AI Stocks & Markets

Today's Stories

1. 1

   Booking Holdings faces analyst concerns about AI cutting out travel intermediaries, but the company argues its market position and customer relationships insulate it from this threat.

   Analysts have raised concerns that artificial intelligence could bypass travel booking platforms like Booking Holdings, allowing travelers and hotels to interact directly without intermediaries. The company is defending its role in the market. For Booking Holdings investors and competitors, this debate signals uncertainty about whether AI will erode the traditional commission-based travel booking model. If AI disintermediation gains traction, it could reduce Booking's role as a connector between customers and suppliers.

   Booking Holdings' ability to retain its customer base and prove that its network effects and brand loyalty are stronger than the pull of direct AI-powered booking alternatives will be central to validating or disproving the disintermediation thesis.

2. 2

   Micron's Blockbuster Earnings Quiet the AI Doubters

   Micron's Blockbuster Earnings Quiet the AI Doubters

3. 3

   Nvidia dominates AI chip revenue with $75.2 billion（約12兆円） in data center sales, far outpacing AMD, Broadcom, and Marvell, while trading at a cheaper valuation than its peers.

   Nvidia's data center division generated $75.2 billion（約12兆円） in Q1, up 92% year over year, significantly larger than AMD's $5.8 billion（約9300億円）, Broadcom's $10.8 billion（約1.7兆円）, and Marvell's $1.8 billion（約2900億円） in their respective recent quarters. Broadcom's AI semiconductor division grew 143% year over year, while Marvell's data center division grew 27% year over year. The four companies compete in two segments—Nvidia and AMD design general-purpose GPUs, while Broadcom and Marvell design custom AI chips (ASICs) for major cloud providers. Broadcom management expects its AI unit to generate over $100 billion（約16兆円） in revenue in 2027 as hyperscalers (large cloud providers) launch custom chips, and Marvell expects its annual revenues to double in fiscal 2028, indicating a structural shift in how AI workloads are deployed.

   Nvidia trades at 23 times forward earnings, making it cheaper than AMD and Marvell and offering better value than Broadcom. Nvidia's new Vera Rubin architecture will start shipping to customers later this year, which should help sustain its elevated growth rates.

4. 4

   Nvidia and AMD are positioned differently in AI's next phase: Nvidia dominates training, but AMD may have more growth runway in inference and agent AI despite trading at a premium valuation.

   A strategic analysis compares Nvidia and AMD's AI chip positions. Nvidia built dominance in AI model training through its CUDA software platform and has expanded into a complete AI infrastructure company through acquisitions including Mellanox and Groq. AMD, historically behind in training, has strengthened its position for inference (where AI produces answers) and announced an acquisition of MEXT, a memory optimization company, to compete in cost-effective AI server solutions. The two chips address different parts of AI's value chain. Nvidia's forward price-to-earnings ratio stands at under 16 times fiscal 2028 estimates, while AMD trades at a 39.5x one-year forward P/E, reflecting different growth expectations. AMD's smaller market cap—less than $900 billion（約140兆円） compared to Nvidia's $5 trillion（約800兆円）—and positioning in two emerging areas (inference and agentic AI, which requires CPUs) suggest it may face less headwind from sheer scale, though at a higher current valuation.

   AMD has two $100 billion（約16兆円） GPU inference deals in place and sees a $120 billion（約19兆円） addressable market in agentic AI as the GPU-to-CPU ratio shifts from 8:1 for training to 1:1 for agentic AI. The inference market is expected to eventually grow larger than the training market. Both companies remain well-positioned for long-term AI infrastructure demand, but their near-term growth drivers differ significantly.

5. 5

   Apple raises Mac and iPad prices by $100 to $300, marking its first major shift to pass higher AI memory costs directly to consumers.

   Apple increased prices on Mac and iPad models by $100 to $300 depending on the device. CNBC's MacKenzie Sigalos characterized the move as the company's first real attempt to pass the AI memory crunch directly onto consumers. The price increases signal Apple is responding to rising costs associated with AI capabilities. For customers and businesses relying on Apple's products, this marks a shift in how the company handles the cost pressures of advanced AI features rather than absorbing them internally.

   The iPhone could be next in line for price increases, suggesting Apple may extend this pricing strategy across its product lineup as AI memory demands continue to grow.

6. 6

   Salesforce and Palantir both posted strong revenue growth this spring, but the market reacted differently to each—signaling that investors are now judging AI software companies on financial fundamentals, not just headline growth numbers.

   Salesforce reported $11.13B in revenue, up 13.3% year-over-year, driven by Agentforce momentum, while Palantir delivered 85% revenue growth and a Rule of 40 score of 145. Despite strong results, the market's reaction to each stock diverged sharply. The divergent market response suggests investors are looking beyond raw growth rates and examining the deeper financial metrics—like profitability balance and efficiency scores—that indicate whether an AI-powered company can sustain value long-term. This shift could reshape how enterprise software businesses are evaluated.

   The Rule of 40 score, which combines revenue growth and profitability metrics, is emerging as the cold, hard financial metric investors use to compare enterprise software giants in the AI era.