How AI Will Impact Our Lives in 2025 and Beyond

To help you understand how AI and other new technologies are affecting energy consumption, trends in this space and what we expect to happen in the future, our highly experienced Kiplinger Letter team will keep you abreast of the latest developments and forecasts. (Get a free issue of The Kiplinger Letter or subscribe). You'll get all the latest news first by subscribing, but we will publish many (but not all) of the forecasts a few days afterward online. Here’s the latest…

Artificial intelligence will affect everything from drug discovery to military strategy, and will boost productivity, automate work and disrupt industries.

Here’s a look at big opportunities of generative AI, tech that responds to plain English (and other prompts) to create text, audio, images, video, computer code and more. We highlight the looming challenges, too.

From just $107.88 $24.99 for Kiplinger Personal Finance

Become a smarter, better informed investor. Subscribe from just $107.88 $24.99, plus get up to 4 Special Issues

CLICK FOR FREE ISSUE

Sign up for Kiplinger’s Free Newsletters

Profit and prosper with the best of expert advice on investing, taxes, retirement, personal finance and more - straight to your e-mail.

Profit and prosper with the best of expert advice - straight to your e-mail.

Sign up

The big AI bet

Astronomical tech spending is laying the foundation for the coming AI age. Tech giants’ capital expenditures have exploded since the launch of ChatGPT, OpenAI’s AI chatbot, in 2022, and heated competition will push spending higher. The combined yearly capital expenditures of Amazon, Google, Meta and Microsoft will soar past $300 billion in 2025, up from $154 billion in 2023, according to estimates from Morgan Stanley Research.

Most of the money is for Nvidia’s AI chips, the graphics processing units (GPUs) for huge cloud data centers owned by those tech giants and others. Other spending is for memory chips, networking gear and related data center tech, boosting Dell, Cisco, Oracle, Arista Networks, SK Hynix, Micron, Marvell and others.

The frenzied spending has sparked fears of an AI bubble, à la the late-1990s dot-com debacle. Today’s pricey AI requires lofty returns on investment to pay off. Though a correction or more painful period may occur, tech giants have massive cash piles and user bases to weather ups and downs in a decade-long AI journey. They are betting on a shift to the next big tech platform, as important, if not more, than the transition from PCs to mobile phones, which makes moving fast imperative.

Green shoots are already emerging. Microsoft-backed OpenAI has more than 300 million weekly users for its chatbot. Meta has 600 million AI users on its apps and other tools and says AI is improving ads and content suggestions. Big Tech companies are seeing fast-growing cloud sales for AI computing and tools. Consulting firm Accenture has already booked $3 billion in AI sales. AI start-ups are scoring users quickly. But the real promise is long-term. For example, OpenAI expects to lose billions yearly until it finally projects a profit in 2029 on an astounding $100 billion in yearly sales, according to a report by the New York Times.

Harnessing novel tech

Generative AI stems from huge amounts of data and computing power, using the brute force of GPUs on a mountain of internet and other data. The AI systems are known as large language models that learn patterns in language to predict the next word, enabling remarkably complex and humanlike responses. The inner workings of these digital neural networks are not fully understood. Flaws, such as errors and unreliability, are still common. But expect steady improvements.

Among the companies building leading AI models: OpenAI, Anthropic, Meta, Google, Mistral AI, Cohere, Amazon, xAI and Stability AI. It costs hundreds of millions of dollars to get AI models to work, a process known as training, and takes months of constant data center computation on hundreds of billions of words or other data. That cost will go into the billions with an ever-growing demand for computing power.

AI's growing physical footprint

AI software is backed by physical infrastructure. Start with Nvidia’s dominance, with millions of its high-end GPUs, costing $30,000-$40,000 each, in data centers. Nvidia’s lead is poised to last for years, as its proprietary hardware and software are best in class. The company’s revenue is forecast to hit nearly $200 billion per year in its 2026 fiscal year, up from $27 billion in fiscal 2023, according to Morningstar.

Competitors are trying to find Nvidia alternatives. Amazon, Meta and Google are building their own silicon. AMD and Intel have AI chips. “In the long run, we expect tech titans to strive to find second-sources or in-house solutions to diversify away from Nvidia in AI, but most likely, these efforts will chip away at, but not supplant, Nvidia’s AI dominance,” writes Brian Colello, Morningstar strategist, in a recent report on the company. Nvidia won’t stop improving as it seeks new customers in anticipation of cooling sales.

U.S. data center growth is booming, reaching a pace of $30 billion in private spending this year, up from $12 billion in 2022, according to Census data. Northern Virginia is by far the largest market, but rapid growth is spread across the U.S. in states such as Iowa, Idaho, North Dakota and Texas. Examples of the trend include Meta investing $10 billion in a new 4 million sq.-ft. facility in Louisiana, and xAI planning to expand its Memphis, Tennessee, data center to 1 million GPUs.

AI data centers are huge energy hogs, so expect soaring electricity demand. Data centers could consume up to 9% of U.S. electricity generation by 2030, more than double today’s amount, says energy research outfit EPRI. To achieve this will require tens of billions of dollars of investment in energy and grid expansion. Data centers will spur demand for reliable power, especially sourced from natural gas. Set to benefit: Constellation Energy, Vistra, ExxonMobil, Entergy and Talen Energy.

Projected energy needs are so big that tech giants are serious about using nuclear to power nearby data centers. Google wants to add 500 megawatts of nuclear energy with Kairos Power’s small modular reactors by 2035. Amazon has similar agreements in Virginia and Washington and is investing in nuclear firm X-energy. Meta has a proposal to add 1-4 gigawatts of nuclear capacity by the early 2030s. In a $1.6 billion deal, Microsoft partnered with Constellation Energy to restart a Three Mile Island reactor. Data center operator Equinix has signed a deal with microreactor maker Oklo.

Even with big energy investments, power shortages will be a bottleneck for AI development. Chips are sure to get more energy-efficient, but not fast enough to curb demand in the near term. Nuclear won’t make an impact until the 2030s.

AI takes hold

For an early sign of what’s to come, look at top AI models that are emerging fast. The tools can create stunning virtual 3D worlds, lifelike voice assistants, Hollywood-like movies, full-length books and more. Google’s recent Genie model can create 3D video games, animations and prototypes by inputting simple text or images, such as writing “Show me a wizard walking through an ancient forest.”

OpenAI, Google, Amazon and others are regularly releasing newly improved tools, from low-cost ones for mobile devices to expensive ones with high-end performance. Meta’s top AI models are open-source, so they are freely available, a contrast with the others’ proprietary efforts, spelling even more competition and choices.

The tech is quickly being embedded in everyday apps and services, including Microsoft 365’s suite of word processing and other apps, web browsers and mobile operating systems from Apple and Android. The tools can write emails, summarize documents, edit photos, create presentations, answer questions, etc.

Unlike social media and web search, the business model can’t be free and reliant on ads because building leading-edge AI is so expensive. Subscriptions are a top option, especially for business users wanting powerful AI for specific tasks. OpenAI announced a $200-per-month subscription for its top AI model, for example.

AI start-ups keep scoring huge sums based on the disruptive potential to shake up entire markets, such as online search, and take on incumbent services, such as sales software. Among the start-ups with valuations in the billions are Glean, Hugging Face, Databricks, CoreWeave, Safe Superintelligence, Poolside and Liquid AI. Venture capitalists and large companies are among the investors. Another example: Tenstorrent recently raised $700 million to build more affordable advanced AI chips.

The race to super AI

Is AI on the path to superintelligence? A lot of top researchers think so. Many companies’ goal is creating so-called artificial general intelligence, an AI that cracks the code of human intelligence, and then improves upon it. OpenAI defines AGI as “highly autonomous systems that outperform humans at most economically valuable work.” Others say it’s a Ph.D.-level virtual genius.

This ill-defined inflection point has an uncertain timeline, which ranges from next year to sometime in the next decade, or being delayed indefinitely, say its advocates.

In theory, AGI poses endless possibilities but also dire existential risks. The promise includes rapid productivity gains and booming economic growth, as scientific discovery is accelerated to cure cancer, solve nuclear fusion, etc. Dangers include self-aware AI going rogue, starting a war or even ending humanity.

Take AGI talk seriously, but with a healthy dose of skepticism. AI hype is partly at play, though the race to AGI will be a driving force for years to come. Past AI hype has fallen flat and decades of research have seen fits and starts. AGI isn’t required for amazing AI tech emerging from steady advances. The pressing risk for society is how humans decide to use powerful AI.

The emerging use cases

There’s a rapidly growing list of uses for generative AI across industries. It will become a powerful tool for knowledge workers. Widely available tools can already transcribe virtual meetings, analyze company data, find cyber threats, research new markets and more. Users interact with AI chatbots like web search typing text to get responses. More questions or commands can refine the results… “provide more details,” “simplify the text for a nonexpert,” “what am I missing?” etc.

There’s an explosion of AI apps and business adoption is ramping up, though it will be a bumpy ride finding the best-use cases and ways to train workers.

Specific areas where AI will gain traction:

• Information technology. Writing and debugging computer code is one of the top uses to boost productivity. Users can paste sections of code to check for errors or input text to build an app. Among the vendors are GitHub, Codeium, Cursor, Harness, All Hands and Factory.
• Media and entertainment. Create a full-length film from a text description. Hollywood will let fans personalize movies, simply entering something such as, “Make a Superman film about sharing, suitable for an 8-year-old,” into a generative AI tool. Movie special effects will be easier. Ditto, commercials, video games, etc. AI tools include Runway, Midjourney, Canva, Pika and Ideogram.
• Retail and advertising. Product descriptions on websites can be automated. Small businesses can cheaply and easily generate high-quality video ads, image ads or marketing text. Facebook is already seeing lots of small firms using its AI ad tools.
• Education. Teachers can create lessons, study guides and essay questions. Ditto for students, who can get bullet points from a long text or help with an essay. Personalized AI tutors such as Khanmigo break down complex topics step by step.
• Healthcare. Tools can transcribe audio of patient visits, send messages or automate notetaking. They can streamline required reporting or help with diagnoses. AI tool Whisper is already widely used. Others include Abridge, Forcura and SmarterDx.

There are many other areas, including legal, finance, pharma, real estate and publishing. AI can pore through scientific papers for insights, or make models of drug molecules. It can give suggestions to deal with a mountain of regulations and compliance tasks.

The Holy Grail is autonomous AI assistants that can do complex work. These “agents” will take control of a computer to pull up documents and websites to book hotels, schedule meetings, do work research and more, completely on their own, while keeping the human in the loop. They’ll call customer service for any issues, and AI may even talk to other AI. Assistants may become intimate virtual friends who understand tone and mood. And every worker could have a tireless assistant.

Intensifying global competition

When it comes to global AI competition, the U.S. has a dominant lead by almost every notable measure, including investment, leading-edge AI systems, data centers, start-ups and jobs, according to a Stanford University AI Index Report. China ranks second in recent power rankings, but as an example of our edge, the U.S. had nearly nine times the private investment in 2023. Rounding out the top 10 rankings: the U.K., India, United Arab Emirates, France, South Korea, Germany, Japan and Singapore.

Still, Chinese tech giants are rapidly launching and improving AI models, including ones from Alibaba, Baidu, ByteDance, Huawei and Tencent. Those tools are already used by hundreds of millions of Chinese users and lots of Chinese firms. No doubt such tech is being widely tested and used by the Chinese government.

The U.S. Dept. of Defense is planning a military AI arms race with China, spurring a flurry of activity by the Pentagon that will only accelerate. There are even concerns China could beat the U.S. to an AGI-type breakthrough. In 2023, the DoD launched a generative AI task force and has since unveiled projects for military chatbots that can assist with acquisitions, sift through intelligence, answer research questions, analyze data, provide insights for decision-making, etc.
Other ideas: Using AI for autonomous weapons, virtual assistants on the battlefield, tracking terrorist activities, cybersecurity defenses and satellite imagery analysis.

Big defense contractors and start-ups are racing to build military AI tools. Lockheed Martin and Leidos are working on secure software products. Anthropic, Meta, Microsoft, OpenAI and others with leading AI models are aiming for contracts, including through partnerships. For example, big data analytics company Palantir has partnered with Anthropic to provide AI to U.S. defense and intelligence agencies to process huge amounts of data, uncover patterns, streamline documents and more. Other vendors in the military market include Shield AI, Anduril, Scale AI, Pryzm and Ask Sage.

Expect military and government leaders to express rising levels of concern on both the risks of using AI and the risks of foreign adversaries embracing the tech. Military leaders are worried about unreliable or inaccurate AI-generated responses, so they’ll tread lightly before relying on AI-generated insight for high-level decisions. The Pentagon’s research arm will spend billions of dollars rigorously testing AI for reliability and trustworthiness. Federal efforts to conduct AI safety tests in classified settings are underway, such as identifying how much info AI models have about nuclear weapons and whether it presents a pressing national security risk.

The many looming challenges

Of course, this breakthrough tech brings a myriad of knotty challenges.

Generative AI will be used for propaganda and to interfere in democracy, since it’s so easy, fast and cheap to create a tsunami of convincing AI propaganda, including text, images, videos and audio. For example, China meddling in Taiwan’s elections. Or Russia having an easier time launching cyberattacks via AI-created e-mails.

AI advances could hit roadblocks. Energy and chip shortages may happen. Companies are running out of data on the internet to train models. There are worries that AI progress may stall as current methods hit a wall. Many worry about the risk of onerous regs slowing progress, spending drying up or looming AI talent shortages. Uses that boost productivity could take longer than expected to gain wide adoption.

Other top concerns relate to privacy while using AI, intellectual property, cybersecurity and misinformation from AI that has gobbled up data with little to no oversight, as well as job losses from AI automation and mindless AI spam flooding the internet.

But many of these problems have solutions and they don’t negate AI's potential.

This forecast is from the team at The Kiplinger Letter, which has been running since 1923. It is a collection of concise weekly forecasts on business and economic trends, as well as what to expect from Washington, to help you understand what’s coming up to make the most of your investments and your money. Subscribe to The Kiplinger Letter.

Learn more about AI

• How to Protect Your Privacy While Using AI
• Major AI Companies You Should Know
• How AI Could Have Positive and Negative Effects on Cybersecurity
• Can Stocks Picked by Artificial Intelligence Beat the Market?
• Best AI Stocks to Buy
• Tech Heavy Hitters Join Forces to Form AI Alliance