Here are Thursday's biggest analyst calls: Nvidia, Tesla, Berkshire Hathaway, Amazon, Texas Instruments & more

Wall Street analyst actions were broadly constructive, with buy reiterations on Tesla, Nvidia, IBM, Alphabet and Amazon, plus upgrades on On Holding, SiriusXM, Texas Instruments and McDonald's. UBS raised Berkshire Hathaway’s price target to $581 from $578, while BofA cited buybacks as a key support factor; the main negatives were downgrades on CSX, Lucky Strike and Madison Square Garden Sports. The article is mostly stock-specific research commentary, with the most notable thematic positives centered on AI, software, semiconductors, and select industrial/consumer names.

Analysis

The tape is telling a coherent story: the market is still rewarding businesses with either algorithmic scarcity or self-funded optionality. AI remains a winner-take-most trade, but the more interesting second-order effect is vendor concentration risk moving down the stack — software and industrial silicon names with real switching costs should see a bid, while “good-enough” AI alternatives will struggle to re-rate until they prove ecosystem pull. That favors NVDA, GOOGL, DDOG, and TXN, but it also implies tighter scrutiny on any supplier whose upside depends on a cyclical recovery rather than structural share gains.

The biggest idiosyncratic dislocation is in the non-tech growth names. ONON and OKLO both screen as premium-duration stories, but their catalysts are different: ONON needs operating leverage and FX relief over the next 2-3 quarters, while OKLO is a multi-year infrastructure financing story with binary execution risk. USAR adds a geopolitical scarcity angle that can extend well beyond the current news cycle; the real trade is not just rare earth pricing, but Western OEMs beginning to pre-emptively dual-source away from China, which could support valuation multiples for domestic supply-chain names even before cash flow inflects.

On the other side, CSX, LUCK, and MSGS are showing the first signs of being “good businesses, expensive stocks.” That usually persists until macro traffic or pricing weakens enough to compress consensus estimates, which is why the short-side opportunity is higher confidence over the next 1-2 quarters than the long-side in these names. MCD is interesting as a potential sentiment reversal: if traffic stabilizes after the value reset, the stock can work as a defensive growth compounder, but any improvement likely comes from mix and units, not a broad demand rebound.

The contrarian miss is that capital returns may matter more than headline growth in this tape. BRK.B and IBM both have the ingredients for multiple support from buybacks, margin discipline, and low expectation sets; these are the kinds of names that outperform when AI and growth leadership get crowded. The market is paying up for narrative, but the cleaner risk/reward may be in cash-generative defensives with explicit capital deployment levers.

Key Decisions for Investors

• Long NVDA / short a basket of AI-adjacent inferiors (or weakly monetizing AI names) over 1-3 months; risk/reward favors the incumbent ecosystem if model-training spend remains concentrated.
• Buy TXN on pullbacks for a 3-6 month trade; the setup is better for share gains than cyclical beta, with downside limited if industrial recovery stalls because the market is already underestimating fab scarcity leverage.
• Initiate a pair: long BRK.B, short CSX for 3-6 months; both are quality compounds, but BRK.B has cleaner capital return support while CSX is vulnerable to multiple compression if growth slips even modestly.
• Go long USAR as a thematic catalyst trade over 6-12 months, but size it smaller than typical because the upside is policy- and supply-chain-driven rather than earnings-driven; use a trailing stop if China supply rhetoric cools.
• Fade LUCK on strength over the next quarter; macro traffic pressure and weak comp math make the downside more immediate than the market is pricing, especially if consumer discretionary spending softens.