Key insights
- The author argues that current tech stock analysis often overlooks the critical factor of sustained enterprise-level AI adoption. The focus should be on AI's value proposition and use cases, not just current financials or valuations. If AI demand doesn't materialize, even companies with strong cash flows or those supplying AI infrastructure could face significant stock price declines, as current high growth expectations may not be met.
I see so many posts discussing if they should buy MSFT, NVDA, MU , Intel, TSM, AMD etc, but almost none discussing the thesis behind these companies: will there be sustained AI adoption at the enterprise level.
In my admittedly novice opinion, investors in tech should be spending their time figuring out that capabilities of AI models, their use cases, how much value they are delivering to companies etc. It doesn't matter what kind of margins or cash flow they have right now, or if MSFT "looks cheap". If MSFT is spending all of its free cash flow on AI data centers and demand doesn't materialize, their stock price is fucked. Even the picks and shovels companies, sure they get paid right now either way, but even the lower PE names like TSM/NVDA are being priced for ~18-20% fcf growth for a decade. Sure you can bet that the cap ex cycle can continue for a while without strong AI adoption, but not indefinitely.
What do you guys think are the best sources of information for this kind of stuff? I've been reading semi analysis, academic studies, mckinsey reports, looking at gov procurement and darpa.