Could AI Make Money Matter Less? What Musk's 2036 Forecast Would Actually Require

Elon Musk has made another unusually ambitious prediction about the future of artificial intelligence. In a July 2026 interview with The Economist, he said that money “won’t matter in 2036,” connecting the idea to a future in which AI and large numbers of robots can provide more goods and services than people are able to consume.

The argument sounds simple: people need money because useful things remain limited and obtaining them usually requires someone’s work, resources or capital. If intelligent machines eventually perform an enormous share of production automatically and the cost of many goods and services falls dramatically, money could become less central to everyday life.

But moving from highly capable AI to an economy where money barely matters would require several enormous changes to happen at the same time. Automation would have to spread far beyond software, physical production would need to become much more abundant, energy and raw materials would still have to be allocated, and societies would need new ways of determining who gets access to valuable resources.

The more interesting question is therefore not simply whether money disappears in 2036. It is what would have to become true before money could begin losing the role it plays today.

Musk’s argument starts with abundance

Musk’s reasoning depends on combining advanced AI with robotics. AI has already become increasingly capable at working with text, software, images, research and other digital tasks. Physical automation is a much larger challenge because producing real-world goods requires machines that can manipulate objects, operate reliably in changing environments and interact with infrastructure built for humans.

In Musk’s scenario, sufficiently capable robots eventually bridge that gap. AI provides the intelligence while machines provide the physical ability to manufacture products, move materials, maintain infrastructure and provide services.

If those systems became inexpensive and widespread enough, the amount of economic output available per person could rise substantially. The important word is if.

Producing software is very different from producing housing, food, vehicles, electricity, medical equipment or new factories. A highly capable AI model may help design these things, but physical production still depends on materials, energy, machinery, land, logistics and time.

Money exists partly because scarcity still exists

Money performs several useful economic functions. It allows people to exchange different goods and services, compare their value and preserve purchasing power for future use. Automation could reduce the scarcity of certain things without eliminating scarcity itself.

Digital products provide a useful example. Once software has been developed, producing another copy can cost very little. The same is true for many forms of digital information. AI could extend this effect by making some types of knowledge work much less expensive to produce. Physical goods behave differently.

A second house still requires land and materials. A semiconductor requires fabrication capacity. An additional vehicle requires components, energy and manufacturing equipment. Electricity requires generation and transmission infrastructure.

Even in a highly automated economy, locations, natural resources, unique products and people’s time could remain limited. As long as people want more of something than can be supplied, some mechanism is needed to determine how that resource is distributed. Money is one such mechanism.

AI could still make many things dramatically cheaper

Money does not have to disappear for Musk’s broader argument to contain an important idea. AI and automation could make many products and services substantially less expensive.

Imagine software development where AI completes much of the routine work, factories where robotic systems operate continuously, automated transportation networks and AI systems capable of assisting with research, education and administrative tasks. The cost structure of those industries could change significantly.

Economists are already examining AI’s potential effect on productivity. The OECD has estimated that AI could meaningfully raise annual labour-productivity growth in advanced economies over the coming decade, although the size of the effect depends heavily on adoption, complementary investment and how rapidly the technology improves.

That is very different from an economy with unlimited production, but even modest annual productivity improvements can become significant when compounded over many years. The realistic near-term question may therefore be whether AI makes certain goods and services cheaper and more abundant, rather than whether it removes prices entirely.

Robots are the difficult bridge between AI and physical abundance

Digital intelligence alone cannot produce the world Musk describes. For an automated economy to transform everyday material production, capable machines would need to operate across an extraordinary range of environments.

Manufacturing is one part of that challenge. Agriculture, construction, maintenance, transport, warehousing, healthcare and household tasks all involve physical conditions that can be less predictable than a controlled factory floor.

Robotics is progressing quickly, but deploying machines at enormous economic scale also requires manufacturing the robots themselves.

That creates an interesting loop. Before robots can produce abundance, companies first need factories capable of producing large numbers of reliable robots, along with batteries, motors, sensors, processors and other components.

Eventually automation could help expand its own production capacity, but reaching that point requires substantial investment and infrastructure first.

Computing itself requires physical resources

Modern AI can feel almost entirely digital, but the infrastructure behind it is physical. Large AI systems require chips, servers, data centres, networking equipment, cooling and electricity. These requirements are already becoming economically important.

The International Energy Agency expects global data-centre electricity consumption to roughly double between 2025 and 2030, reaching around 950 terawatt-hours in its central projection. AI-focused data centres are expected to increase their electricity use even faster.

Efficiency improvements can reduce the energy required for individual AI operations, but rapidly increasing demand can still push total consumption higher.

A future filled with both powerful AI systems and large robotic fleets would therefore depend on continued expansion in energy generation, electricity grids, semiconductor manufacturing and supporting infrastructure.

That does not make Musk’s long-term abundance scenario impossible. It simply shows that intelligence alone cannot create it.

The bigger question may be who owns the machines

Suppose AI and robotics become extraordinarily productive. There would still be another question: who owns them?

If automated factories, computing infrastructure, robotics platforms and energy systems remain concentrated among a relatively small number of companies and investors, enormous production capacity would not automatically mean everyone receives an equal share of its output.

This distinction matters because technology determines what an economy can produce, while institutions and ownership influence how the resulting value is distributed. Current economic research already points to this issue.

IMF analysis has found that the economic benefits of AI may be distributed unevenly, particularly because people who already own significant productive assets can benefit from increased returns to capital while other workers depend more heavily on wages. A highly automated economy could make this distinction even more important.

If machines perform a growing share of valuable work, ownership of the machines, software and infrastructure could become more economically significant relative to supplying labour.

Work could change before money does

Musk has separately suggested that work could eventually become optional for many people. A complete transition of that kind would be an enormous change, but AI does not need to eliminate employment for the labour market to look different by 2036.

The OECD says AI is already affecting work through several overlapping channels. It can automate parts of existing jobs, complement workers performing other tasks and create new roles connected to deploying and managing the technology. That means entire occupations do not necessarily have to disappear for AI to create major changes.

A professional might continue doing the same job while AI handles a growing percentage of repetitive research, drafting or administrative work. A factory employee might supervise automated equipment instead of directly operating every stage of production.

Other jobs could become more valuable because they depend heavily on interpersonal communication, judgment, physical adaptability or expertise that works well alongside AI. The result may be a gradual rearrangement of work rather than one moment when employment suddenly becomes unnecessary.

Productivity gains and access are different questions

Imagine that automation reduces the cost of producing a particular service by 90%. That is a major productivity improvement. It still does not automatically determine who receives the service, at what price or under what conditions. A future economy would need to address both production and distribution.

Those questions become particularly important if human labour represents a smaller share of total economic output. Wages currently give most households purchasing power. If fewer people obtain most of their income through traditional employment, another mechanism may need to replace part of that purchasing power.

Various ideas have been proposed over the years, including broader ownership of productive assets, public benefits and different forms of income support. Which approaches societies eventually choose would be an economic and institutional decision rather than something AI itself determines.

Scarcity would probably move rather than disappear

Another possibility is that extreme automation changes what people compete for. If manufactured products become exceptionally inexpensive, value may shift toward things that cannot be easily replicated.

Land in desirable locations is an obvious example. There is only so much beachfront property or space in a particular city regardless of how many robots exist. Authentic experiences, personal attention, rare physical resources and access to limited events could remain valuable for similar reasons. Even time remains fundamentally limited.

Everyone receives the same 24 hours in a day, and no amount of industrial automation produces additional hours for an individual person. A highly productive future could therefore contain abundance in many categories while scarcity remains important elsewhere.

Money might become less relevant for everyday necessities without becoming irrelevant everywhere.

2036 is an extremely ambitious timeline

Ten years is a long period in technology, but a short period for restructuring an entire economic system.

By 2036, AI systems could be considerably more capable than they are today. Robots could also become much more common across factories, warehouses, transportation and other industries. That would still leave a large distance between widespread automation and a world where money no longer matters.

Physical infrastructure takes time to build. Electricity systems, factories, housing and transportation networks frequently operate on multi-year or multi-decade investment cycles. Institutions can move even more slowly.

Tax systems, pensions, employment law, financial markets, public services and international trade are deeply connected to monetary economies. Replacing those structures would involve decisions extending far beyond the capabilities of AI models or robots.

For that reason, Musk’s 2036 statement is more useful as an illustration of the destination he imagines than as evidence that an exact economic transition has already been established.

A more realistic 2036 could still look very different

Money does not need to disappear for 2036 to feel economically different from today. AI assistants could perform far more professional work. Robotics could become common in manufacturing and logistics.

Automated systems could reduce the cost of transportation and certain services. Small businesses could gain access to capabilities that previously required much larger teams. Some goods might become substantially cheaper.

At the same time, demand for electricity, computing infrastructure and advanced chips could continue increasing. Certain skills could become more valuable while others are used less frequently. Ownership of productive technology could play a larger role in determining economic outcomes.

That would already represent a major transformation. And unlike the complete disappearance of money, many parts of that scenario are extensions of technological and economic changes already underway.

The future of money depends on more than AI

Musk’s prediction ultimately rests on a simple economic idea: if everyone can obtain nearly everything they need without meaningful scarcity, the reason for exchanging money becomes weaker. The difficult part is reaching that condition.

AI may make intelligence inexpensive. Robotics may eventually make certain forms of physical labour inexpensive. Neither automatically makes land, energy, raw materials, infrastructure or unique resources unlimited.

Nor do those technologies decide how ownership and access should work. That means the future of money will depend on a combination of technological capability, productivity, resource availability and economic institutions.

By 2036, money may still look very familiar even if AI has transformed large parts of the economy. The more immediate change could be something subtler: people may discover that certain things which once required significant amounts of money, labour or expertise have become dramatically easier to obtain.

If that happens across enough areas of life, money could gradually become less important for some needs without disappearing altogether. That outcome may be less dramatic than predicting the end of money, but it would still represent one of the biggest economic changes of the AI era.