UBS reiterates Nvidia stock buy rating on AI demand strength By Investing.com

UBS reiterated a Buy on Nvidia with a $245 price target versus a $221.23 share price, citing still-elevated AI-related export demand despite a 28% month-over-month pullback in Taiwan April AI server exports to $23.5 billion after March's outsized rebound. The firm also noted first-quarter exports of $70.9 billion, up 7% quarter-over-quarter, while TSMC first-quarter sales rose 8.4% quarter-over-quarter and beat seasonality. Nvidia also announced a board expansion and new director appointment, alongside strategic AI infrastructure partnerships and Japan server discussions.

Analysis

The market is treating this as a benign “AI demand intact” confirmation, but the more important signal is that the supply chain is normalizing without losing altitude. That is usually the best setup for semis: it removes the risk of near-term inventory overhang while preserving enough end-demand momentum to keep estimate revisions drifting higher. The cleaner read is not just NVDA strength, but that the capex pulse is broadening from hyperscaler buildouts into adjacent infrastructure names, which supports second-order beneficiaries in networking, foundry, and AI power/cooling.

UBS’s commentary also implies a subtle rotation risk inside the AI complex. If export data are high but less explosive, the multiple expansion trade becomes more selective: investors will likely pay up for companies with direct monetization, software attach, or strategic capacity constraints, while lower-quality “AI exposure” names may lag even if the thematic remains strong. That creates an asymmetry where NVDA can keep grinding higher, but the breadth of the trade may narrow, leaving room for pairs against names that have rallied on narrative rather than bookings visibility.

The IREN partnership is strategically interesting because it converts a speculative compute story into something closer to a funded industrial rollout, and that can force repricing across smaller AI infrastructure platforms. However, the upside comes with execution and financing risk: energy access, regulatory approvals, and customer concentration can all become binding constraints over the next 6-18 months. The contrarian takeaway is that the market may be underestimating how much capital intensity this AI infrastructure cycle demands; winners will be those with the cheapest cost of capital and the strongest balance sheets, not necessarily the fastest revenue growth.

Near term, the tape is vulnerable to macro inflation data more than to this micro news. If rates back up, the market may start discounting the duration-heavy AI cohort, even as fundamentals remain healthy, which would create a better entry point than chasing strength here. In other words, the bull case is intact, but the risk/reward has shifted from “buy everything AI” to “own the highest-quality enablers and fade the weaker secondaries.”

Key Decisions for Investors

• Stay long NVDA into the next 1-3 months, but use pullbacks tied to inflation/rates spikes as entries; upside remains intact, but chasing at 52-week highs compresses reward/risk unless revenue estimates reaccelerate.
• Pair trade: long NVDA / short a basket of lower-quality AI infrastructure proxies over 1-3 months; if the theme narrows, quality should outperform while narrative-heavy names lag.
• Buy TSM on any post-data consolidation over the next 2-6 weeks; the read-through from stable leading indicators supports wafer demand, with less valuation risk than the highest-multiple AI beneficiaries.
• Speculative long IREN only as a defined-risk position over 6-12 months; the partnership is a validation event, but financing/regulatory/execution risk argues for calls or a small equity starter rather than outright size.
• Avoid shorting the AI complex outright; instead hedge duration sensitivity with a partial short in long-duration growth or rate-sensitive indices if inflation data surprise hotter, since macro is the cleanest reversal catalyst.