Tesla slid nearly 18% this week, notching its largest single-week decline since 2022; SpaceX fell more than 7% this week.

South Korean President Lee Jae-myung, during his visit to Silicon Valley, held a "beer summit" with global AI leaders, as Samsung Electronics and SK Hynix signed major multi-billion-dollar deals with leading U.S. technology companies. According to South Korea’s Presidential Office Policy Office, local conglomerates have reached agreements with global tech giants including Nvidia for a series of cooperation projects totaling $950 billion. Under the deals, SK Group will supply high-performance semiconductors worth $750 billion to global tech firms including Nvidia, while Samsung Electronics will provide chips worth $200 billion to Broadcom. Nvidia announced this Friday that it will partner with South Korea’s SK Group on an AI initiative valued at over $500 billion, covering large-scale AI data center construction and next-generation memory technology. The plan includes a long-term partnership with SK Hynix, designed to secure next-generation memory supplies for Nvidia and co-develop high-bandwidth memory (HBM) for AI training, AI agents, and physical AI applications. SK Telecom plans to build a 2-gigawatt (GW) AI data center that will use Nvidia’s Vera Rubin chips and SK Hynix’s HBM4 high-bandwidth memory, with the first facility expected to become operational in 2027.

On the evening of July 24 local time, U.S. President Donald Trump attended the 2026 White House Correspondents' Association Dinner at Washington D.C.'s Waldorf Astoria Hotel. The event was originally scheduled for April 25 but delayed until now due to an attack incident. Toward the end of his speech, Trump donned a red "Trump 2028" cap and joked, "I'm pleased to announce my intention—this is an exclusive scoop—I plan to run for a fourth term as U.S. president, and I will do it." He added, "It should be easy; I'm getting better and better at running for president these days." During the segment, Trump repeated his false claim that he actually won the 2020 presidential election. Subsequent media reports explicitly noted "he did not win" to clarify this misstatement. Per the 22nd Amendment to the U.S. Constitution, a president is limited to a maximum of two terms. Trump is currently serving his second and final constitutionally allowed term, making him ineligible in theory to legally run for a "fourth term."

According to market sources, Samsung Electronics and Broadcom have signed a memorandum of understanding (MOU) for a five-year, $200 billion advanced storage chip foundry agreement.

South Korea’s presidential office policy chief Kim Yong-beom announced that Samsung Electronics, SK Hynix, and global tech giants including Nvidia (NVDA.O) have reached an agreement to carry out a series of cooperation projects totaling $950 billion. According to Kim, these projects include a long-term deal under which SK Group will supply high-performance semiconductors worth $750 billion to global tech companies including Nvidia. In addition, Samsung Electronics will provide chips worth $200 billion to Broadcom (AVGO.O).

The World Foundation(@worldcoinfnd) raised $52.5M by selling $WLD. On-chain data shows it transferred out 217.4M $WLD, suggesting a sale price of $0.2415 per $WLD. At the same time, the World Foundation has received 47.5M $USDC.

AMD has announced the launch of Instella-MoE, a fully open-source Mixture of Experts (MoE) model. Boasting a total of 16 billion parameters with 2.8 billion activated per token, it is claimed to deliver leading performance among open-source language models of the same scale. AMD stated that Instella-MoE was trained entirely from scratch using its own AMD Instinct MI300X and MI325X GPUs, along with the ROCm software stack. It adopts architectural innovations including Gated Multi-head Latent Attention (Gated MLA) and FarSkip-Collective to enhance training and inference efficiency. Performance tests show that the base model Instella-MoE-16B-A3B has an average score of 76.7, ranking among top open-source models, outperforming models such as SmolLM3-3B and OLMo-3-7B, and competing with larger-scale models despite activating only 2.8 billion parameters. Additionally, the model supports 64K-token long context processing, and has completed a full training pipeline covering pre-training, mid-training, long context extension, supervised fine-tuning (SFT), direct preference optimization (DPO), and reinforcement learning (RL). AMD is simultaneously releasing all of Instella-MoE’s model weights, training configurations, data ratios, intermediate checkpoints, and inference code to advance open AI model research and reproducibility. The company noted that Instella-MoE demonstrates the capability to train large-scale MoE models based on AMD hardware and its open software ecosystem, and will continue developing larger-scale, more capable, and more efficient open-source language models in the future.