Micron Technology led Tuesday's top-traded stocks, closing 7.62% higher with $322.63 billion in volume. Bank of America reiterated its Buy rating and $1,550 price target, calling the recent share price decline a "compelling buying opportunity." The bank expects memory chip prices to normalize by 2027-2028 and says the company's fundamentals remain strong. BofA forecasts fiscal 2028 earnings per share of around $150, with a bearish scenario of roughly $100, still well above the peak of about $12 in the prior cycle.
Nvidia finished 2.56% higher on $272.96 billion in volume. Reports emerged Tuesday that SpaceX is partnering with Nvidia to develop a satellite computing payload called "Starmind AI1." Separately, Nvidia CEO Jensen Huang posted on X that the company officially launched Alpamayo 2 Super, a cutting-edge open-source inference model for autonomous vehicles. Beyond basic visual perception, Alpamayo can understand and reason about complex environments—thinking deeply before acting. It serves as a core pillar for the long-tail market of robotaxis, trucks, shuttles, freight logistics vehicles, tractors, and various mobile robots, aiming to power billions of autonomous machines. To allow development teams to deeply review, fine-tune, and deploy the model, Nvidia released it under the OpenMDW-1.1 license for commercial use, believing open-source makes technology safer and more reliable. The next wave of AI is robotics, and autonomous driving is where it all begins.
Palantir surged 29.45% on $272.27 billion in volume. Under CEO Alex Karp, Palantir's stock soared after reporting strong second-quarter results. The data analytics company beat expectations on accelerating revenue growth and robust demand. Deutsche Bank analyst Brad Zelnick said this "further reinforces our view that Palantir is several steps ahead of other software companies in translating AI into customer value." Zelnick noted that while many software providers are still figuring out how to integrate AI, "Palantir increasingly looks like a time traveler that has already arrived at the AI future others are still aspiring to."
AMD rose 7% on $224.72 billion in volume. After Tuesday's close, AMD reported second-quarter revenue of $115.4 billion, beating the consensus estimate of $112.8 billion. The company expects data center sales to accelerate in the second half of 2026.
SanDisk gained 10.84% on $200.72 billion in volume. On Tuesday, SK Hynix and SanDisk jointly released the first standard specification for High Bandwidth Flash (HBF) at the FMS 2026 flash memory summit, along with a complete performance roadmap, ecosystem partnerships, and mass production timeline. HBF is positioned as a new storage tier between HBM and SSDs, offering both near-HBM data transfer speeds and significant capacity expansion through NAND flash. As AI inference applications proliferate, HBF is attracting attention for its ability to meet both bandwidth and capacity demands.
SpaceX closed 9.43% higher on $156.01 billion in volume. The company released its first earnings report as a public company after its record IPO in June, with second-quarter revenue far exceeding expectations. Revenue was $78.1 billion, above the $69.3 billion forecast. The loss per share was 9 cents, compared to the average analyst estimate of a 26-cent loss. SpaceX said revenue grew 92% from $41 billion in the same period last year. Since opening at $150 on June 12, SpaceX stock has fallen 16%, with market capitalization significantly shrinking. The company lost $49 billion last year, primarily due to heavy investment in AI infrastructure. It merged with Musk's xAI in February, stating a vision to build data centers in space. Even its launch business, which includes NASA contracts, is currently operating at a loss.
Alphabet (Class A) rose 1.11% on $133.25 billion in volume. A Barclays report suggested that if Google adjusts its AI chip business model, its TPU could generate up to $252 billion in revenue by 2028. Barclays believes TPU, one of the most widely used custom AI chips, could expand its ecosystem if Google forms a joint venture with Blackstone, Apollo Global, and Broadcom. Currently, Google's AI chip business does not sell TPUs directly to data center operators like CoreWeave, but instead provides TPU computing power through Google Cloud. In its second-quarter report, Google Cloud revenue grew 82% year-over-year to $248 billion.
Broadcom rose 6.61% on $121.15 billion in volume.
Intel gained 10.84% on $117.77 billion in volume. Reports indicate progress in Intel's EMIB-T advanced packaging technology, with yields approaching 90% and costs about 50% lower than TSMC's CoWoS solution. Intel plans to offer the packaging service at scale by 2027, but substrate yields are only around 50%, a major bottleneck for capacity expansion. Suppliers are working to improve output and yields. Separately, Intel is pushing forward with its Ohio megafab campus, targeting full operations by 2031. The facility, spanning about 1,000 acres, will produce 18A and 14A process nodes, with mass production of 14A planned for 2031.
Marvell Technology closed 12.81% higher on $72.3 billion in volume. Marvell announced an expansion of its AI memory infrastructure portfolio, introducing solutions for server-level AI storage, rack-level CXL memory expansion and pooling, and cluster-level optical interconnect shared memory. These innovations aim to address growing memory capacity and bandwidth bottlenecks in agentic AI inference. As AI models scale, context windows lengthen, and KV cache demand surges, the traditional tightly coupled compute-memory architecture is becoming a key barrier to AI inference efficiency. Marvell's "memory disaggregation" architecture allows memory resources to scale independently from compute, reducing GPU idle time and improving overall infrastructure utilization.

