Why there’s a “huge vibe divergence” between tech and finance on AI

Not since Biggie Smalls and Tupac has there been an East Coast versus West Coast schism like the current perception gap over artificial intelligence.

On Wall Street, the only type of AI exposure that’s worked lately is owning companies benefiting from shortages (memory chips) or ones that will benefit from expanding capacity of these scarce resources (semicap equipment). The theme has been a net negative for the market this year, based on how much software stocks thought to be at risk of severe disintermediation by AI have slumped.

In Silicon Valley, commentators and VCs are pounding the table that AI has now proven its ability to transform computer-based work, thanks in large part to recent progress from Anthropic. That is, AI has gotten better at doing things — things that save time and have commercial value.

This apparent discrepancy is primarily being recognized by tech types who claim their counterparts in finance are short-sighted and fail to appreciate the scale of recent breakthroughs.

But AI is certainly asking a lot more of investors.

It’s asking them to forgive a lack of free cash flow generation as money gets piled into massive capex instead of buybacks.

It’s asking them for a lot more money, both through debt issuance in private and public markets and, soon, a heck of a lot of equity supply from the likes of SpaceX/xAI, Anthropic, and OpenAI.

The valuations of these privately held companies have increased by about $550 billion since September.

It would be somewhat disingenuous for AI boosters to say that all these privatized tech gains, and the underlying progress upon which they’re based, wouldn’t also be someone’s pain.

The Takeaway

Not every great company is a great stock where it’s priced, and not every great stock is a great stock all the time. In the short term, initial conditions and positioning matter more than fundamentals in explaining price action. The hangover from capex binges tends to be a very difficult period for the big spenders (see: the aftermath of shale investment or the dot-com bubble) — and that’s not an indictment of the technology. If AI is a flop in ROI terms, it would not be the first time Silicon Valley vastly overestimated the commercial, real-world applicability of a new technology.