The trade has changed.
For years, any company with a credible AI narrative could count on a valuation lift. Broadcom’s networking chips fed hyperscaler buildouts. Palo Alto Networks embedded AI into
its security platform. CrowdStrike made AI-powered threat detection a selling point. Proximity was enough. It isn’t anymore.
Saxo Bank’s analysis published June 4
frames the shift directly: today’s investors are more selective, demanding stronger evidence
that companies are actually positioned to benefit from AI rather than simply exposed to it. That’s a behavioral change, not a cyclical one. It’s also a harder standard to clear –
because “positioned to benefit” requires a company to show a specific mechanism, not just a
product roadmap.
The three companies named represent three distinct AI exposure profiles, which makes the
pattern more interesting than any single name. Broadcom’s AI revenue is infrastructure-level:
it supplies custom AI accelerators and networking silicon to hyperscalers. Its AI story is
real but capital-intensive and dependent on a small number of very large customers. Palo Alto’s AI exposure is product-embedded: it has integrated AI across its security
platform, but investors are asking whether that integration drives pricing power or just
feature parity. CrowdStrike’s case is arguably the most complex, AI is both a core
capability claim and a factor in the competitive pressure it faces from Microsoft and others
who are also AI-equipping their security products.
Broadcom shares have come under pressure in recent weeks, according to market reports,
though specific figures from this package aren’t independently verified and shouldn’t be
treated as confirmed. What’s clear from Saxo Bank’s commentary and corroborating market
analysis is that the narrative alone isn’t holding valuations the way it did in 2023 and
2024.
The real story is this: the AI premium is under audit. The market ran a long experiment in
which AI-adjacent exposure was rewarded almost unconditionally. That experiment is ending. What replaces it is a more granular assessment, how much of this company’s revenue is
directly attributable to AI demand, and is that demand durable or front-loaded by
infrastructure build-out cycles?
That question matters differently for each of the three companies named. Broadcom’s AI
revenue is more visible and more verifiable (it shows in chip shipments and hyperscaler
capex). Palo Alto’s is embedded in a platform renewal cycle that’s harder to isolate. CrowdStrike’s is competitive, it has to demonstrate AI advantage, not just AI presence.
Context: this shift isn’t new in markets, but it’s arriving later for AI than for prior
technology cycles. Cloud stocks went through the same transition around 2022, when investors
stopped rewarding growth-at-all-costs and started asking about path to profitability. AI-adjacent equities are now entering an analogous phase, the framing is shifting from
“is this company in the AI story?” to “what does AI actually deliver for this company’s
economics?”
What to Watch
Watch for: earnings calls from Broadcom, Palo Alto, and CrowdStrike, whenever they fall
in the calendar, for the first hard numbers on AI-attributable revenue versus total revenue. Analyst price target revisions that cite AI monetization clarity (or the lack of it) will
be the leading indicator of whether this selectivity hardens into a structural derating or
resolves into a new growth narrative.
The catch is that the companies most exposed to this scrutiny are also the ones best
positioned to answer it, if their AI integrations are performing. The question isn’t
whether AI is working in their products, it’s whether the market will wait for proof at
the current multiples.