Former Google CEO Eric Schmidt addressed a crowd of thousands of students at the University of Arizona, stating that artificial intelligence is poised to infiltrate every profession, classroom, hospital, and laboratory. Hardly had he finished his words when a chorus of boos broke out from the audience. Since May 2026, this marks the third commencement speech on a U.S. campus that has been met with jeers due to the mere mention of AI.

Since ChatGPT made its grand entrance at the close of 2022, NVIDIA's stock price has skyrocketed, multiplying by 12 times. This meteoric rise has transformed even entry-level roles into lucrative 'gold mines.' Contrary to popular belief, however, a select group of key NVIDIA employees have chosen to forgo millions in unvested stock options, opting instead to ride the wave of the AI startup boom. While NVIDIA's staggered stock vesting system is designed to retain top talent, some employees view the fleeting opportunity presented by the AI revolution as invaluable, prompting them to strike out on their own and launch their ventures. The AI expertise these entrepreneurs have amassed can now be fully leveraged, and NVIDIA's esteemed reputation, along with Jensen Huang's management approach, provide a solid foundation for their future endeavors.

The company that once laid the crucial foundation for AI development is now facing the loss of its core talent. On June 18, Noam Shazeer, a key author of the Transformer paper and co-lead of Google Gemini, announced on the social platform X that he would be leaving Google to join OpenAI, which has secretly filed an initial public offering (IPO) application with the U.S. Securities and Exchange Commission (SEC). Noam Shazeer is one of the eight authors of the 2017 paper "Attention is All You Need," which laid the technical groundwork for modern large language models. OpenAI CEO Sam Altman promptly reposted and commented, stating that Noam has been one of the people he most wanted to collaborate with since the founding of OpenAI, and mused, "It only took a decade."

According to tech media reports, OpenAI is expected to launch its next-generation flagship model, GPT-5.6, later this month (June 22-28), featuring mini, standard, and Pro versions. This model achieves leapfrog improvements in core capabilities such as coding, Agent workflows, and 3D generation, with the context window expanded to 1.5 million tokens, a 43% increase, and optimized long-session coding capabilities and Codex response speed. Meanwhile, the token pricing for GPT-5.6 may be just one-third of that of its competitors, aiming to disrupt the market with a low-price strategy and accelerate enterprise customer migration.

Microsoft has recently announced its decision to eliminate the highly acclaimed Drop file-sharing functionality in the upcoming iteration of the Edge browser. Notably, Edge version 149 had already seen the sequential removal of the sidebar and collections features. This sequence of actions underscores Microsoft's strategic pivot towards enhancing Copilot and driving the browser's AI transformation, thereby championing a full-fledged AI integration within the browser.

On June 21, 2026, reports emerged that WeChat’s AI assistant, “Xiaowei,” has commenced a limited grayscale rollout. For select users, an eye-shaped icon has appeared in the top-left corner of the main WeChat interface, acting as the gateway to the “Xiaowei” beta. According to Tencent’s customer service, “Xiaowei” is an in-house AI assistant currently undergoing testing by the WeChat team in a controlled environment. It enables users to perform native WeChat functions and access Mini Programs through text or voice commands, such as sending messages to friends, checking Moments, scheduling services, and more.

The head of Anthropic has asserted that artificial intelligence (AI) companies must attain annual revenues in the realm of 'hundreds of billions of dollars'; otherwise, they will confront significant risks to their continued existence.

Reports indicate that Tesla has formally filed an application to register the trademark 'Amazing Abundance'. This strategic move is widely regarded as a pivotal stride for the company in pursuing its long-term growth ambitions, particularly in the realms of artificial intelligence, humanoid robotics, and autonomous driving systems. Furthermore, Tesla envisions leveraging automation technology to substantially slash production costs.

John Jumper, the recipient of the 2024 Nobel Chemistry Prize and a senior scientist, has declared his intention to leave Google DeepMind, where he has dedicated nearly nine years of service. He plans to make this transition in June 2026 and join the AI startup Anthropic. As one of the co-developers of AlphaFold, Jumper's addition to Anthropic is anticipated to significantly bolster the company's prowess in AI safety and the R&D of state-of-the-art technologies.

In the wake of its public debut, SpaceX is set to issue bonds worth a minimum of $20 billion, a strategic move aimed at bolstering its foray into artificial intelligence and the establishment of orbital data centers. This bond issuance is earmarked to settle a $20 billion bridge loan that comes due in 2027. Projections indicate that by 2031, SpaceX's net debt will soar beyond the $400 billion mark, accompanied by capital expenditures exceeding $1 trillion, predominantly directed towards AI endeavors and the construction of space-based data centers.

Yesterday at noon, Tesla founder Elon Musk addressed a query from a netizen on a social media platform, suggesting that China’s AI development could achieve "Fable" level—a term often used to describe highly advanced, almost mythical AI capabilities—by the first quarter of 2027. Later that evening, Tang Jie, the chief scientist at Zhipu AI, countered Musk’s prediction, stating that such advancements would likely occur sooner than anticipated. It is noteworthy that Zhipu AI unveiled and open-sourced its latest flagship model, GLM-5.2, on June 17. This model is tailored for extended tasks, boasting key features such as 1M lossless context retention, enhanced coding efficiency, and other advanced functionalities.

On June 20, 2026, the Tencent WeChat team announced the commencement of a small-scale grayscale test for its in-house AI assistant, named 'WeChat Xiaowei.' This innovative feature is currently in its testing phase. WeChat Xiaowei enables users to engage with it through both text and voice commands, facilitating seamless operation of native WeChat functions. These functions encompass adjusting settings, sending messages, and initiating calls, among others.

From a technical standpoint, WeChat Xiaowei leverages Tencent's proprietary AI models, complemented by high-quality open-source models sourced from the industry. This integration ensures a robust and versatile AI assistant. Presently, the grayscale testing phase has encompassed multiple operating systems, ensuring broad compatibility. Participation eligibility is determined based on the actual page displayed to users, ensuring a tailored and efficient testing process.

Metal-organic frameworks (MOFs) have attracted significant interest within the scientific community, primarily due to their customizable structures and versatile applications. While X-ray diffraction technology is a well-established tool for material characterization, efficiently analyzing powder X-ray diffraction (PXRD) data and predicting the crystal structures of MOF materials in high-throughput experimental and self-driving laboratory settings continues to pose challenges for researchers. Professor Pan Feng's team, based at the School of Advanced Materials at Peking University's Shenzhen Graduate School, specializes in graph-theoretic structural chemistry, AI for Science (AI4S), and materials genomics. They have successfully leveraged artificial intelligence technology to analyze XRD data and have innovatively developed a generative artificial intelligence framework, Xrd2Mof, utilizing a diffusion model. This framework utilizes PXRD patterns, metal nodes, and organic ligand information as inputs, directly outputting MOF structures and thus achieving a significant technological breakthrough.

Recently, the Intelligent Photonics Team from the Beijing Institute of Technology (BIT) has achieved significant milestones in the realm of "Large Language Models + Nanosynthesis." The team has meticulously assembled an extensive nanocrystal database, encompassing nearly 160,000 high-quality, meticulously aligned data entries. This database provides a comprehensive overview of crucial aspects, including synthesis steps, reactants, and the physicochemical properties of nanocrystals. Leveraging this rich repository, the team has developed the NanoExtractor large language model, which significantly boosts information extraction accuracy to an impressive 92% through the implementation of four sophisticated data augmentation strategies. Additionally, they have constructed the NanoDesigner generative inverse design model. This innovative model is capable of directly generating specific, actionable synthesis routes tailored to target products, constrained reactants, and desired properties, achieving an average experimental success rate of 80%. This research represents a groundbreaking departure from the inefficiencies of traditional "trial-and-error" methodologies, paving the way for intelligent, on-demand customization of nanomaterials. The related findings have been published in the esteemed journal ACS Nano.

On June 18, the Tsinghua News Network reported that comprehending the constitutive behavior of materials under extreme conditions is essential for the modeling, design, and utilization of advanced metals and superalloys. Presently, the majority of constitutive models are based on phenomenological approaches. Although Crystal Plasticity Finite Element (CPFE) modeling can integrate physical equations at the microscale level, its extensive computational demands hinder its widespread adoption in engineering applications. Traditional mechanics research typically integrates theoretical analysis, numerical simulation, and experimental observation. Nevertheless, when confronted with high-dimensional parameter spaces, intricate multiscale coupling, and nonlinear experimental data, it often encounters obstacles such as exorbitant computational costs and stringent model assumptions. The incorporation of artificial intelligence (AI) and deep learning (DL) technologies is facilitating a transition in mechanics research from model-driven to data-driven methodologies.

John Jumper, the Nobel Prize laureate in Chemistry and a key contributor to the development of AlphaFold, has declared that he is leaving Google DeepMind—a company where he has dedicated nearly nine years of his career—to become part of the AI startup Anthropic. Shortly after earning his Ph.D. in 2017, Jumper embarked on his journey with DeepMind. Although he had limited experience in deep learning at that time, his extensive knowledge of protein physics paved the way for him to lead the AlphaFold team. Under his leadership, AlphaFold successfully cracked the protein folding problem, marking a groundbreaking achievement in the field of computational biology. For this remarkable feat, Jumper, along with Demis Hassabis, was honored with the 2024 Nobel Prize in Chemistry. In his new role at Anthropic, Jumper is set to bolster the company's endeavors in the life sciences domain, expediting the processes of drug discovery and the design of biological experiments.

US President Trump has expressed that, in his view, Anthropic PBC does not constitute a national security threat. Nevertheless, just a few days prior, the Trump administration, citing national security concerns, barred foreign governments, businesses, and individuals—including foreign nationals residing within the United States—from accessing Anthropic's most cutting-edge AI models, namely Mythos 5 and Fable 5. This action compelled Anthropic to suspend access to these two models for all users.

On June 16, 2026, a US federal judge dismissed xAI's lawsuit against OpenAI for stealing trade secrets, with xAI losing the case and being barred from filing another lawsuit on the same grounds.

On June 19 local time, senior research scientist John Jumper announced on the social platform X that he would conclude his nearly nine-year tenure at Google DeepMind and join the artificial intelligence startup Anthropic. It is reported that John Jumper, along with his Google DeepMind colleague Demis Hassabis and David Baker from the University of Washington in Seattle, jointly won the 2024 Nobel Prize in Chemistry.

The Norwegian government has decided to nearly fully ban the use of generative artificial intelligence tools in primary schools starting from the new school year in autumn 2026, and to strengthen restrictions on their use in junior high and senior high schools. This move aims to prevent the technology from undermining students' foundational learning abilities and to avoid further deterioration in educational performance.