Artificial Intelligence: From Hype to Proof

Written by the William Allan team · Published 2025-09-11

Artificial Intelligence (AI) has dominated the headlines for the past two years. From boardrooms to dinner tables, few topics have captured as much attention — or driven as much market enthusiasm. Large technology companies have committed substantial sums to AI infrastructure, attracting investor attention and raising expectations. The investment case still requires company-specific analysis of costs, financial results, and valuation.

But in 2025, the story is shifting. Investors are beginning to move beyond the initial stage of broad optimism — the “sounds good, I’m in” mentality — toward a more measured demand for tangible results. This transition is critical to understand, because it has significant implications for the way capital is deployed and how companies will be judged going forward.

Looking Back: Lessons From the Dot-Com Era

This isn’t the first time investors have been swept up in the promise of transformational technology. The late 1990s saw a similar wave of enthusiasm during the rise of the internet. Companies raced to establish websites, build infrastructure, and stake their claim in the digital world. Firms like Cisco supplied the routers and switches that powered internet adoption, and valuations climbed rapidly.

Of course, not every company survived. The dot-com bubble burst when investor expectations exceeded what businesses could deliver in the near term. Yet, the internet itself didn’t disappear. Instead, it evolved, and the companies that integrated digital capabilities into their models reshaped entire industries. Retailers, financial firms, healthcare providers, and countless others leveraged technology to become more efficient, more connected, and more profitable over the following decades.

The lesson? Hype alone does not build durable business models. But when real-world adoption takes root, the long-term impact can be transformative.

Where AI Stands Today

AI today may be at a similar turning point. In the early phase, the focus was on infrastructure: data centers, chips, and cloud computing capacity. Companies such as Nvidia have become widely recognized for their role in supplying the hardware capable of running AI applications at scale. For example, Nvidia’s August 2025 earnings release reported continued data-center revenue growth. That is evidence of demand for infrastructure, not proof that every customer will earn an attractive return on its AI spending.

Now, however, investors are beginning to ask a tougher question: where are the profits? The excitement over AI’s potential must translate into measurable results — improved efficiency, new products, and real returns on capital.

We believe this next phase — moving from hype to proof — is a healthy step. It separates durable opportunities from speculative excess.

Beyond Tech: AI’s Broader Impact

While most attention has centered on technology firms, history suggests the long-term impact of AI will extend far beyond Silicon Valley. Just as the internet reshaped industries from retail to transportation, AI has the potential to touch virtually every business.

Consider a few possibilities already emerging:

Healthcare: Faster drug development through AI-powered modeling could lead to improved health outcomes.

Education: Personalized learning tools can adapt content to each student’s pace and style.

Consumer Goods: Companies can use AI to refine supply chains, forecast demand, and reduce waste.

Service Industries: Fast-food restaurants are testing AI at drive-thrus with the aim of improving efficiency and changing how employees spend their time. Results depend on the implementation.

These examples highlight an important point: the most significant winners of the AI era may not be the companies building the tools, but those that find the most effective ways to use them.

Why This Matters for Investors

Periods of technological change often spark periods of volatility. Valuations swing as investors try to gauge which companies will emerge as leaders. In the process, markets can overshoot in both directions — first with excessive optimism, and later with excessive pessimism.

For long-term investors, volatility can create opportunities, but it also exposes portfolios to losses and requires attention to diversification and cash needs. When a price falls, disciplined investors can reassess the business and its valuation. A lower price alone does not establish that an investment is attractive or suitable.

William Allan may use individual companies as well as funds. Direct ownership can provide position-level control and selectivity, but it also creates company-specific risk, concentration, trading costs, and the possibility that an investment thesis is wrong. Funds may offer broader diversification and operational efficiency. Neither structure assures better returns or lower total cost.

The Takeaway

AI is not a passing trend. We believe it represents a structural shift, much like the internet revolution before it. But as with every major technological wave, the challenge for investors is separating hype from reality.

The coming years will likely bring both excitement and disappointment as companies prove — or fail to prove — their ability to use AI to create lasting value. For us, the goal is clear: remain disciplined, focus on businesses with strong fundamentals, and invest when valuations make sense.

That approach has guided us through many periods of change, and we believe it will continue to serve investors well as AI moves from promise to proof.

Companies and securities mentioned are examples for discussion only, are not recommendations to buy or sell, and may or may not be held in client accounts. Any investment decision depends on the investor's objectives, risk tolerance, time horizon, tax situation, and other circumstances.

This commentary is for informational purposes only and is not investment, tax, or legal advice. All investing involves risk, including the possible loss of principal, and past performance does not guarantee future results.