CNBC Daily Open: The U.S. is flexing its economic power — drawing retaliation
Source: CNBC

Canada retaliated “dollar for dollar” to U.S. tariffs, imposing new duties on 700+ U.S. goods worth about $20B, including 50% tariffs on American steel and aluminum (doubling the prior rate). In parallel, the U.S. warned it could cut off firms helping Iran evade sanctions, while China modestly hedges with expanded use of CIPS to reduce dollar-centric exposure. Despite the trade/sanctions escalation, U.S. indexes rose overnight and chip stocks rallied ahead of Nvidia results (+2%; AMD +4.9%, Micron +2.5%), while Apple also announced AI-upgraded Mac Mini/Mac Studio models with prices starting at $899 (+$100).
Analysis
The immediate signal is cost inflation moving back up the stack, but the more important read-through is bargaining power. Hardware names that can reprice fast will protect margins; those with sticky demand but limited pricing power will see margin leakage before volume shows it. That makes AAPL more interesting than the tariff noise suggests: the market should test whether higher ASPs are a deliberate margin defense or the first sign that consumer elasticity is finally biting.
TSM remains the cleaner structural winner because leading-edge capacity is still scarce and AI silicon demand is still node-intensity driven. The risk is not the product cycle; it is policy and logistics: any broadening of sanctions enforcement, shipping disruption, or energy spike would hit the semiconductor supply chain through freight, substrates, and fab operating costs before it shows up in revenue. The second-order beneficiary of the trade war is actually domestic metals/industrial input suppliers, not the chip complex.
Into earnings, NVDA, AMD, and MU are being bid as a single AI factor, but the setup is fragile. If guidance is merely consistent rather than accelerating, multiple expansion can stall because the market is already paying for uninterrupted hyperscaler capex and flawless monetization. The contrarian angle is that OpenAI turnover and Apple’s price increases both point to AI demand still being narrow and expensive; that is bullish for infrastructure suppliers only until buyers start rationing spend.
AllMind Terminal
AI-powered research, real-time alerts, and portfolio analytics for institutional investors.
Request TrialMarket Sentiment
Overall Sentiment
mixed
Sentiment Score
-0.05
Ticker Sentiment
Key Decisions for Investors
- Relative value: long MU / short AAPL for 1-3 months. Memory pricing should have more direct leverage than Apple’s ability to pass through higher component costs; fade AAPL strength if gross-margin commentary does not re-accelerate.
- Event trade: buy a modest NVDA upside call spread into earnings only on a pullback, not after a squeeze. The setup is asymmetric if guide raises AI capex visibility, but the position should be capped because a merely 'good' print can still be sold.
- Keep TSM on the watchlist as a buy-the-dip structural long over 6-18 months. The thesis fails if export-control risk widens or if management signals node mix is not converting into pricing power.
- Avoid chasing AMD on the first pop. Use it as the short leg against NVDA if the market starts pricing in broad AI share gains without evidence of sustained datacenter wins.
More News
- Nvidia GPUs are everywhere. Here are the ways companies are accessing them
- As companies pour billions into Earth-based AI infrastructure, Google is taking the data center race off-planet
- Trading expert sets date when Micron (MU) stock will crash to $400
- Wall Street is pitching data centers as a major real estate bet. The risks are piling up
- Global PC shipments crater 20% as rising prices hammer demand
- AI predicts Palantir stock price for end of Q4 2026
From AllMind Research
- Anthropic IPO Preview: Valuation, Timing, and What to Watch
- Shein After the IPO: Venue, Valuation, and What Must Be Proved
- What AI Research Tools Should a Small Hedge Fund Buy First?
- AI in Asset Management: 2026 Statistics That Hold Up
- What is Broker Research and RMS Systems (And How to Actually Use Them)