CNBC Halftime Report
2026-07-02 · Hosted by Scott Wapner · CNBC
Executive Summary
The committee debated the second-half setup as the Dow logged its best first half since 2021 and the Russell 2000 its best first half since 1991 (+22% YTD). Despite semiconductors down ~4% and AI infrastructure off near double digits, the broad market held — the Mag 7 (Meta, Microsoft, Apple, Tesla) stepped up to blunt the decline, epitomizing 2026’s back-and-forth rotation. Meta jumped ~10% on its cloud/compute plan, which Steve Weiss called the answer to “how are they going to generate return on capex.” The panel split on whether the market has genuinely broadened (Belski’s small-cap and financials case) or is still fundamentally an AI trade (Weiss: 42 of the S&P’s top 50 stocks are AI-linked). Goldman sees double-digit S&P earnings growth persisting through 2027 (consensus ~24.3%), supporting targets near/above 8,000.
Key Stories & Changes
1. The Second-Half Setup & Rotation Debate
Dow best first half since 2021; Russell best first half since 1991; YTD: Dow +9%, S&P +10%, Nasdaq +13%, Russell +22%
S&P 500 hasn’t had a negative July since 2014; NASDAQ 100 up 17 of last 18 Julys (avg +4.23% since ‘08 per BTIG’s Krinsky)
Joe Terranova: even with semis/AI infra down, Mag 7 laggards (Microsoft, Meta, Apple, Tesla) picking up the slack; expects rotation to continue, not a regime change
Steve Weiss: broadening is overstated — of the top 50 S&P names, only 8 aren’t AI-associated (energy, industrials like Caterpillar now count as AI plays)
Brian Belski: genuine broadening — financials +double digits, small-cap +14% and Russell +22% in Q2, small-cap earnings +28% for 2H
$2.3T of Mag 7 market cap lost in June (biggest ever), raising the question of why July seasonality hasn’t been front-run
2. Meta’s Cloud Business
Meta +~10%, best day since January, on plans to build out cloud to sell excess AI compute
Steve Weiss: answers the #1 question he got — how Meta generates ROI on capex (“Now they have cloud”)
Neo-clouds Nebius and CoreWeave fell on the news; Belski owns Nebius (up ~200% 1H), calls it a one-event reaction with a strong balance sheet
Terranova cited the SpaceX-Anthropic ~$20B compute rental deal as a template; sees a “paradigm shift” and proof point for Meta’s spend
3. Financials & Private Equity
KBE bank ETF +12.4%, best first half since 2021; committee buying financial underperformance (S&P Global +7%, exchanges ICE/CME/NASDAQ, Wells Fargo all higher)
Belski: banks still “massively under-owned,” expects consolidation; likes both large- and small-cap banks
Private equity/BDCs the exception — Brent Talkington sold Apollo (~$128); cites lack of monetizations, fund gates (Apollo capped draws at 5%), and negative sentiment; Weiss avoids until a peer/sentiment turn
4. Nike & Athletic Wear
Nike +4% after beating estimates (had been down 12%); no committee member owns it
Weiss: too much competition, turnaround taking too long, “stuck in the mud”; Terranova (as a trade) sees it washed out, could work against the ~$40 June low toward the 200-day (~$57)
Preference for Dick’s Sporting Goods (target $260) and Lululemon as better second-half stories
5. Best Ideas & Calls of the Day
B of A Q3 picks: IBM, Spotify (owned), Walmart (owned, technically weak on soft near-term trends), Nucor, Snowflake, Lennar
BTIG 2H picks: AppLovin (+10%, “reasonable valuation”), Edwards Lifesciences, Expedia, Prologis (owned)
Palo Alto Networks: best quarter ever, +112% in Q2; Terranova sees a “mega winner,” Weiss says already getting more credit than deserved
Lockheed Martin upgraded to Buy at Citi (defense names “not working”; Terranova skeptical — “hope is not a strategy,” but Talkington calls Lockheed the “crème de la crème,” 27-year dividend grower)
Other calls: CBRE (Belski bullish, data-center transactions), Simon Property downgraded at Wolfe (owned/held), American & United Airlines targets raised at Bernstein, Old Dominion upgraded
Trends Identified
1. Rotation Is the Defining Feature of 2026, Not a Trend Break
Joe Terranova framed the year as the market “vacillating back and forth” — early-year cyclical broadening, then Mag 7 leadership during the Middle East conflict, then a June reversal where the Mag 7 suffered. Trying to time the peak “will frustrate the heck out of you”; the setup favors owning both the S&P equal-weight and market-cap indices rather than picking one.
2. “It’s Less About Sectors and More About Stocks”
Weiss argued the second half will be stock-specific rather than sector-driven — Microsoft, Google, and Apple’s individual second-half positioning matters more than headline sector labels, since even industrials (Caterpillar), energy, and financials are increasingly viewed through an AI lens.
3. Meta’s Cloud Move Validates AI Capex Returns
The panel treated Meta’s compute business as more significant than a new revenue line — it supplies the “proof point” for enormous AI capex, echoing the SpaceX-Anthropic rental model and reframing the ROI question that dogged Meta after earnings.
4. Small-Caps and Financials Broadening (Contested)
Belski’s core thesis — vindicated Q2 small-cap and financials outperformance with attractive price-to-free-cash-flow and ROIC — points to genuine broadening. Weiss’s counter (top-50 names are AI-linked) frames this as the central second-half debate. —-
Sentiment Analysis
Overall Market Sentiment: Bullish
Few on the desk are bearish; the debate is about the mechanism of continued gains (broadening vs. AI concentration) rather than direction.
Risk Factors Highlighted
Hidden AI concentration: Weiss warns the market is more AI-dependent than indicators suggest — a peak in AI fundamentals would hit broadly.
Semiconductor/memory pullback: Semis down ~4% and DRAM at the 21-day moving average; a break lower threatens the momentum trade.
Private equity overhang: Fund gates (Apollo’s 5% cap) and lack of monetizations weigh on PE/BDC names.
Communication services weakness: Netflix and AT&T dragging; only Google strong in the sector.
Nike competitive erosion: Loss of footwear share to On, Hoka, Adidas makes a turnaround hard to time.
Defense-name underperformance: Lockheed/RTX/Northrop “not working” despite the Citi upgrade.
Elevated earnings bar: The coming season demands very high results from core AI names (Santoli’s Micron caveat).
Timing the peak: Attempting to call the market top amid constant rotation risks costly frustration.
This episode was covered in today’s The Market Signal — 2026-07-02, a cross-source synthesis of multiple podcast reports.