Alex Karp Grouped Palantir With These 3 Unstoppable Stocks as the Only True Artificial Intelligence (AI) Infrastructure Winners

Palantir CEO Alex Karp is cited as placing Palantir among AI infrastructure leaders alongside Nvidia, Micron, and SK Hynix, using the Rule of 40 (revenue growth + operating margin). Palantir’s Rule of 40 score rises from 83% in Q1 2025 to 145% one year later, attributed to accelerating AIP-driven commercial adoption and expanding operating margins. The article is mildly cautionary on valuation for Palantir (described as “richer” with less margin of safety) while highlighting Nvidia as the most compelling risk-reward among the four.

Analysis

The market takeaway is less “AI is broadening” than “AI capex is still being monetized most cleanly by the picks-and-shovels layer.” That keeps NVDA and, to a lesser extent, MU/SKHY as the highest-conviction beneficiaries because their revenue is tied to already-committed infrastructure spend, while PLTR remains a longer-duration story dependent on budget conversion and deployment depth after the initial pilot. In a risk-off tape, software names with rich multiples tend to de-rate faster than the suppliers selling into non-discretionary buildouts.

Second-order, the biggest spillover is not into other AI software vendors but into adjacent infrastructure and component bottlenecks: memory, networking, power, and packaging stay constrained longer if hyperscalers keep adding racks. That is supportive for MU and SKHY over the next 1-3 quarters, but it also raises the chance that incremental capex shifts from “must buy” to “optimize what we already bought,” which would cool the operating leverage narrative in 6-18 months. The most vulnerable peer set is the long-tail of AI application companies where customer adoption is still proving out and valuation is already pricing a multi-year bull case.

Contrarian view: the article treats Rule of 40 as a durable moat, but for cyclical suppliers it can peak just as the market starts paying up for it. For PLTR specifically, the consensus risk is underestimating how quickly growth expectations can compress if enterprise buying slows even modestly; for NVDA/MU, the consensus may be over-discounting how long supply tightness can support margins before capacity additions normalize pricing. The key falsifier is any evidence that hyperscaler capex growth or enterprise AIP conversion decelerates over the next two earnings cycles.

Key Decisions for Investors

• Prefer NVDA over PLTR on a 1-3 month horizon: use a long NVDA / short PLTR pair if AI spend stays intact but the market continues to punish higher-multiple downstream software. Best entry is on PLTR strength after any positive commentary; thesis fails if PLTR growth re-accelerates without multiple compression.
• Add MU on pullbacks for a 6-12 month HBM/DRAM tightness trade. Risk/reward improves if memory pricing and gross-margin guidance keep inflecting; invalidate if management signals utilization or ASP normalization earlier than expected.
• Use PLTR as a watch-list short rather than a fresh outright short unless there is evidence of slower commercial bookings or customer expansion. The stock is most vulnerable when the market shifts from celebrating efficiency to demanding evidence of durable seat expansion.
• If you want broader AI beta with less single-name risk, favor NVDA over a basket that includes downstream software. The structural upside is in the supply-constrained layer; the downside is slower but less severe unless capex rolls over.
• Set an alert on hyperscaler capex commentary and enterprise software renewal trends over the next two earnings seasons; that is the cleanest falsifier for the whole stack trade.