Stop Piling Into the AI 'Model Wars'

Big Tech companies have dominated the AI conversation for years...

Every headline focuses on who has the most powerful model... Is it OpenAI's ChatGPT, Alphabet's (GOOGL) Google Gemini, or Anthropic's Claude?

It's hard to say, but the overall goal is clear. Each company is racing to build faster, bigger, and smarter AI models. Yet the industry is starting to shift its focus...

Venture capitalists are pouring cash into smaller, nimbler companies building innovative tools on top of existing AI infrastructure.

And it's paving the way for a new batch of winners in Silicon Valley – companies that create AI-powered applications for end users, rather than foundational models.

Today, we'll explain why investors should stop focusing solely on AI infrastructure and start paying attention to the companies using this technology to create real, scalable products.

The first tech boom followed a similar playbook...

Back in the 1990s, investors flocked to infrastructure stocks like Cisco Systems (CSCO). Its switches and routers powered the backbone of the Internet.

Cisco reached a $555 billion market cap and briefly became the world's most valuable company. But it wasn't the biggest winner in the tech space.

When Amazon (AMZN) entered the scene, it used Cisco's infrastructure to build a simple e-commerce site on top of the Internet hardware.

This compelling consumer experience made Amazon a trillion-dollar titan. A similar dynamic is now unfolding in the AI space...

Companies like OpenAI and Alphabet are locked in an arms race...

They're spending billions of dollars on cloud infrastructure, research and development, and compute power.

Yet the payoff remains uncertain. Application-focused startups, on the other hand, are much leaner, faster, and more practical...

Legal-tech firm Harvey, for one, created an AI product that streamlines contract reviews and litigation support. These tasks once cost lawyers hundreds of dollars an hour. The startup earned $50 million in annual recurring revenue in 2024. And it already counts major law firms among its customers.

Anysphere, which launched the Cursor AI tool in 2023, helps engineers write code using natural-language prompts. The company hit "unicorn" status (a $1 billion valuation) with fewer than 50 employees.

These startups don't train their own models. They use what's already out there – like OpenAI's GPT-4o... Meta Platforms' (META) Llama 3... Alphabet's Google Gemini. And they build AI-powered tools customers actually need.

The true AI winners will deliver tools that solve real-world problems...

These companies are developing products and services on top of powerful AI models. So they don't need to worry about costly infrastructure. That ensures lower costs and a faster time to market.

Current Hidden Alpha holding AppLovin (APP) is a perfect example of that...

This company has a simple business model… AppLovin offers an advertising platform for mobile apps. It uses AI to train its advertising software (called Axon) to improve ad placements based on existing models.

We recommended AppLovin last June, and five months later, the stock doubled. By December, it had quadrupled. Since 2023, AppLovin has also quadrupled ad spending on its platform.

This is just one example of how AI-powered apps are revolutionizing the tech industry. Investors who get in early on these winners stand to benefit the most from the AI era.

Regards,

Joel Litman
August 6, 2025