Bernstein reiterates Richemont stock rating on likely sales upside By Investing.com

Bernstein reiterated an Outperform rating on Richemont with a CHF200 price target and said the company could surprise to the upside in Q4 fiscal 2026 constant-currency growth, especially in Japan and Europe. The note cites strong fundamentals, including 66% gross margins and 7.2% trailing 12-month revenue growth, while consensus still assumes 9% group constant-currency growth. BofA also upgraded Richemont to Buy, but cut its target to CHF175 from CHF190.

Analysis

This is less a single-stock luxury note than a read-through on how fragile the AI trade has become when a macro headline can simultaneously hit semis, data-center capex proxies, and index momentum. NVDA is the obvious barometer: even modest policy uncertainty around chip flows or tax treatment can force systematic de-risking because the name is crowded, option gamma is high, and ownership is still tightly linked to the “AI spend is linear” narrative. The second-order effect is that suppliers and adjacent hardware names typically absorb a larger percentage drawdown than the direct policy target because investors immediately price in delayed procurement, not just margin impact.

The market is likely over-anchoring on near-term sentiment while underestimating how quickly this can reverse if policy language stays aspirational rather than executable. For semis, the key variable is not whether rhetoric is hawkish, but whether it alters enterprise spending decisions over the next one to two budget cycles; if not, the move becomes a positioning event rather than a fundamentals event. That creates an attractive setup for a bounce in the highest-quality AI beneficiaries if the headline risk fades, especially because the broader selloff can force short covering once dealers re-hedge.

Contrarian view: the selloff may be justified as a reset in multiple compression rather than a true demand shock. After a multi-quarter rerating, the market is telling you that incremental bad news now has more impact on valuation than on earnings, which means the best risk/reward is usually not outright long-beta but a relative-value expression versus semis with weaker pricing power or lower AI exposure. If the tax scare evolves into real capex friction, the underappreciated loser is the equipment and memory chain, where order timing is more elastic than for leading-edge compute.

Key Decisions for Investors

• Buy NVDA 1-3 month out-of-the-money calls on a 1-2 day post-selloff stabilization; target a 2-3x payoff if the policy headline fades and crowded shorts cover, but keep premium small because implied vol should remain elevated.
• Put on a pair trade: long NVDA / short a weaker AI-linked semi or equipment name over the next 4-8 weeks; the long leg benefits from best-in-class earnings durability, while the short leg captures the higher downside if capex reprices.
• If the selloff extends another 3-5% without new policy follow-through, add to quality semis on weakness rather than chasing the first rebound; the best entries usually come after forced selling, not on the headline day.
• Use a tight stop on any outright long NVDA position if the market starts pricing in actual procurement delays rather than tax rhetoric; that would convert this from a multiple event into a fundamentals event.
• For hedged portfolios, reduce exposure to second-derivative beneficiaries of AI capex in the supply chain for the next 1-2 quarters, as they are more vulnerable to order pushouts than the mega-cap platforms.