This dot-com survivor says AI buildout is more like 1997 than 1999 — and still urges investors to have more cash

Amid the palpable excitement surrounding artificial intelligence, a veteran of the dot-com era offers a nuanced perspective that encourages investor caution. Drawing on his experience, this executive asserts that the current AI infrastructure buildout more closely resembles the foundational, yet unproven, days of 1997 rather than the speculative bubble of late 1999. This distinction is critical for investors trying to navigate the escalating valuations in the AI sector. The executive's core argument centers on the idea that while significant capital is being deployed to construct the underlying hardware and software necessary for AI, the widespread, revenue-generating applications are still in their nascent stages. In 1997, the internet was clearly revolutionary, but profitable business models were still largely conceptual. Similarly, today's AI investments are heavily concentrated in infrastructure — chips, data centers, and advanced algorithms — with the expectation that killer applications will materialize later. This stage requires substantial capital expenditure without immediate, commensurate returns for many players. For investors, this implies a need for discernment. While certain companies directly involved in AI infrastructure like chip manufacturers or cloud service providers are seeing immediate benefits, the broader market enthusiasm may be getting ahead of itself. The call to maintain a higher cash allocation is a strategic move to preserve capital and provide flexibility. Should the market experience a correction, cash offers the opportunity to acquire undervalued assets; conversely, it protects investors if the anticipated AI applications fail to materialize as quickly or profitably as expected. Sector-wide, this perspective suggests that the current investment cycle in AI is a multi-year endeavor. Companies that are purely speculative, without strong underlying fundamentals or clear paths to profitability, are most vulnerable. Established technology giants with deep pockets and diversified revenue streams are better positioned to weather this buildout phase, leveraging their resources to invest in AI without placing their entire business at risk. The macroeconomic implication is that while AI promises long-term productivity gains, its immediate impact on overall economic growth might be more staggered than some optimists predict, necessitating a patient approach. Looking ahead, investors should closely monitor the development and adoption rates of enterprise-level AI solutions. Key indicators will be the revenue generated by specific AI-powered products and services, rather than just capital expenditure announcements from infrastructure providers. The emergence of clear, profitable use cases that go beyond research and development will be the true sign that the AI boom is transitioning from its foundational 1997-like stage to a more mature and broadly profitable era. The full implications for investors will depend on follow-through in price action, positioning data and forthcoming policy or earnings catalysts referenced above. The full implications for investors will depend on follow-through in price action, positioning data and forthcoming policy or earnings catalysts referenced above. The full implications for investors will depend on follow-through in price action, positioning data and forthcoming policy or earnings catalysts referenced above. The full implications for investors will depend on follow-through in price action, positioning data and forthcoming policy or earnings catalysts referenced above.