Wolfe Research stays bullish on equities, sees AI, semiconductors driving H2 gains

Wolfe Research remains constructive on U.S. equities into the second half of 2026, citing easing oil prices, resilient earnings growth, and continued AI investment as support for further gains. The firm continues to favor technology, especially semiconductors, and expects the Fed to hold rates steady rather than tighten, even as U.S.-Iran tensions and higher-for-longer rates remain key risks. It highlighted broad market leadership from tech, with 33 of the top 50 S&P 500 performers this year coming from the sector.

Analysis

The market implication is less about “AI is good” and more about who captures the marginal dollar of incremental compute spend. The clearest winners are the high-leverage infrastructure enablers: NVDA and AVGO should continue to earn upward revision momentum because the spend mix is still skewed toward accelerators, networking, and custom silicon rather than a broad software monetization cycle. That matters because capital is now chasing throughput and latency, not just model size, which keeps hardware demand more durable than a typical capex fad.

The second-order winner is MSFT, but for a different reason: it monetizes AI through distribution and enterprise workflow lock-in, so it should be more resilient if corporate buyers become disciplined and force ROI scrutiny. By contrast, the near-term relative losers are the “aspirational AI” names with less proven unit economics; when management teams shift from experimentation to payback, multiple compression usually starts there first. SMCI and APP can outperform in bursts, but they are more exposed to sentiment reversals and order-timing noise than the platform leaders.

The main risk is that the trade is crowded and therefore vulnerable to a fast factor unwind if oil spikes, rates reprice higher, or there is any disappointment in hyperscaler spending cadence. The timing matters: the next 4-8 weeks are about flow and positioning, while the next 2-3 quarters are about whether 2027 capex estimates keep grinding up. If those estimates flatten, the market may still like AI structurally, but the beneficiaries will narrow sharply and the low-quality beta names could de-rate quickly.

Contrarian view: consensus is treating AI capex as a one-way ratchet, but the more likely surprise is a rotation within the complex rather than a collapse in the theme. That argues for owning the cash-rich platforms and semis with pricing power, while fading the names that need perfect sentiment and uninterrupted risk appetite. In other words, the trade is not “buy all AI”; it is “own the toll collectors, avoid the speculators.”

Key Decisions for Investors

• Long NVDA vs short SMCI for the next 1-3 months: keep exposure to AI infrastructure demand while shorting the higher-beta execution/sentiment name; target a 1.5-2.0x relative move if capex remains firm.
• Overweight AVGO on any 2-4% pullback over the next 2 weeks: best risk/reward in the group if investors continue to prefer custom silicon and networking exposure; use a trailing stop under the prior swing low.
• Pair long MSFT / short a basket of lower-quality AI beneficiaries (SMCI, APP) over 1-2 quarters: monetize the shift from AI enthusiasm to ROI discipline; MSFT should hold up better if enterprise spend gets selective.
• Buy NVDA call spreads 3-6 months out into any volatility spike: structure for continued upward earnings revisions without paying full premium for crowded positioning; exit if hyperscaler capex commentary stalls.
• Avoid initiating fresh longs in APP until there is evidence of sustained fundamental monetization, not just factor momentum: best viewed as a trading vehicle, not a core exposure, with asymmetric downside if risk appetite fades.