What to buy now with the market at all-time highs, according to UBS

UBS highlighted two stock baskets for investors despite market highs: inflecting fundamentals and quality laggards, with Broadcom and Accenture among the named opportunities. Broadcom’s forecast CFROI is 77.1% and is expected to exceed 80% in 2027, while Accenture’s CFROI is projected at 43.7% despite a 34% YTD decline and a record valuation discount. The note also flagged Nvidia, Micron, Palantir, Reddit, Microsoft, Adobe, Take-Two Interactive and Salesforce as screened ideas.

Analysis

This setup is less about chasing momentum and more about exploiting a valuation gap created by investors over-penalizing business models with visible AI exposure. The market is paying for AI beneficiaries that can show near-term earnings beta, but it is still discounting the second-order monetization layer: ASIC/infra enablers like AVGO that compound through inference demand, and high-quality services/enterprise software names where AI may compress headcount but can also increase wallet share, implementation complexity, and switching costs.

The bigger signal is that quality is becoming a factor again, but only where expectations remain low. AVGO’s rerating can continue because its earnings path is tied to a multi-year capex cycle, not just one product cycle; that means upside should persist as long as hyperscaler and networking spend stays elevated. By contrast, ACN looks like a classic sentiment dislocation: the market is pricing an AI-driven margin reset before the evidence exists, creating a favorable asymmetry if enterprise AI adoption is slower to disintermediate than feared.

The main risk is timing mismatch. These are not “days” trades; the catalyst window is more likely 1-3 quarters for multiple expansion and 12+ months for fundamental inflection to be recognized. If AI spend pauses or enterprise customers begin delaying transformation budgets, AVGO could de-rate on forward growth deceleration, while ACN remains vulnerable to a narrative-driven de-rating even if fundamentals hold. The consensus is likely underestimating how long it takes for AI to move from prototype to budget-line-item displacement, especially in consulting and workflow-heavy software.

Key Decisions for Investors

• Long AVGO on any 3-5% pullback; hold 3-6 months. Risk/reward favors continuation if AI infrastructure spend remains constructive, with downside primarily if hyperscaler capex inflects lower.
• Build a long ACN position in 3 tranches over 4-8 weeks; 6-12 month horizon. Use it as a mean-reversion trade on sentiment overhang, with the caveat that the stock can stay cheap until the market sees stable bookings.
• Pair trade: long AVGO / short a basket of over-owned AI beneficiaries with less visible cash-return support (e.g. NVDA or PLTR) for 1-2 quarters. The thesis is relative rerating toward names with stronger embedded expectations and better quality momentum balance.
• Buy ACN call spreads 6-9 months out instead of stock if you want convexity to a sentiment reversal. This limits capital at risk while capturing a potential re-rating if AI-disintermediation fears prove premature.
• Avoid initiating new shorts in MSFT/ADBE/CRM here; the crowding risk is high and the better expression is relative long ACN versus the group, not outright bearishness on enterprise software.