" will appear, but can we have greater resilience and ability to respond when it does? How to transform unpredictable shocks into manageable risks? This is yet another "entropy reduction" problem.

In the past, we relied on the wisdom of elites, accumulated experience, and hierarchical organizations to resist entropy increase. The brain of an experienced veteran mayor was itself an "entropy reduction machine," capable of filtering out noise by instinct and grasping the crux of the problem. An efficient bureaucracy, through standardized processes, ensured the orderliness of policy execution.

But today, these traditional methods are failing. Because the rate of entropy increase we face has far exceeded the processing capacity of the human brain and traditional organizations.

This is the historical inevitability of the intelligent computing center's emergence. It is a "convergence engine" we have specifically built to combat the era of great entropy increase.

"Convergence" is the key to understanding the core value of the intelligent computing center. Its workflow is essentially a continuous process of convergence:

1. Input divergence: The input end of the intelligent computing center is extremely divergent and high-entropy. It connects to millions of cameras and sensors across the city, the ever-changing public opinion data on the internet, real-time transaction data from financial markets, and cloud images from meteorological satellites. Before being processed, this data is chaotic, raw, and full of uncertainty. It is the rough material of the real world.
2. Computational processing: Computing power is the motor of this engine. It drives various complex algorithms and neural network models to clean, label, correlate, learn from, and reason about this raw data. This process is like a precise alchemical furnace, using immense energy (electricity and computation) to refine pure gold from ore. Every matrix operation, every parameter update, is about eliminating uncertainty in the data and increasing its information content.
3. Output convergence: The output end of the intelligent computing center is highly convergent and low-entropy. It no longer outputs messy data, but specific, executable "products of certainty."
   • It converges the city's chaotic traffic flow data into an optimal set of traffic light timing plans.
   • It converges vast numbers of medical images into a precise identification of early lesions.
   • It converges complex power grid load data into a minute-by-minute power dispatch command.
   • It converges thousands of research papers into a potential new drug development pathway.

Whoever can compute faster and more accurately can stabilize the situation and seize the initiative. Behind this statement lies a profound strategic logic.

"Computing faster" means gaining an advantage in time. While competitors are still holding meetings, doing research, and making decisions through traditional methods, you have already simulated tens of thousands of possibilities through computing power and found the optimal solution. In the fast-paced financial market, being 0.01 seconds faster means the difference between winning and losing. In the golden 72 hours of disaster relief, planning a rescue route one minute faster could save one more life. This speed, brought by computing power, grants us the "initiative" in responding to crises.

"Computing more accurately" means gaining an advantage in quality. It reduces the "cost of trial and error" in decision-making. In the past, planning a subway line might take years of demonstration, only to find after construction that it did not match actual population flows, resulting in enormous sunk costs. Today, through intelligent computing centers, we can simulate the city's population movement and industrial changes for the next 20 years, planning a "more accurate" route. In healthcare, AI-assisted diagnosis has the potential to help physicians reduce missed and mistaken diagnoses (contingent on clinical validation and physician oversight), an important support for the quality of life. In manufacturing, more accurate yield prediction can save substantial costs and enhance the competitiveness of the entire national industrial system.

"Computing more accurately" is the ultimate optimization of resources and the precise avoidance of risks. It means we are moving from an era of "extensive" governance to an era of "precision" governance. In the past, our policies and investments were, to some extent, like "flood irrigation," hoping to cover all areas in need, but inevitably resulting in enormous waste and ineffective investment. Decisions based on powerful computing power are more like "precision drip irrigation," putting every resource where it is most needed, achieving maximum benefit at minimum cost. In the current context of slowing global economic growth and tightening resource constraints, the significance of this cannot be overstated.

When "speed" and "accuracy" combine, they form what we call "composure."

At the economic level, this composure manifests as industrial resilience in the face of global economic cycles. When external market demand weakens, can we use computing power to quickly analyze the weak links in the supply chain, precisely provide credit support to small and medium enterprises, and guide capital to the most promising specialized and sophisticated enterprises? When a critical technology is blocked, can we use the computing platform to coordinate the nation's research forces for joint research and breakthroughs, simulating and accelerating the R&D of new materials and processes? An economy with a powerful intelligent computing center is like a giant ship equipped with advanced navigation computers and dynamic stabilization systems, capable of maintaining course and moving steadily even in rough seas.

At the social governance level, this composure manifests as anti-fragility in responding to sudden public events. Anti-fragility means not just remaining unchanged in the face of impact (that is "robustness"), but becoming stronger and more orderly after experiencing chaos and pressure. When a torrential rainstorm hits, a "fragile" city system descends into paralysis; but a smart city with "anti-fragility" can, through its intelligent computing center, predict all flooding points across the city within minutes, dispatch drainage resources in real time, dynamically plan emergency rescue routes, and even afterward, by reviewing the full data of the event, optimize the design of the city's pipeline network, so it performs better the next time a rainstorm comes. Computing power transforms the city from a passive body enduring disasters into a living organism that can learn and evolve from crises.

This composure ultimately manifests at the highest level of national security and international strategic interplay as a strategic deterrent we call "cognitive sovereignty."

This is a crucial but often overlooked dimension. In the era of artificial intelligence, we must clearly recognize that algorithms and models are never neutral. They are products of specific data, specific values, and specific objectives. A general-purpose large model trained by a Western tech giant necessarily has its "worldview" based on Western data, Western context, and Western interests.

Imagine a terrifying scenario: if our nation's core infrastructure -- financial risk control systems, public opinion analysis systems, tactical command systems -- relied on externally supplied underlying decision models. This would mean that the way we see the world, the framework through which we analyze problems, and the basis on which we make decisions, are all being shaped by others without us knowing. The "reality" we see is the "reality" that others want us to see. The "optimal solution" we converge on through computation may be the "optimal solution" that benefits others the most.

This is the "Trojan horse" of the cognitive domain. Physical borders can be guarded by armies, but the cognitive frontier must be guarded by our own computing power and our own models.

Therefore, building an autonomous and controllable national intelligent computing center and training our own foundational large models have significance far beyond the commercial and technological domains. This is about building a national "moat" for the 21st century. We must not only possess our own "two bombs and one satellite," but also our own "computing power and large models." Because the latter determines our "right to survive" and "right to develop" in the digital world.

Whoever masters the most powerful "convergence engine" holds the power to define facts, interpret the world, and shape the future. This is the ultimate meaning of "computing power is composure." It concerns a nation's position in the future global landscape and a people's destiny in the evolution of civilization.

Chapter 3: This Book's Promise -- A Practical Red Book for Stewards

So far, we have moved from the infrastructure positioning of "water, electricity, and coal," through the era background of "resisting entropy increase," to the strategic height of "cognitive sovereignty." These grand narratives are necessary, as they help unify our thinking and clarify the historical coordinates and future significance of the work we are engaged in.

However, the mission of this book is not merely to provide a grand theoretical framework.

Because we deeply understand that our readers -- those sitting in the seats of NDRC directors, big data bureau chiefs, state-owned enterprise executives, and technology park leaders -- face not abstract philosophical debates in their daily work, but concrete, headache-inducing practical problems:

• How to approve the budget? When reporting to superiors, how can you clearly explain the fundamental difference between investing tens of billions in an intelligent computing center versus building an industrial park or a highway in the past? What is its return cycle and measurement standard?
• How to save on electricity? The intelligent computing center is notoriously power-hungry. Under the hard constraints of dual carbon goals and dual energy consumption controls, how do you balance the growth of computing power with green development? Is PUE really the only indicator? Are there overlooked pitfalls?
• How to allocate computing power? After the center is built, computing resources are limited. Do you prioritize the internet giants that can pay top dollar to quickly recoup costs? Or should you, like "targeted poverty alleviation," use "computing vouchers" to subsidize the struggling but promising tech startups? This art of "dividing the cake" directly determines the future of the local digital economy ecosystem.
• How to ensure security? How can data be made "usable but invisible" without leaving its domain? How do you keep core national data secure against increasingly rampant ransomware and APT attacks? When AI-generated content begins to confidently produce nonsense or even spread harmful information, where are the boundaries of regulation?
• How to produce results? This is the most central question. Within one term of office, how can the intelligent computing center, this "gold-devouring beast," quickly produce visible, tangible, and quantifiable results? Is it by the number of listed companies incubated? By the percentage point reduction in the city's traffic congestion index? Or by the improvement in the efficiency of government departments? How do we build a scientific assessment system that can clearly demonstrate its value to superiors and the public?

Each of these questions is extremely real and extremely thorny. They are the gaps between the grand blueprint and the harsh reality. Bridging these gaps is the core promise of this book.

This Book's Promise: No Empty Talk, Only How to Turn "Computing Power" into "Productivity" and "Competitiveness"

We will completely discard the technical jargon and academic models that leave practitioners confused. We will not delve into the mathematical principles of the Transformer architecture, nor will we get entangled in the details of different chip fabrication processes.

Instead, we will use the most straightforward language, the most vivid cases, and the most actionable methodologies to provide you with a must-have "practical red book" for your desk.

• In the "Build Well" section, we will tell you that site selection is not just an economic calculation of electricity and land prices; it is also a "thermodynamics of computing power" concerning water, wind, light, and heat. The choice of architecture is not just a technical debate, but a "political decision" concerning national supply chain security.
• In the "Manage Well" section, we will dissect the causes of "computing power hegemony" and provide specific strategies like "tide scheduling" and "time-sharing leasing" to ensure fair and efficient allocation of computing resources. We will also explore how to build a "model supermarket," upgrading from selling "flour" (bare computing power) to selling "bread" (model as a service), raising the intelligent computing center's position in the value chain.
• In the "Use Well" section, we will focus on the implementation of application scenarios. From how the "city brain" achieves the leap from "seeing" to "foreseeing," to how to design "computing vouchers" and "data sandboxes" to open the door to the AI era for small and medium enterprises.
• In the "Guard Well" section, we will confront the challenges of security and ethics head-on. How to regulate generative AI and hold our "narrative position"? How to design "human-in-the-loop" processes to ensure humans retain the final say in critical decisions?

More importantly, this book will provide you with a set of "core talking points" that can be directly used for reporting, decision-making, and internal briefings. When you need to apply for a budget, you can say: "What we are investing in is not servers, but this city's future 'right to develop' and 'right to define.'" When you face questions about idle computing power, you can explain: "This is not idleness; it is 'strategic redundancy' and a 'computing power trump card' for dealing with black swan events."

We firmly believe that managing a pioneering endeavor like the national intelligent computing center requires not just engineers and scientists, but a group of "stewards" who understand strategy, industry, and governance.

This book is written for these future stewards.

Our ultimate goal is to help you, and the institutions you lead, successfully transform the intelligent computing center from a huge cost center into a powerful value-creation center; from a passive service provider into an active ecosystem builder.

Let us together start this powerful engine called the "intelligent computing center," and calculate the most certain path to prosperity and stability for our cities and our nation through the uncertain future.

This is the era of composure, and computing power is the firmest reins in our hands.