On Thursday, the stock with the highest trading volume was Micron Technology, which closed down 0.53% with a turnover of $45.475 billion. After a staggering 684% surge over the past year, Micron Technology still possesses significant upside potential, driven by AI chip demand and a recovery in the memory cycle.
Analysts note that its fiscal second-quarter revenue and earnings per share significantly exceeded expectations, accompanied by a robust third-quarter outlook, propelling its market capitalization toward the trillion-dollar mark. High Bandwidth Memory (HBM) technology is fundamentally transforming the business model of the memory industry, shifting Micron from a cyclical commodity supplier to a quasi-infrastructure enterprise. By securing five-year Strategic Customer Agreements (SCAs), the company has locked in long-term pricing and shipment volumes, substantially enhancing the stability and predictability of future earnings.
The second-highest volume stock was NVIDIA, which closed up 0.78% with a turnover of $29.68 billion. Prior to Thursday, NVIDIA's stock had declined for four consecutive trading sessions. It is reported that CEO Jensen Huang will visit South Korea next week. This marks his return after approximately seven months, signaling NVIDIA's efforts to strengthen its partnership with South Korea.
Industry observers anticipate that during his visit, Huang may discuss collaborations in HBM, next-generation AI performance hardware, and foundry services with Samsung Electronics and SK Hynix. The industry believes that for NVIDIA to ensure the competitiveness of its next-generation products, it needs to enhance cooperation with South Korean firms in HBM, advanced packaging, and foundry sectors. Additionally, he may explore potential supply chain collaborations in cloud computing and physical AI with companies like LG and Naver.
The third-highest volume stock was Microsoft, which closed up 3.47% with a turnover of $19.303 billion. Microsoft is set to unveil multiple proprietary AI models at its Build developer conference next week, with a key focus on a model dedicated to code generation to bolster the competitiveness of GitHub Copilot. Currently, this tool's market share is being eroded by competitors like Cursor and Claude Code. Furthermore, Microsoft will also release models specialized for speech-to-text, reasoning, and image generation, among other applications.
This launch represents a critical test of Microsoft's in-house AI capabilities. Due to restrictions in its agreement with OpenAI, the company has been unable to develop top-tier models; this limitation is set to expire in April next year. Currently, Microsoft's Copilot products primarily rely on models from OpenAI and Anthropic, but the company will lose access to OpenAI's technology in approximately six years, creating an urgent need to build its own AI ecosystem.
Microsoft plans to capture market share with a pricing strategy undercutting similar offerings from OpenAI and Anthropic. The company aims to reduce operational costs through its proprietary models and attract developers to procure them via the Azure cloud platform. If its code model can serve as a cost-effective alternative to Claude, it could help win back price-sensitive developer users.
The fourth-highest volume stock was Western Digital, which closed up 3.25% with a turnover of $16.328 billion. CTO Alper stated that the global AI race is increasingly becoming "memory-centric," potentially exacerbating supply shortages for memory chips.
He pointed out that large language models and key-value caching consume substantial memory, while Mixture of Experts models introduce additional memory demands, making memory a core component of AI competition. He noted it is the first time he has seen customers so resolutely locking in long-term supply; Western Digital has recently signed five supply agreements with terms up to five years, expected to generate at least $42 billion in revenue.
Additionally, Western Digital is collaborating with SK Hynix to develop stacked flash memory, with samples anticipated by year-end and full products launching next year. This technology holds promise as a novel memory solution for AI inference scenarios.
The fifth-highest volume stock was Advanced Micro Devices, which closed up 4.55% with a turnover of $16.04 billion. On Thursday, Thunderobot unveiled a comprehensive AI workstation product matrix covering tower, mini-PC, and mobile form factors, marking the industry's first complete coverage across these three categories. At the launch event, an executive from AMD's Greater China region stated that AMD is building a full-stack computing power system spanning cloud, client, and edge, aiming to co-build intelligent computing infrastructure through an open ecosystem. Reflecting on years of collaboration with Thunderobot in gaming hardware, the two have now formed strategic synergy in the AI computing track. AMD will provide robust computing power support for Thunderobot's AI workstations, from flagship supercomputing to portable terminals, leveraging its full-range chip portfolio.
The tenth-highest volume stock was Marvell Technology, which closed up 3.09% with a turnover of $10.543 billion. Marvell Technology reported record first-quarter revenue of $2.418 billion, a 28% year-over-year increase, slightly exceeding expectations; data center business revenue reached $1.83 billion, up 27% year-over-year. Although net profit declined significantly due to one-time charges related to acquisitions, non-GAAP earnings per share met expectations.
For the second quarter, the company anticipates revenue of approximately $2.7 billion, representing about 35% year-over-year growth, surpassing analyst expectations. Regarding full-year guidance, the company raised its fiscal 2027 revenue forecast to about $11.5 billion, implying roughly 40% growth. Given strong AI orders, the fiscal 2028 revenue target was increased to approximately $16.5 billion, about $1.5 billion higher than prior guidance, corresponding to approximately 45% year-over-year growth.
The twelfth-highest volume stock was Snowflake, which closed up 36.48% with a turnover of $9.363 billion. Amazon.com announced that Snowflake has committed to purchasing $6 billion worth of services and technology from AWS over five years, including Graviton general-purpose chips and AI GPUs. This represents the largest single investment in Snowflake's history.
Concurrently, Snowflake's first-quarter results showed adjusted earnings per share of 39 cents and revenue of $1.39 billion, both exceeding expectations. The company also announced the acquisition of AI agent platform Natoma, which will be integrated into its AI Data Cloud. Following these announcements, Snowflake's stock surged 37.62% in after-hours trading.
Analysis suggests that AI has not eliminated SaaS demand but has instead amplified corporate needs for system capabilities like data governance and observability. Snowflake is positioned on a critical path for the scaled deployment of AI, and sustained capital expenditure by major AI companies indicates that the current AI market trend has not yet peaked in the short term.
The sixteenth-highest volume stock was Palantir Technologies Inc., which closed up 8.17% with a turnover of $7.175 billion. Palantir Technologies' stock rose today, primarily driven by optimism surrounding the finalization of long-term commercial contracts and continued strength in its AI and defense businesses. The company has extended its partnership with Thomas Cavanagh Construction through 2035, demonstrating strong commercial client stickiness and the predictability of enterprise-level revenue. Additionally, the market is optimistic about its deep integration in the defense sector and the growth potential driven by AI security demands.
However, the stock still faces negative pressures: some analysts question its high valuation and warn that competition is intensifying from rivals like OpenAI and Anthropic; furthermore, recent insider sales exceeding $43 million have raised investor concerns about valuation risks.
Analysts state that, overall, Palantir's gains largely depend on sustained market confidence in its growth narrative, but high valuation and competitive pressures are also capping upside potential, resulting in a mix of optimism and caution.
The eighteenth-highest volume stock was NEBIUS, which closed up 8.62% with a turnover of $5.782 billion. Nebius Group's stock rose today, propelled by multiple positive catalysts. A fund managed by prominent AI investor Leopold Aschenbrenner disclosed a stake of approximately 5.6% in the company, involving an investment of about $2.6 billion, significantly boosting market confidence.
Simultaneously, the company's latest earnings report showed revenue surged 684% year-over-year to $399 million, with gross margin expanding to 74% and EBITDA turning positive, demonstrating strong operating leverage. Additionally, the company has reached an agreement with Meta Platforms, Inc. and established power and infrastructure partnerships with entities like Bloom Energy, which could accelerate data center expansion.
Despite the overall optimistic market sentiment and media buzz surrounding its "Cloud AI" concept, some analysts caution that the rapid stock price increase may expose investors to execution risks; if the company cannot consistently expand capacity, secure power, and obtain customer contracts, its valuation could face pressure. Overall, the market is reacting positively to this series of strategic developments.
Comments