Trump meets with AI execs, Treasury yields pressure stocks, Alaska's luxury push and more in Morning Squawk

Rising Treasury yields pressured equities, with the Dow falling nearly 350 points, the Nasdaq down almost 1%, and the S&P 500 posting its largest one-day decline in more than a month. AI developments dominated headlines: AMD agreed to acquire World Labs for $8.2 billion, OpenAI shelved a model that failed internal safety standards, and President Trump is meeting major technology executives in Washington. Separately, Mesabi Metallics plans a $15 billion Iowa steel plant targeting 7.5 million tons of annual output, while MongoDB shares plunged 18.5% after CEO CJ Desai departed for Meta.

Analysis

The AI policy meeting is unlikely to alter near-term monetization, but it raises the probability that compliance, safety testing and provenance requirements become a fixed cost of model deployment. That favors hyperscalers with distribution and balance sheets (META, GOOG) and NVDA, whose ecosystem can absorb regulatory complexity, while raising the funding hurdle for independent model developers. The non-obvious risk is that safety-driven release delays shift enterprise spend from frontier-model experimentation toward proven inference, data-management and workflow products over the next 6-18 months.

MDB's drawdown should not automatically be bought: a CEO departure during an AI-platform transition creates both execution risk and a potential reset of sales leadership, product roadmap and forward-margin expectations. The relevant 1-3 month catalyst is whether the interim CEO reaffirms Atlas consumption growth, net revenue retention and FY operating-margin targets; absent that, the stock can remain de-rated even if the underlying database market is intact. META's enterprise initiative is strategically credible as a distribution channel, but investors should treat it initially as another capex-intensive product investment rather than a discrete earnings catalyst.

AMD's World Labs purchase expands its exposure to embodied AI and simulation, but the acquisition is more valuable as a developer-stack differentiator than as a near-term revenue contributor. If deal consideration is materially stock-funded or management signals elevated R&D and integration spending, AMD's relative multiple could compress versus NVDA despite the strategic narrative. Higher real yields remain the immediate cross-asset constraint: they disproportionately pressure long-duration AI valuations and discretionary premium-travel demand, making Alaska's investor-day capital-allocation and unit-revenue targets more important than cabin-upgrade headlines.

Key Decisions for Investors

• Maintain a 1-3 month relative long NVDA / short AMD position rather than adding outright AMD: NVDA retains the cleaner monetization and software-distribution setup, while AMD faces acquisition-integration and capital-allocation uncertainty. Reassess if AMD discloses limited dilution and a concrete revenue roadmap for World Labs; stop if AMD outperforms NVDA by 10% following transaction terms.
• Do not bottom-fish MDB before management's first post-departure operating update. Establish an alert to buy only if Atlas consumption growth, net revenue retention and full-year margin guidance are reaffirmed; a guidance cut or a further leadership reshuffle supports a tactical short versus ORCL or SNOW over the next quarter.
• Hold META versus GOOG only as a selective 6-12 month AI-distribution overweight, sized for elevated capex risk. Add after evidence of paid enterprise adoption or incremental ad-targeting/productivity monetization; reduce if 2027 capex guidance rises without a corresponding revenue or margin framework.
• Treat ALK investor day as a conditional catalyst, not a pre-event long. Buy only if management quantifies premium-seat revenue uplift, fleet-capex funding and ROIC above its cost of capital; otherwise, premium retrofits risk becoming a margin drag during a weakening yield environment. A long ALK / short JETS structure is preferable if those targets are delivered.