The Digital Economy Is Becoming an Energy Economy

For almost three decades, the digital economy was described as if it had escaped geography. Information travelled instantly, software crossed borders almost without friction, and businesses moved from offices to websites and then to the cloud. The vocabulary itself reinforced the illusion. We spoke about cyberspace, virtual infrastructure and cloud computing—as though economic activity had somehow detached itself from land, electricity and physical infrastructure. But the next phase of the digital economy is revealing almost the opposite. The more digital the world becomes, the more important physical geography may become. Behind every AI model, streaming service, financial transaction, e-commerce platform and cloud application stands an expanding industrial infrastructure of data centres, electricity networks, fibre cables, cooling systems, land and increasingly sophisticated power-management systems.

From Factory Geography to Computing Geography

The industrial revolution created its own economic map. Textile mills initially gravitated toward water and labour. Heavy industries clustered around coal, ports and railways. Petroleum later reorganised industrial geography around refineries, pipelines and shipping routes. Manufacturing globalisation created another map centred on container ports, industrial clusters, highways and efficient supply chains.

The AI and cloud era is creating yet another map: the geography of computation.

A data centre may appear very different from a steel plant, but economically the location problem is surprisingly familiar. Operators need reliable inputs, infrastructure, connectivity, security and regulatory certainty. The difference is that one of the most important raw materials is electricity and one of the most important transport networks is fibre.

This creates an important contradiction in the idea of a borderless digital economy. Data can cross the world in milliseconds, but the infrastructure processing that data must still exist somewhere.

And where that somewhere is located is becoming economically strategic.

Electricity Is Becoming the Industrial Raw Material of AI

The most important shift may be the convergence between digital competitiveness and energy competitiveness. Earlier generations of policymakers often treated telecommunications, electricity and industrial development as separate policy areas. That distinction is becoming increasingly artificial.

Large-scale cloud computing and particularly AI require enormous computational capacity. Computational capacity requires chips. Chips require servers. Servers require electricity. Electricity generates heat. Heat requires cooling. Cooling requires additional infrastructure and, depending on technology and climate, potentially significant water resources.

The digital value chain therefore eventually reaches something very physical:

AI → Compute → Chips → Data Centres → Electricity → Grid → Energy Infrastructure.

This means countries seeking leadership in artificial intelligence cannot think only about algorithms, startups and semiconductor access. They must think about transformers, transmission lines, generation capacity, storage, substations, fibre networks and land availability.

The next AI bottleneck may therefore emerge not inside a laboratory but inside an electricity grid.

The New Location Economics

Traditional industries often searched for cheap labour. Data-intensive industries increasingly search for something different: cheap, reliable and scalable electrons combined with fast connectivity.

The ideal data-centre location sits at the intersection of several economic systems—abundant power, resilient electricity networks, fibre connectivity, suitable land, manageable cooling conditions, physical security, political stability and predictable regulation.

Few locations possess all these advantages simultaneously.

A location may have cheap renewable electricity but poor international fibre connectivity. Another may possess excellent connectivity but face land constraints. A third may have abundant land but an unreliable electricity system. A fourth may have excellent infrastructure but increasingly restrictive environmental or water regulations.

This creates a new form of location competition.

Governments once competed to attract automobile factories, semiconductor fabs and export-processing zones. Increasingly, they will compete for hyperscale data centres, cloud regions, AI computing clusters and digital infrastructure investment.

But there is a critical difference.

A large manufacturing plant usually creates visible factory employment and supplier ecosystems. A highly automated data centre can involve enormous capital investment without generating comparable direct employment.

Governments therefore need to ask a harder question than simply how much investment is arriving:

How much domestic economic capability is being created around that investment?

Data Centres Are Not Automatically Development Centres

This distinction will become increasingly important for emerging economies.

Celebrating billions of dollars of announced data-centre investment can easily become another version of the old industrial-policy obsession with counting factories rather than measuring linkages.

The real development impact depends on what develops around the facility: domestic cloud services, cybersecurity capabilities, AI businesses, engineering services, renewable-energy systems, cooling technologies, equipment maintenance, fibre infrastructure, digital startups and skilled employment.

Without these linkages, a country risks becoming merely a landlord and electricity supplier to the global digital economy.

The strategic objective should therefore not be simply hosting data centres.

It should be building digital-industrial ecosystems around them.

Water May Become the Hidden Digital Constraint

Energy receives most of the attention, but cooling introduces another dimension to digital geography.

Data centres generate substantial heat. Different cooling technologies have different water requirements, but in water-stressed regions the interaction between digital infrastructure and local resource availability could become politically sensitive.

This creates an uncomfortable future possibility: communities, agriculture, industry and computing infrastructure may increasingly compete indirectly for the same underlying resources.

The location economics of data centres could consequently become partly a climate-economics problem.

Cooler regions, locations with abundant renewable electricity, advanced cooling technologies and strong water-management systems could gain advantages that were barely considered in earlier theories of digital competitiveness.

Climate geography may therefore influence digital geography.

Sovereign Data Will Reinforce the Map

Another force is pushing the digital economy toward physical geography: sovereignty.

Governments increasingly care about where sensitive data are stored, which legal jurisdiction governs them and who controls the infrastructure processing them. Financial data, government records, health systems, defence information and strategically important commercial datasets are unlikely to be treated simply as globally mobile commodities.

This may gradually produce regional or national computing ecosystems.

The global internet will remain interconnected, but the infrastructure beneath it could become increasingly territorial.

The result may resemble energy infrastructure more than the original open-internet ideal: interconnected globally, but organised around national security, regulation, resilience and strategic control.

The Data Centre Could Become the Factory of the Intelligence Economy

The industrial economy converted energy and raw materials into physical goods.

The emerging intelligence economy converts electricity, chips and data into computation.

That makes the data centre something more important than a warehouse filled with servers. It is increasingly becoming the factory floor of the intelligence economy.

This changes how economic geography should be understood.

Regions possessing abundant clean power, strong grids, fibre connectivity, political stability and digital skills could become the industrial centres of the AI age even if they were never major manufacturing centres.

Conversely, some traditional technology centres may discover that talent alone is no longer sufficient if electricity, land and infrastructure cannot expand with computational demand.

The competitive unit may gradually shift from the technology company to the technology-energy ecosystem.

India: From Data-Centre Capacity to Compute Competitiveness

For India, the opportunity is larger than simply attracting global cloud companies. The country has a huge digital population, expanding digital public infrastructure, rapidly growing renewable-energy capacity, a large technology workforce and rising demand for cloud and AI services.

But the strategic challenge is integration.

Data-centre policy cannot remain separated from power-sector reform, renewable generation, storage, grid modernisation, semiconductor strategy, fibre infrastructure, urban planning and skill development.

India should therefore think beyond data-centre parks toward compute corridors—locations where renewable power, transmission infrastructure, fibre networks, cloud infrastructure, AI computing capacity, universities, startups and technology services reinforce one another.

This could create a completely new form of industrial cluster.

Yesterday’s cluster might have contained factories, warehouses and suppliers.

Tomorrow’s cluster could contain data centres, renewable-energy plants, battery systems, AI companies, semiconductor-linked services, cybersecurity firms and research institutions.

The Next Economic Map Will Be Drawn by Electrons and Data

The deepest lesson is that technology has not abolished geography. It has redesigned it.

The nineteenth-century economic map was shaped by coal, rivers and railways. The twentieth century was shaped by oil, highways, electricity grids, airports and container ports. The early twenty-first century appeared to be shaped primarily by digital networks.

The next phase may combine all of them.

Power, fibre, chips, water, land, climate, regulation and data sovereignty will increasingly determine where computational capacity concentrates.

Countries that understand this convergence early may capture much more than data-centre investment. They could capture the infrastructure of the intelligence economy itself.

Countries that treat data centres merely as another real-estate category may discover something much later: the cloud was never really floating above the economy.

It was sitting on land, connected to fibre, consuming electricity—and quietly creating a new geography of global economic power.

#DataCentres #ArtificialIntelligence #DigitalEconomy #EconomicGeography #CloudComputing #Energy #DigitalInfrastructure #AIInfrastructure #India #FutureEconomy

When Economic Independence Moves from Oilfields to Algorithms

The strategic resource of the twentieth century was oil. The strategic infrastructure of the twenty-first century may be computation. For most of modern economic history, countries measured vulnerability through physical dependence. Who supplied the oil? Where did the food come from? Who controlled the shipping lanes? Could factories obtain steel, machinery and energy during a crisis? The digital economy has quietly changed the meaning of dependence. A country may possess ports, factories, banks, universities and even large foreign-exchange reserves, yet much of its economic system can still depend on semiconductors designed elsewhere, cloud infrastructure operated elsewhere, operating systems controlled elsewhere, payment architecture influenced elsewhere and increasingly artificial-intelligence models developed elsewhere. Economic sovereignty is therefore acquiring a new layer: digital sovereignty.

From the oil shock to the chip shock — strategic dependence has changed its address. The oil crises of the 1970s taught governments that efficiency without supply security could become dangerous. Countries subsequently created strategic petroleum reserves, diversified energy suppliers and invested in domestic energy capabilities. The semiconductor shortages following the pandemic delivered a similar lesson in a different form. A tiny component could delay automobiles, electronics, machinery and communications equipment across continents. The lesson was larger than chips themselves. Modern economies contain digital chokepoints that are difficult to see until they stop functioning. The difference is that oil dependence was relatively easy to measure in barrels. Digital dependence is distributed across intellectual property, fabrication equipment, processors, software, data centres, submarine cables, satellites, cybersecurity systems, cloud platforms and technical standards. The twenty-first-century strategic reserve may therefore include not only barrels of petroleum but computing capacity, trusted data infrastructure and technological knowledge.

The cloud is becoming the new industrial estate. Industrialisation once required land, electricity, roads, factories and ports. Increasingly, businesses also require cloud computing, cybersecurity, digital identities, data storage, APIs and AI services. When thousands of domestic enterprises depend on a small number of external technology platforms, those platforms cease to be ordinary commercial services. They begin to resemble economic infrastructure. This creates an uncomfortable question. If electricity grids, banking systems and telecommunications networks are considered strategically important, why should the computing infrastructure on which all three increasingly depend be viewed purely as another marketplace?

Yet digital sovereignty can easily become confused with digital isolation. Trying to reproduce every chip, operating system, cloud platform and AI model domestically would be enormously expensive and, for most countries, unrealistic. The objective cannot sensibly be technological autarky. The more intelligent objective is strategic optionality: no critical economic function should depend on a single external technological gatekeeper without credible alternatives, interoperability or contingency arrangements.

AI changes the sovereignty equation again. Earlier digital dependence centred largely on hardware and software. Artificial intelligence introduces dependence at the level of economic intelligence itself. Businesses may increasingly use AI systems to design products, write software, optimise factories, screen applicants, detect fraud, analyse markets and support scientific discovery. If the underlying models, computing infrastructure and training ecosystems are concentrated outside a country, an unusual economic hierarchy could emerge. Nations may generate enormous quantities of data while importing the intelligence required to extract economic value from that data.

This could reproduce an old development problem in a remarkably modern form. For centuries, poorer economies exported raw commodities while industrial economies captured higher margins through processing and manufacturing. In the emerging digital economy, data can become the raw material, computing the processing plant and AI the value-added layer. Countries that merely generate users and data while importing algorithms and computational capability could become the digital equivalent of commodity exporters.

The semiconductor race reveals the cost of sovereignty. Governments across major economies are supporting semiconductor capacity because markets optimised primarily for efficiency created highly concentrated supply chains. But semiconductor manufacturing demonstrates an important limitation of nationalist industrial policy: sovereignty cannot simply be purchased by announcing a fabrication plant. Advanced chips depend on extraordinary networks of specialised equipment, materials, design software, intellectual property, engineering talent and global suppliers. Building one factory does not automatically create technological independence.

The same principle applies to AI. Announcing a national AI model without sufficient computing infrastructure, electricity, research capability, high-quality datasets and commercial adoption can produce technological symbolism rather than technological power. Digital sovereignty therefore requires ecosystems rather than monuments.

The next trade barriers may be written in code rather than tariffs. Twentieth-century protectionism was visible at customs: tariffs, quotas and import licences. Digital protectionism can be far less visible. Data-localisation rules, cybersecurity certification, cloud-procurement conditions, AI standards, digital-service regulations, technical protocols and platform requirements can influence market access without a container ever crossing a border. Standards may consequently become instruments of industrial strategy.

This matters especially for smaller economies and MSMEs. A multinational corporation can maintain separate data systems and compliance teams for different jurisdictions. A small exporter cannot easily maintain multiple digital architectures merely to sell into different markets. If digital sovereignty evolves into incompatible national technology systems, the world could replace one global internet economy with several partially connected digital economic zones. The compliance cost of this fragmentation could become the new non-tariff barrier.

Payment sovereignty may become as important as monetary sovereignty. Money itself is becoming digital infrastructure. Instant-payment networks, digital wallets, central-bank digital currencies, stablecoins and cross-border payment systems are increasingly intertwined with commerce. Historically, states defended the authority to issue currency. In the future, they may also worry about who owns the technological rails through which that currency moves. A country can theoretically possess monetary sovereignty while becoming technologically dependent on external payment infrastructure. That distinction will become increasingly important as physical cash becomes less central to everyday transactions.

India’s opportunity is not to copy every technology but to build strategic layers. India’s scale gives it unusual advantages: a huge domestic market, deep software capabilities, expanding digital public infrastructure, a large entrepreneurial base and growing demand for data centres, electronics and AI. But scale alone does not guarantee sovereignty. The critical question is where India wants to possess capability, where it can rely on trusted international partnerships and where dependence represents an unacceptable systemic risk.

This requires moving beyond the simplistic slogan of making everything domestically. A more sophisticated strategy would distinguish between technologies India must control, technologies it must understand, technologies it must manufacture competitively and technologies it can safely source internationally from diversified partners. Those are four very different policy objectives.

The challenge also extends far beyond large technology companies. Millions of MSMEs are moving onto digital accounting systems, marketplaces, cloud applications, logistics platforms and AI tools. Their digital transformation could improve productivity enormously, but it could simultaneously create new forms of platform dependence. Digital sovereignty therefore cannot remain an elite discussion about supercomputers and semiconductor fabs. It must eventually address whether ordinary enterprises retain portability of data, interoperability of systems, competitive choice and affordable access to digital infrastructure.

The paradox of the coming decade is that greater sovereignty will require deeper cooperation. No serious technological economy can independently reproduce the entire digital stack. Semiconductor supply chains alone demonstrate the depth of international specialisation. Countries seeking absolute independence could therefore make themselves poorer without necessarily becoming safer. The emerging strategy will instead be selective interdependence: domestic capability in critical layers combined with diversified international partnerships elsewhere.

This may produce a new geography of globalisation. Countries could increasingly choose technology partners according not only to price but also to security, legal compatibility, geopolitical trust and continuity of supply. The cheapest supplier may no longer automatically be considered the economically safest supplier. Efficiency will remain important, but resilience will acquire a price.

The ultimate sovereignty question is not where the server stands, but who retains the power to choose. A domestically located data centre running entirely foreign technology does not necessarily constitute sovereignty. Nor does a nationally branded AI model dependent on imported processors automatically create technological independence. Genuine digital sovereignty lies deeper: technical capability, competitive alternatives, skilled people, interoperable infrastructure, cybersecurity, institutional capacity and the ability to continue essential economic functions when external relationships are disrupted.

The world spent much of the twentieth century building strategic petroleum reserves because governments understood that economies could not function without energy. The twenty-first century may require a broader idea of strategic reserves: compute capacity, semiconductor access, trusted clouds, secure communications, resilient payment systems, critical datasets and human technological capability.

The countries that understand this early may gain more than technological independence. They may gain negotiating power.

And this is where the digital sovereignty race becomes fundamentally different from the technological races of the past. The competition is no longer simply about who invents the next technology. It is increasingly about who controls the infrastructure through which everyone else must use it.

In the industrial age, power belonged disproportionately to those who controlled oilfields, factories, ports and shipping routes. In the digital age, power may increasingly belong to those who control chips, compute, clouds, standards, networks and algorithms.

Economic sovereignty is not disappearing. It is migrating—from territory to technology, from barrels to bandwidth, and increasingly from machines to intelligence.

#DigitalSovereignty #ArtificialIntelligence #Semiconductors #DigitalEconomy #EconomicSovereignty #CloudComputing #Technology #India #MSME #FutureEconomy

When the Most Valuable Assets Can No Longer Be Touched

The Factory Is Still There, but Value Has Moved Somewhere Else

For most of economic history, wealth was reassuringly visible. Agricultural power could be measured in land and harvests. Industrial power appeared as coal mines, steel plants, machines, warehouses, railways and factories. A company’s strength could almost be photographed. The balance sheet reflected this physical world: land had a price, machinery had a replacement cost, inventories could be counted and buildings could be offered to a bank as collateral.

The twenty-first-century economy is quietly breaking this relationship between what can be seen and what creates value.

Increasingly, the valuable part of a business may not be its factory, office or inventory. It may be software controlling production, an algorithm predicting customer behaviour, a database accumulated over years, a patented process, a trusted brand, a product design, a distribution system or simply the organisational knowledge that allows thousands of people to work together efficiently.

We are moving from an economy of owning things to an economy of knowing things.

From Land to Machines to Knowledge

Every major economic transformation has changed the meaning of capital. In an agrarian economy, land was the decisive asset. The Industrial Revolution shifted power toward machinery, factories and infrastructure. The twentieth century added mass production, managerial systems and global brands.

The digital age is producing another transition.

A modern company can become enormously valuable without owning proportionately enormous physical assets. Software can be reproduced millions of times at negligible additional cost. A successful design can travel internationally without a container ship. An algorithm developed in one location can influence transactions across dozens of countries almost instantly. A brand can command a premium even when competing products emerge from remarkably similar manufacturing systems.

This creates an unusual economic paradox: the economy is becoming more valuable while parts of its productive capital are becoming less visible.

And that invisibility is not merely an accounting curiosity. It challenges some of the basic institutions through which capitalism has traditionally operated.

The Balance Sheet May Be Looking at Yesterday’s Economy

Industrial accounting was designed for an industrial world.

Buy a machine and accountants recognise an asset. Construct a factory and investment appears clearly. Spend heavily building organisational capability, training employees, developing proprietary databases, experimenting with software or strengthening a brand, and significant portions of that expenditure may be treated differently.

This distinction becomes increasingly problematic when competitive advantage itself moves toward intangible capability.

Two companies may possess factories of roughly comparable physical quality yet produce dramatically different economic results because one possesses better software, stronger design capability, deeper supplier knowledge, superior data and a more trusted brand.

The machines may look similar.

The productive intelligence surrounding the machines is not.

Therefore, future industrial policy cannot simply ask: How much manufacturing capacity has been created?

It must increasingly ask: How much knowledge has been embedded inside that capacity?

The New Geography of Trade May Be Invisible

Traditional international trade statistics were built around goods crossing borders. A container carrying garments from India to the United States is relatively easy to record. Its origin, destination and declared value can be identified.

But imagine a product designed in Italy, engineered through software developed in India, manufactured in Vietnam, marketed through an American digital platform, supported by cloud infrastructure elsewhere and sold under intellectual property registered in another jurisdiction.

Where exactly was the value created?

The physical product crosses one border. Its economic intelligence may have crossed several.

This could become one of the defining problems of international economics. Countries that dominate manufacturing volumes may not necessarily capture the largest share of value. Countries controlling design, technology, standards, software, brands, platforms and intellectual property can potentially capture substantial margins without undertaking the majority of physical production.

The old debate was about who manufactures the product.

The emerging debate will be about who owns the intelligence inside the product.

Data Is Becoming Capital—but Strange Capital

Data illustrates why conventional economic categories are becoming uncomfortable.

A machine deteriorates when used. Data can become more useful when repeatedly analysed and combined with other information. A warehouse occupies a specific geographical location. A database can be accessed simultaneously across borders. A physical asset can normally be sold to another owner. The commercial value of data often depends heavily on context, scale, permissions and the analytical systems surrounding it.

Algorithms create similar complications.

Their economic value may not lie in the lines of code themselves but in accumulated learning, proprietary datasets, continuous experimentation and integration into business processes.

This means the future corporation may increasingly resemble an institutional memory system rather than simply a collection of physical assets.

Banking Could Face an Intangible Collateral Problem

There is another consequence that deserves far more attention.

Banks understand buildings, land and machinery because these assets can usually be valued and pledged as collateral. But what happens when the most productive assets of a small enterprise are software, designs, customer relationships, technical knowledge or intellectual property?

A technologically sophisticated company can therefore become economically valuable but financially difficult to understand through conventional lending models.

This could produce a strange financing divide.

Asset-heavy businesses may continue receiving traditional credit because lenders can see their collateral, while knowledge-intensive businesses depend increasingly on equity, venture finance, specialised lenders or cash flows.

The financial system could therefore underestimate precisely the enterprises that represent the emerging economy.

For MSMEs, this question becomes particularly important. The next generation of cluster development cannot stop at common facilities, industrial estates and machinery upgrading. Clusters will increasingly require shared design intelligence, testing knowledge, digital systems, databases, branding capability, intellectual-property support and mechanisms for converting knowledge into finance.

The future cluster may contain fewer common machines and more common intelligence.

Taxation Will Chase Value That Has Become Mobile

Factories are difficult to move overnight. Intellectual property is considerably more mobile.

As corporate value becomes increasingly intangible, governments face a fundamental taxation challenge: economic activity can occur in one country, customers can reside in another, intellectual property can be located elsewhere and profits can potentially be attributed through complex corporate structures.

The international tax debate is therefore not merely about tax rates. It reflects a deeper problem: the geographical location of economic value itself is becoming harder to define.

The industrial economy tied companies to places.

The intangible economy partially separates value from geography.

Governments will spend the coming decades trying to reconnect the two.

Productivity May Also Be Misread

There is an even more uncomfortable possibility: parts of the productivity puzzle may reflect measurement systems struggling to capture organisational transformation.

Installing an expensive machine is visible investment. Redesigning an entire production system around data, artificial intelligence, worker knowledge and supply-chain coordination can be much harder to quantify.

Yet the second transformation may eventually matter more.

Artificial intelligence could accelerate this problem dramatically. AI will increasingly become embedded not only in products but in organisational decision-making—procurement, quality control, forecasting, design, maintenance, logistics, marketing and management.

A factory may therefore become significantly more productive without looking radically different from outside.

The intelligence layer will change before the concrete layer does.

The Great Divide May Become Tangible versus Intangible

The most important future inequality may not simply be between manufacturing and services or between developed and developing economies.

It may emerge between economies that produce physical goods and economies that own the knowledge governing those goods.

A country can manufacture millions of products and still capture limited margins if foreign companies control the technology, product architecture, branding, standards, distribution platforms and customer relationships.

This is particularly important for emerging economies.

Industrialisation remains essential because physical production creates employment, supplier networks, engineering capabilities and technological learning. But manufacturing without intangible accumulation can become a trap. Countries may build factories while remaining permanently dependent on somebody else’s technology, brands and market access.

The next stage of development therefore cannot simply be Make in India—or its equivalent elsewhere.

It must increasingly become Know in India, Design in India, Patent in India, Brand from India and Make in India.

The Future Factory Will Have Two Floors

One floor will remain physical: machines, workers, materials, energy and logistics.

The other will be invisible: software, algorithms, designs, patents, databases, standards, brands, organisational routines and accumulated knowledge.

The competitiveness of the first floor will increasingly depend upon the sophistication of the second.

This changes the meaning of industrial strategy. Governments have historically competed through infrastructure, industrial land, electricity, tax incentives and capital subsidies. These will remain important. But future competitiveness will depend increasingly on research ecosystems, universities, technical talent, intellectual-property institutions, digital infrastructure, design capabilities, standards, data governance and the ability of firms to convert knowledge into commercially scalable assets.

Countries building factories without simultaneously building intangible capital may discover that they have constructed the hardware of industrialisation while importing its operating system.

The Invisible Economy Will Require New Economic Eyes

The central problem of the intangible economy is therefore not that economic value is disappearing.

It is that value is becoming harder to see.

Our statistics still largely observe yesterday’s economy. Banks often lend against yesterday’s definition of assets. Trade systems were designed around yesterday’s borders. Tax systems search for yesterday’s geographical connection between production and profit. Industrial policies frequently subsidise yesterday’s forms of capital.

But tomorrow’s competitive advantage may increasingly sit inside code, designs, databases, organisational routines and human knowledge.

The great economic race of the coming decades may consequently not be about who owns the largest number of factories.

It may be about who owns the invisible architecture that tells those factories what to produce, how to produce it, how to improve it—and who ultimately captures the value.

That is the real arrival of the intangible economy.

#IntangibleEconomy #FutureOfBusiness #KnowledgeEconomy #ArtificialIntelligence #Innovation #IntellectualProperty #Manufacturing #MSME #DigitalEconomy #India