New year holidays are popular time to make predictions for the new year. While making predictions, we humans tend to over estimate what can happen in the short term and under estimate what could happen in the longer horizon. Technology industry is currently obsessed with generative AI and LLMs. Although lot of the optimism about generative AI is deserved, it does feel like it is over hyped. There are number of other technologies that are equally or more promising but are ignored at the moment. Here are some of those promising technologies that will have large impact on our lives in this decade.
Energy
Cost of energy produced by Solar+Battery will a be fraction of fossil fuels and they will continue to fall on a Moore’s law like curve. Cost of electricity produced by Wind+Solar+Battery will be one third the cost of electricity produced using Coal, Natural gas or any other fossil fuels without any subsidy by the end of the decade. This is largely based on the work of Tony Seba. If you are even remotely interested in the future of humanity, you should definitely follow Tony Seba. Watch all of his talks and read all the ebooks he has on his site. They are quite informative and eye opening.
According to International Energy Agency, cost of electricity produced by Solar and Wind is already cost competitive with all other fossil fuel sources. But when you add the cost of battery storage, the cost becomes somewhat borderline. For example in California where day time cost of electricity is ~$0.10 and evening and night time electricity cost is ~$0.30, adding battery storage is no brainer with LFP batteries. you can recover the cost of installing and operating a battery storage plants in 5 years. But the cost equation is not as appealing everywhere else on the planet yet. California is an extreme example partly due to subsidies associated with residential solar. But this equation will significantly improve with sodium ion batteries.My own prediction is: cost of Solar+Battery based energy will continue to decrease by 40% every three years for the foreseeable future. Most of these cost gains will come due to massive Solar cell and Battery factories being built in China. This continuous drop in energy costs will usher in the biggest productivity gain humanity has ever seen. Bigger than the productivity gains seen during 1880s-1920s period. It will impact more people (~4-5 billion people) than the post war growth period. This will play out over couple of decades. But we will see major impacts in this decade.
Nuclear fusion will drive the cost of energy further down in the next decade. But it will not have the scale and will not be cost competitive by the end of this decade. Expect some fusion energy plants to enter commercial production before the end of the decade. Fusion will probably make the long standing dream of “energy too cheap to meter” a reality. But most likely in the next decade, not this one.
Meanwhile the traditional Nuclear fission based production will go sideways. Even though modern nuclear technologies like Small Modular reactors, Molten Salt Reactors and Thorium reactors are all very safe and make lot of sense, the regulatory framework will derail these technologies. One bright spot might be the use of Thorium reactors by container shipping industry. Expect Asian container shipping firms to aggressively deploy thorium reactors. Western shippers like Maersk will follow them closely.
Transportation and logistics
EV transition for consumer vehicles will be complete by the end of the decade. There will be no new fossil fuel based consumer vehicles made by the end of the decade. EVs will be cheaper than all ICE vehicles within the next two years. EVs with 1000 mile range will be the norm by the end of the decade. This will seal the fate of ICE vehicles. Biggest beneficiaries of this transition will be the Chinese EV companies. Expect most of the car companies in US, Germany and Japan who dominate the market today to cease to exist in their current form by the end of the decade. This will have significant impact to the suppliers in the automotive sectors. Expect sea changes to manufacturing sector due to death of western car companies and the dominance of Chinese EV/Battery companies. Western countries will try to erect trade barriers to Chinese competition in automotive sector. But this will not be able to turn the tide on Chinese EV/Battery manufacturing might.
Expect most of the public transportation sector to transition to EV as well. The short haul trucking might move to EV as well. But Long haul trucking and freight trains might be the slowest to transition to Electric. River barges and ships will probably not move to EV in this decade.
Autonomous driving will become the norm within next three years. We have been beta testing autonomous vehicles for more than 15 years. DARPA urban autonomous driving challenge was in 2007. Google has been working on autonomous vehicles at least since 2008. Autonomous technology seems to be ready for prime time finally. Tony Seba has some good insights on how Transportation-As-A-Service will change the way we own and use cars. My own prediction is, TaaS is not for everyone. TaaS will probably dominate the lower end of the market and rental industry. But many of the middle class and upper middle class folks who are germaphobic and not comfortable sharing their personal vehicles will stick to vehicle ownership. I expect bigger autonomous EV vans with comfortable seating and big screen TVs to drive us around while we relax (and maybe work) inside. With autonomous vehicles, we will no longer be required to sit in uncomfortable drivers seats and be frustrated by traffic jams.
Autonomous logistics will become the norm. The autonomous technology will change how we move goods, manage the warehouses and logistics in general. Robot use in the logistics sector will probably see a Moore’s law like curve where it is doubling every few years. Most of these robots will not be humanoid robots. They will be mundane robots like the Kiva robots used in Amazon fulfillment centers. This will start with long haul trucking industry moving to autonomous drivers. Soon enough, the loading/unloading, sorting, packaging and order fulfillment operations will will be automated. Imagine a world where the autonomous cranes unload container ships, ship the container to a warehouses using autonomous trucks. Autonomous unloaders at the warehouse unload the pallets and store them in storage areas. From there the pallets get shipped to fulfillment centers or super stores. All of this done autonomously without need for human operators. Only human involvement would be the sales folks taking the orders and supply chain strategists planning the logistics.
Autonomous transportation will lead to the creation of Gigacities. With the combination of autonomous technology and working from home, the urban centers will slowly disappear and the suburban sprawl will at-least double in radius. Current suburban sprawl emerged in the 50s as most American families bought cars. We were so far limited in the size of the sprawl by commuting distances. With autonomous technology, commuting distance will no longer be a limiting factor. Story of civilization has been to build bigger and bigger cities where larger number of people can come together, work together and collaborate closely. This trend will continue.
Below ground transportation will become cheaper than above ground transportation. The Gigacities will need new forms of transportation to get around quickly across the city. Common sense says, building roads on the ground should always be cheaper than tunnels. Even in congested areas, building double decker or triple decker viaducts should be cheaper than tunnels. But, building roads or rails above ground requires land acquisition which can only be done by governments at scale. This means governments have a monopoly on building above ground infrastructure. Governments rarely have the need to drive down the cost of building infrastructure. This has gotten so bad in most of the developed countries that building new road in urban areas cost 20x more than rural area. For example in California it costs $62 million per mile per lane of road in urban and suburban areas. In rural areas, it could be as low as $2.5 million per per lane. Today the below ground transportation is quite expensive. Building tunnels costs over $100 million per mile per lane. A two lane tunnel costs about $250 million per mile. The Boring company tunnels in Las Vegas have been costing $15 million per mile. They are relatively narrow single lane tunnels. Cost of tunneling will rapidly decrease to around $4 million per mile per lane. There are number of problems to be solved here including efficient concrete segment production & installation, spoil disposal, air circulation, tunnel network operation etc. Who is better to figure out these complex problems than Elon Musk.
Manufacturing
Cheap energy and autonomous industrial robots will create a manufacturing renaissance. Manufacturing is already heavily automated. If you tour a modern semiconductor factory or a battery factory, you will notice there are no humans involved in the manufacturing process itself(this is true even in Asia). Humans are mostly inspecting the screens to make sure nothing goes wrong. So the incremental automation does not have the same level of ROI as some of the other sectors. As logistics becomes fully autonomous and the cost of energy keeps dropping(See [1] above), companies will be incentivized to move to lights-out-factory model.
There are sectors like electronics assembly, Apparel industry and heavy manufacturing where the human labor is heavily used, there is more need for automation. But these factories have already move to low wage countries. So the progress here might be slow.
Technology
Autonomous operations will have bigger impact than Generative AI. There is lot of excitement in the technology industry about Generative AI, LLMs and AI/ML in general. Broader theme of AI/ML changing how we operate as a society is probably correct. I personally think autonomous logistics and industrial robots will have bigger impact on overall human productivity. An AI enthusiast will gloss over this and say this is all AI. But, large scale autonomous operations requires a large scale integration of systems, which we are not very good at.
Imagine a textile supply chain, cotton grown in Brazil might be shipped to a factory in Indonesia to be made into textile. From there it might go to a factory in China which is processing a fashion clothing order from France. This factory might be importing machinery from Germany, zippers from Japan and buttons from Korea. Then the finished cloth might be shipped to US so that a consumer can buy it on his phone as an impulse buy. There might be thousands of companies, suppliers, warehouses, shippers, brokers involved in the whole process today. If the whole thing has to function autonomously, we need to figure out how to operate such planet scale systems. How to coordinate millions if not billions of robots and autonomous systems.
Busting of bureaucracy will result in massive productivity gains. This is one of the most underrated areas that people are not paying attention to. What we humans call civilization is our ability to cooperate at a large scale. According to Yuval Harari, this is what separated us homo sapiens from Neanderthals. We homo sapiens could organize at much larger groups than any other species that ever existed. Every civilization from the Egyptians, to the Romans to Chinese empires, to Mongols, to British empire, what made them successful was their ability to coordinate at larger and larger scale than previously possible. With industrial revolution and assembly line, we could build bigger and bigger factories and coordinate supply and distribution at scale. We reached the peak of this with industrial scale killing armies of WWII. An estimated 60 million dead in a span of six years using global scale coordinated armies of WWII. People called this Managerial Revolution at the time. We had reached a scale where we could coordinate the economic activity of entire continents. Which lead to our decades long experiment with communism to see if we could run entire countries on planned economy(which failed). Today the internet and software allows us to coordinate our activities at much larger scale than we could imagine before.
During 1930s Ronald Coase asked the question, why does a firm exist and keeps getting bigger. If we know the communism doesn’t work because of lack of incentives, and we know free-market economies are the best way to organize the economies, why do we have large corporate conglomerates? Aren’t they same as running mini communist organization? Most corporations are structured exactly like a communist bureaucracy. They are top down, centrally planned, most people in the firm lack clear incentive. In fact most corporate employees think they are doing a bullshit job. The answer had to do with the cost of coordination. There was high cost of coordination across different companies and lower cost of coordination within the companies. If enterprises had to go to open market for every little thing, cost of coordination would be too high. We built large corporate bureaucracies to overcome this coordination problem. Departments with specific tasks and everyone know how to do their little part. We have also built large government bureaucracies to solve the same coordination problems.
With the internet, our cost of coordination has gone down significantly and continues to go down. Imagine two employees of a large multi national corporation who are working from home coordinating their work using digital tools, it makes little difference whether they are working for same company or different one. Internet and software provides us fundamentally different ways to coordinate at large scale. In the near future I expect most of the bureaucracy to come undo. Dismantling of this bureaucracy will result in large scale productivity improvement.
You can see some of it happening already. For example Uber displaced a taxi bureaucracy in most cities. Airbnb allows property owners, hosting companies, cleaning companies to collaborate very differently to provide accommodation compared to a hotel chain like Marriott. The traditional enterprise organization structure, tools, processes make collaborating with people inside the org easy while making working with people outside the org difficult. We will see the emergence of new kind of software tools that make these lines very blurry.
We will see the emergence of two-pizza companies from two-pizza teams. Amazon is famous for creating two-pizza teams(small teams that can be fed by two pizzas). These are teams of 8-10 people who have a complete responsibility for one service or one product. Amazon is also well known for giving small teams complete P&L responsibility to running its own operations the way they see fit. It is much easier to align incentives in smaller teams than big complicated organizations. Haier has famously created hundreds of micro-enterprises where each micro-enterprise owns a specific product category offered by Haier. Each micro-enterprise is a startup with entrepreneurial spirit and with its own decision making power and complete responsibility for P&L of the product category. We will see more of these dis-aggregated enterprise model emerge.
Today, our accounting practices, tax rules and capital markets favor larger and larger companies. Soon this model will reverse as we realize all these large enterprises have the same problems as centrally planned communist countries and the modern digital tools allow us to have smaller companies with much better aligned incentives. Saying all of them will be small companies of 8-10 people is stretching it bit too much. But they will be fraction the size of today’s fortune 500 companies which employ 100,000s of people.
Financial services
Most financial services firms will cease to exist as we know them today. The business model of most of these firms has become either central bank arbitrage or investing in unproductive asset classes like residential mortgages and government bonds. This will collapse on its own weight. World economy can not continue investing in unproductive assets forever. We will look at each sector of financial services separately.
Traditional banking will become digital-only commoditized business. Traditional banking today is a branch heavy business. The big five banks all have 10s of thousands of branches across the country. Only reason to maintain these branches is so that they can sell other more profitable services like mortgages or investment services. In near future all three of these will become digital services. We will all be getting mortgages from online brokers, getting investment services from digital investment services. Traditional banks will go through massive pain. The cost structure of traditional banks which employ people in branches is just unsustainable. They will try to use regulation as a defense mechanism. But this will not last long. Maybe handful of the large banks might survive as custodial services for digital banks. Most other smaller local banks and credit unions will disappear. There is a hard truth that the traditional banking sector employs closes to 3 million people and they add no real value in a digital world. Now, they are all glorified paper pushers who exist only because of government regulation. They served a real purpose when person to person was the only way to do banking. Not anymore.
The loan underwriting business of these banks will be divested from these banks. Most of these loan underwriting will move to private equity like entities. More on this later.
Payment networks will disappear and all payments will be smart-phone/device based. Another cash cow of banking business is payment processing. Visa and master card process $5 trillion in purchases. Globally, this is closer to $20 trillion. In the near future the payment cards will disappear and all payments will be based on smart phones and other smart devices. Payment terminals are already disappearing and being integrated into POS terminals. This will end the payment network monopoly. What will replace this will look somewhat closer to China’s WePay or India’s UPI. This might even be based on CBDCs or central bank payment rails. But this is going to be a free or almost-free payments processing without any commission. The digital-only banks will be offering free payment services for an opportunity to offer the businesses and consumers other services. None of these payment services will be based on Bitcoin or blockchain. They will be based on payment rails that are blessed by central banks and legislature. They will be subject to all the traditional KYC and AML rules. Governments are not going to let go of these controls any time soon.
Insurance industry will go through disaggregation. Insurance is a diverse business with many sub category like health insurance, life insurance, auto insurance, property insurance etc. Each of these sub categories have their own dynamics thats too long to get into in this essay. But all of these have commonalities like, agents or sellers, underwriters and claims processing. All three of these will become separate entities. Insurance sellers or agents will be replaced by digital marketplaces. Most current insurance companies might survive as underwriters. Underwriting will be a purely data science business. Claims processing will be outsourced to independent businesses. Most existing insurance companies will struggle to keep up with this disruption because of their cost structures. One example of this in healthcare sector is, around 40% of the costs associated with healthcare is going to bureaucrats who manage the program instead of healthcare workers. This seems overly excessive and unsustainable.
Low cost robotic lending funds will takeover standardized consumer loan businesses like mortgages and auto loans. Most of standardized lending businesses are extremely inefficient and assume almost no risk. Most of the risk is taken by government agencies like FannieMae or backed by real assets. We don’t need an army of mortgage brokers and loan processors doing unnecessary drudgery of paperwork. This is better done by simple digital marketplaces and algorithms.
Private equity firms will take over most of corporate lending. Most of western world has been in capital investment drought for the last 30+ years. Banks have forgotten how to lend to corporations or invest in productive asset classes. As the banks lucrative businesses disappear and they go through turmoil more private equity style firms will step up to fill the void in corporate lending space.
Tokenization of assets will happen, but they will not be based on blockchain. The tokenization effort will be driven by regulators. They will look similar to mobile driving license and digital identity standards driven by regulators. They will be very much centralized regulated efforts with government sponsored entities acting as maintainers of centralized ledgers.
Government services
Budget shortfalls will force the governments to break up their bloated bureaucracies and embrace digital services. There is general consensus that the easy money policies of central banks has enabled deficit spending by governments and made them bloated. US federal government employs close to three million civilians. The state and local governments employ another 20 million people. It is anybody’s guess if we really need this many people to run the government programs we have. Governments themselves have no way to measure the efficiency of this workforce and no incentive to increase efficiency.
Take the example of social security administration. It is a good candidate for becoming a digital-only service. It collects social security tax, invests in government bonds and pays out to beneficiaries. All of these could be done digitally or algorithmically. Social security administration employs more than 60,000 people, which probably made sense when they had to manually mail the checks to beneficiaries and retirees had to file for social security benefits using paper forms. A digital only social security administration should be able to do all of these with less than 2,000 employees.
Closing thoughts
Making predictions is easy, anybody could do it. As Steve Jobs said: ideas are worth nothing unless executed on. So here is to the makers, the builders, the dreamers, the crazy ones, the rebels and the entrepreneurs who will make all of these come true.