Many believe artificial intelligence (AI) will be the next “gold rush,” akin to railroads or the internet. However, Jerry Neumann cautions that AI won’t make everyone wealthy. The key lies not just in the value a technology creates but in who captures that value.
History offers lessons. In the 19th century, railroads and later electrification generated millions of jobs and thousands of companies. The internet in the 1990s sparked over 7,000 IPOs in the U.S. alone. But not all transformative technologies follow this path. Containerization in logistics, for example, slashed costs and quadrupled global trade from 1960 to 1980, yet it didn’t spawn countless new corporations. The winners - companies like Walmart and IKEA - were those that best adapted to the new landscape.
Neumann argues that AI resembles this latter model. Unlike personal computers or the internet, which empowered thousands of small developers to innovate, today’s large language models are controlled by a handful of giants - OpenAI, Google, Anthropic, Meta.
Building a comparable model costs billions: training GPT-4 reportedly cost $100 million, and newer models demand over $1 billion in investment. This effectively shuts out newcomers. The AI model market is no longer in its early, open phase; it’s now a stage where dominant players are solidifying their positions.
The U.S. stock market currently lists about 4,000 companies, half the number from 1996, partly because large corporations acquire potential rivals. AI may follow suit: innovative startups are likely to be swallowed by tech giants.
The AI application market, however, is set to grow. Forecasts suggest the “AI-in-gaming” sector will rise from $2.3 billion in 2023 to $28 billion by 2033. In healthcare, automation could save up to $200 billion annually in the U.S. alone. PwC estimates AI could boost global productivity by 14%, or $15 trillion, by 2030 in professional services. But who will claim these gains? It’s unlikely to be thousands of new companies.
The primary beneficiaries will likely be established players integrating AI into existing operations. Banks, for instance, could automate credit checks, saving hundreds of millions annually, while consulting firms might streamline data analysis, cutting personnel costs. In these cases, AI boosts margins but doesn’t birth new industries.
There’s another possibility: in some sectors, AI could radically reshape business models. In healthcare, it might democratize diagnostics; in education, it could enable personalized “AI tutors” for millions. Companies that pursue these transformative paths could unlock entirely new markets, not just enhance existing ones.
Neumann’s conclusion is clear: AI will transform the economy, but it’s not a guaranteed ticket to riches for every investor or entrepreneur. Like containerization, its benefits will be unevenly distributed. The technology itself matters less than who applies it and how.
Also read:
- AI Vulnerabilities Are Becoming Critically Dangerous
- Scientists Confirm: Nature Walks Outperform Therapy for Boosting Mood and Reducing Anxiety
- Rest is a Fiction: New App Generates Your Fake Vacation Photos
- No News from the US: War Enters Dull Phase, Enemies Identified, Lines Drawn, Artillery Preparation and Reconnaissance Underway
Despite the grim outlook for AI model development, certain areas hold promise. Middleware companies (interfaces between large AI models and users), firms focused on privacy and data, and niche, domain-specific models are likely acquisition targets for corporations seeking ready-made solutions.
Rather than betting on core AI models, investors should focus on sectors where automation enhances productivity: professional services, healthcare, education, and creative industries. AI excels at finding patterns in ambiguous data, generalizing, and providing probabilistic answers - making it invaluable where information is complex.
Those who can reorient their businesses to leverage these opportunities will thrive. For everyone else, AI may simply become another line item in the budget.
