The launch of ChatGPT in November 2022 sparked a frenzy around AI stocks, as illustrated by the sevenfold rise in NVIDIA’s stock price in less than two years. The first phase of the AI cycle, which centered on firms providing infrastructure and those with the most direct AI use cases, is nearing an end. However, we still believe AI will prove to be the biggest productivity enhancer for the global economy since electricity and that the technology opportunity set remains a rich one.

Productivity‑enhancing technologies often fuel speculative asset bubbles. During the dot‑com bubble of the 1990s, the NASDAQ Composite Index—home to most newly listed dot‑com firms—rose from a level of around 750 in early 1990 to a peak of more than 5,000 by March 2000. The index then crashed, plunging 78% by October 2002, contributing to a recession in the U.S.

The rise of AI is not the dot‑com bubble revisited

Parallels are often drawn between the dot‑com bubble and the recent surge in AI, but we believe there are important differences between the two. Unlike the dot‑com bubble, this AI cycle has been driven by a surge in earnings rather than by speculation. Back in March 2000, for example, technology giant Cisco was one of the most popular stocks on the U.S. equity market, with a next 12 months (NTM) price‑to‑earnings (P/E) ratio of more than 125x.

By contrast, and as an example, NVIDIA’s NTM P/E at the end of October was 35X (Figure 1). Wall Street’s estimate for NVIDIA’s full‑year 2026 earnings had increased from USD 0.62 in November 2022 to USD 4.07 at the end of September this year, meaning that its sevenfold share price increase was almost entirely driven by earnings growth estimates rather than by market sentiment alone. Rather than seeing AI as a bubble that is about to burst, we regard it as a multiyear investment cycle in which the initial period of incredibly rapid growth is now giving way to a period of moderating, yet still impressive, growth.

Another major difference between the AI investment cycle and the dot‑com bubble is the funding source for much of the infrastructure spend. During the 1990s, the fiber infrastructure was primarily debt funded, issued by companies such as WorldCom. Today, NVIDIA’s growth has been driven by selling the linchpin AI technology to some of the most cash flow‑generative companies in history (Microsoft, Google, Amazon, Meta). The key question is whether NVIDIA can retain its dominant status as we potentially go from an estimated USD 45 billion AI chip market in 2023 to an estimated USD 500 billion AI chip market in 2028,¹ or will it be challenged by new competitors.

Innovative linchpin companies offer strongest growth prospects

Investors looking to navigate the next phase of the AI investment cycle responsibly should seek to identify those key linchpin companies that are innovating within secular growth markets. Improving fundamentals are also crucial—firms with accelerating revenues, increasing operating margins, and/or improving free cash flows are worth looking at closely. Finally, it is important to ensure that any valuations paid are reasonable.

When analyzing AI’s impact on the wider global technology field, investors should include the semiconductor industry, where new kinds of semiconductors are being designed for AI applications. Also, AI is being used to improve the existing chip design and manufacturing processes. Among software firms, data infrastructure companies, vertical application vendors, and cybersecurity vendors are well placed to capitalize on the advancements in AI. Finally, in financial technology (fintech), generative AI is being used to improve the customer experience in areas such as personalized banking, fraud detection, and credit risk assessment.

Key takeaway
Investors seeking to navigate the next phase of the AI investment cycle should look for key tech firms that are innovating within growth markets.