The Quiet Revolution: Here's why the next AI giants won't need giant models!

For this article, let’s talk about a coaching comment my friend and colleague, Professor Joel Litman, presented to his workforce at Valens Research. Curious?

Continue reading the article below to know more.

The Quiet Revolution: Here's why the next AI giants won't need giant models!

In every technological revolution, there comes a moment when the spotlight shifts.

The noise fades from the headline-makers, and quietly, new leaders emerge—not by building the biggest machines, but by using them better.

Professor Joel Litman , Chairman and CEO of  Valens Research  and Chief Investment Officer of  Altimetry Financial Research, believes that moment has arrived in the world of artificial intelligence (AI).

While much of Wall Street and Silicon Valley remains fixated on massive AI models and their escalating capabilities, Professor Litman argues that the  real  opportunity lies elsewhere.

For years, the world has watched tech giants like  OpenAI,  Alphabet, and  Anthropic  battle for dominance in what he calls the “AI Model Wars.”

Billions are poured into training larger, faster, and supposedly “smarter” systems. Each company races to outdo the other with sheer scale and computing muscle.

However, Professor Litman says the game is changing, and smart investors need to change with it.

From Builders to Users

According to Professor Litman, venture capital is flowing into a new breed of companies—smaller, agile startups not building AI infrastructure from scratch but using existing models to create powerful, scalable applications.

He compares today’s AI boom to the dot-com era of the 1990s. Back then, investors piled into infrastructure stocks like  Cisco Systems, whose routers and switches powered the Internet’s foundation.

Cisco even became the world’s most valuable company for a time, boasting a USD 555 billion market cap.

… but the true giants of that era didn’t build the Internet; they  built on it. Companies like  Amazon  leveraged Cisco’s infrastructure to create services consumers actually used.

That’s where the trillion-dollar value was ultimately created.

According to Professor Litman, history is repeating itself. 

While everyone’s watching the companies training the largest models, the real money is being made by those building practical, profitable tools on top of them.

The distinction is simple but profound: Foundational model developers face astronomical costs and uncertain returns, while application-focused companies can move faster, spend less, and generate real revenue sooner.

Professor Litman even points to  Harvey, a legal-tech startup that uses existing AI models to automate contract reviews and litigation support—tasks that used to cost lawyers hundreds of dollars per hour.

By 2024, Harvey had reached USD 50 million in annual recurring revenue and attracted top global law firms as clients.

… or take  Anysphere, the company behind the  Cursor AI  tool for engineers. It helps programmers write and debug code using natural-language prompts.

Within a year of launch, Anysphere hit “unicorn” status, valued at over USD 1 billion and with fewer than 50 employees.

Neither company trains its own massive AI models. Instead, they rely on existing ones like  OpenAI ’s GPT-4o,  Meta ’s Llama 3, or  Google ’s Gemini—and build innovative solutions customers actually need.

For Professor Litman, these are the businesses investors should be watching. They’re not burning billions on compute power but on proven technology and focusing on solving real-world problems.

Another one of the clearest examples of Professor Litman’s thesis in action is  AppLovin (APP), a company currently held in his firm’s Hidden Alpha portfolio.

AppLovin runs an AI-driven advertising platform called  Axon, which optimizes mobile ad placements using existing AI models. It doesn’t train its own; it simply uses available AI technology to improve marketing results and boost client performance.

When Professor Litman recommended AppLovin in June 2025, it doubled within five months.

By December, it had quadrupled. Since 2023, ad spending on AppLovin’s platform has surged fourfold.

Professor Litman argues this is what the next wave of AI success will look like:  Not infrastructure builders, but practical innovators applying AI to drive measurable business outcomes .

The bottom line? 

Professor Litman’s message is clear, and it’s a contrarian one amid the hype.

Investors chasing the biggest AI models are essentially funding an arms race with uncertain winners.

The smarter move is to look for companies that apply these models in ways that generate immediate, scalable value.

Don’t get caught up in the noise!

The next generation of trillion-dollar companies won’t be the ones building the largest AI models. Rather, they’ll be the ones building with them.

If you’re willing to look beyond the headlines, you’ll find that that’s where the future is being written… one application at a time.

Hope you’ve found this week’s insights interesting and helpful.

Stay tuned for next Wednesday’s The Independent Investor!

Management doesn’t always know best… and this geothermal giant shows us why.

Learn more about  the impact of incentives  in next week’s article!