Navigating the AI Exuberance Without Losing Our Bearings

At Nexus, we’re excited about AI’s potential but remain grounded – investing with discipline, paying attention to valuation, and maintaining diversification.

Stock Market Euphoria: When AI Mentions Move Billions

The euphoria around artificial intelligence has reached new heights, not just in Silicon Valley, but on Wall Street too. Share prices of companies touting AI-related deals or exuberant forecasts are soaring. Recently, Oracle’s stock surged 36% in a single day after it reported a significant increase in orders for its technology used to power AI models. The majority of Oracle’s new orders are expected to be driven by one company – OpenAI, the maker of ChatGPT. Similarly, the share price of AMD, a semiconductor chip maker, jumped 24% in a day after announcing a major chip supply agreement, again with OpenAI. These aren’t small speculative companies, they’re large, well-established businesses with market values in the hundreds of billions. Moves of this magnitude in a single day are rare.

Financial Creativity Returns: When Suppliers Fund Their Own Demand

It’s not just the share price moves that are unusual – the structures of these deals are unusual too. AMD has given OpenAI the right to buy AMD shares in the future for a token price (about one cent per share) if certain technical or business goals of their deal are met. Giving shares to a customer for the privilege of selling your products? That’s a head-scratcher. Similarly, Nvidia has committed an eye-popping $100 billion equity investment in OpenAI over several years, which it will likely use to buy Nvidia chips. Effectively, Nvidia is financing its own sales. Look a little deeper, and the creativity goes further. New companies like CoreWeave and Nebius Group will build and rent large AI data centers, running on Nvidia systems. That isn’t odd, except that Nvidia holds major equity stakes in both. At the same time, the incumbents are using special-purpose vehicles to fund massive data center projects and keep the debt off their balance sheets. When suppliers start backstopping their own demand, you know exuberance is morphing into creative deal engineering, reminiscent of the vendor-financing deals of the late 1990s.

The Trillion-Dollar Question: Can OpenAI Afford its Ambition?

Not only are these deals creative, but the scale of these ambitions defies belief. The Financial Times estimates that OpenAI has committed to spending over $1trillion dollars on data center capacity, centered around its deals with Oracle, Nvidia, and AMD. OpenAI today has roughly 800 million weekly users, 5% of whom are paid subscribers, $13 billion in annualized revenue, and an $8 billion loss in the first half of 2025. That begs the question: how does OpenAI fund ambitions of that size?

Add to that the ambitious plans of the other technology heavyweights like Meta, Microsoft, and Alphabet, and total global data center spending could reach up to $3 trillion by 2029. AI-related capital expenditure has already grown so large that it accounted for nearly half of U.S. GDP growth in the first half of this year, while much of the rest of the economy has been merely muddling through.

Echoes of the 1990s: When Capacity Outran Reality

A few surmise that this extraordinary investment in data centers will make AI models smarter, creating new use cases of them. The thinking goes: build enough supply, and demand will eventually follow. And the first one to get to the smartest AI model will reap most of the benefits. Right now, the demand doesn’t match the hype, and the return on investment is difficult to envision. Then there’s the practical challenge of execution – these data centers require vast amounts of electricity, skilled labour, and advanced chips, the latter mostly made in Taiwan, layering on geopolitical risk. To us, it appears a risky bet, one that could prove rewarding for some but costly for others if expectations don’t materialize.

At Nexus, we’re optimistic about the long-term potential of AI. It is a transformative technology, and we’ve begun integrating AI tools into our daily work to improve our productivity. That said, we’re not sure things will move quite as fast as the market seems to think. During the internet boom of the 1990s, the fiber-optic buildout yielded enormous benefits, but it took years for that capacity to be fully used. The same was true during the “Railway Mania” of the 1800s, when capacity expansion outpaced economic reality. We think something similar may be happening again today.

Our Philosophy: Patience, Prudence, and Preservation of Capital

As such, we continue to follow our investment discipline – investing in high-quality businesses at reasonable valuation and keeping portfolios well diversified. Today, the U.S. equity market has become heavily concentrated around a single theme: AI. The same eight large technology stocks now make up nearly 38% of the S&P 500 and 70% of the NASDAQ, creating meaningful concentration risk. By contrast, as of October 15, 2025, the top eight stocks in the Nexus Equity Fund accounted for 26% of the portfolio and spanned a broad range of sectors. As such, our portfolios aren’t concentrated around a single theme.

As long as AI stocks keep marching higher, index returns will look strong because of that concentration. But when momentum fades, investors may face real losses – a risk we’re not prepared to take. That doesn’t mean we’ve avoided AI altogether. We continue to own companies such as Meta, Microsoft, and Alphabet, all AI beneficiaries. But, our investments in them were made for their strong core businesses that generate meaningful free cash flow, and were made long before the current excitement began.

Investors today face a unique challenge: balancing enthusiasm for AI with the realities of market cycles.

While technology adoption can drive exceptional growth, history reminds us that periods of rapid innovation are often followed by corrections. By focusing on companies with strong fundamentals, diversified revenue streams, and proven management teams, investors can participate in AI’s long-term potential without being swept up in short-term exuberance. This approach allows portfolios to benefit from innovation while minimizing exposure to speculative excess.

It’s hard to predict when a boom turns into a bubble and we don’t plan to guess. The rally in AI stocks may continue, and if it does, our portfolios may trail the major indices. That’s perfectly fine with us. Our focus has always been on delivering strong absolute returns that help our clients achieve their financial goals – not on chasing index performance, especially when doing so could put your capital at risk. We’re fortunate to have clients who understand our disciplined approach and share our conviction that lasting wealth is built through patience and prudence.