Pulse

US President Donald Trump's visit to China included a delegation of over a dozen American business leaders, including CEOs from major tech companies like Apple, Nvidia, and Tesla. The group, with a combined net worth exceeding $1 trillion, aimed to foster deals and discuss trade, investment, and supply chains. Discussions also touched upon artificial intelligence as a critical battlefront in the tech war between the two nations, with US export controls on semiconductors and China's rare earth exports being key strategic points. AI

Chinese AI startup DeepSeek is reportedly seeking to raise up to 50 billion RMB (approximately $7 billion USD), which would be the largest funding round for a Chinese AI company. The company has already established R&D centers in Silicon Valley, London, and Singapore to support its global expansion. DeepSeek gained attention for its V3 model, which demonstrated performance competitive with leading global AI models while incurring significantly lower training costs. AI

Wispr Flow, a voice AI startup, is experiencing accelerated growth in India following the launch of its Hinglish language support. The company is strategically focusing on the Indian market, despite the inherent complexities of its diverse linguistic landscape and fragmented monetization. Wispr Flow aims to expand its multilingual voice capabilities and local hiring to reach a broader user base beyond white-collar professionals. AI

B.AI has announced a new initiative to distribute 100 billion tokens, aiming to boost its presence in the AI execution scaling sector. This limited-time token drop is designed to establish B.AI as a significant player in the evolving AI landscape. The program focuses on providing resources to support AI development and deployment. AI

A North American watchdog has issued a Level 3 alert, warning that data centers, particularly those supporting AI workloads, pose a significant threat to the power grid's stability. The North American Electric Reliability Corporation (NERC) cited the rapid fluctuations in power demand from these facilities as a major risk that current response protocols cannot adequately manage. NERC has mandated that grid operators submit mitigation plans by August 3rd to address these concerns, which also impact cryptocurrency mining and traditional data centers. AI

Oracle has secured a significant artificial intelligence contract with the Pentagon, a development that analysts believe could positively impact the company's stock performance. The specifics of the deal and its potential financial implications are being closely watched. AI

The African Lion 2026 military exercise in Morocco is integrating advanced AI, FPV drones, and autonomous robots into simulated combat scenarios. Concurrently, Unusual Machines is showcasing its drone ecosystem at XPONENTIAL 2026, highlighting sensors, AI, and autonomous capabilities. Both developments indicate a strong push by the US market to industrialize drone warfare and integrate AI into future military operations. AI

Subquadratic has unveiled a new model with a 12 million token context window, significantly expanding the amount of information an AI can process at once. This breakthrough shatters previous records for context window size. The development promises to enable more complex and nuanced AI applications by allowing models to retain and analyze much larger amounts of data. AI

A Chinese AI research lab has reportedly developed a new model with an astonishing one trillion parameters. This development signifies a significant leap in model scale, potentially enabling more complex reasoning and understanding capabilities. Further details on the model's architecture, training data, and performance benchmarks are anticipated. AI

Anthropic has partnered with SpaceX to significantly increase the computational power available for its Claude models, doubling usage limits for paid subscribers. This collaboration is seen as crucial for advancing AI development within the industry. Separately, a new study indicates that popular chatbots like ChatGPT, Claude, Grok, and Perplexity are sharing user data with advertising brokers such as Meta and Google, raising substantial privacy concerns. AI

Executives linked to Supermicro allegedly used a Thai government-affiliated entity, Obon Corp., to smuggle restricted Nvidia AI GPUs to China, with some shipments potentially reaching Alibaba. This operation, estimated at $2.5 billion in sales, involved circumventing U.S. export controls through deceptive practices like relabeling server chassis. While Alibaba denies involvement, the case highlights concerns about the effectiveness of export control enforcement in Southeast Asia. AI

Nous Research's Hermes Agent has surpassed OpenClaw to become the leading open-source AI agent on OpenRouter's global rankings as of May 10, 2026. Hermes is currently processing 224 billion daily tokens, exceeding OpenClaw's 186 billion, and represents a different architectural approach focused on self-improvement through a "do, learn, improve" loop. This shift is notable as OpenClaw's founder joined OpenAI and the project transitioned to an independent foundation sponsored by OpenAI. AI

Baidu has released ERNIE 5.1, an updated foundation model built on ERNIE 5.0, which improves capabilities in search, reasoning, Q&A, writing, and agent functions. This new model boasts significantly enhanced efficiency, utilizing only about 6% of the pre-training cost compared to similar models. Separately, Mark Andreessen criticized the tendency to label those with positive AI experiences as having 'AI psychosis,' distinguishing it from 'AI cope' and highlighting societal misunderstandings of AI interaction. He also noted that younger generations are gaining 'AI superpowers' through AI, signaling a new era of augmented human capabilities. AI

Google has released Gemini 3.1 Flash, a cost-effective model optimized for high-volume, low-latency AI applications. This new model aims to provide efficient performance for tasks requiring quick responses and affordability. It represents Google's effort to cater to a wider range of AI use cases where cost and speed are critical factors. AI

Chinese AI startup DeepSeek has released its new V4 model, aiming to compete with global leaders. The company previously disrupted the market by offering advanced AI capabilities at a significantly lower cost than its rivals. This latest release marks their return after a year, signaling continued ambition in the AI race. AI

Transsion has invested in Future Smart, a company that produces AI meeting earbuds powered by iFlytek technology. This collaboration is focused on developing advanced AI Agent hardware for global markets. The partnership aims to leverage Future Smart's existing user base of 1.5 million and expand into regions like Africa, Southeast Asia, and Latin America. AI

AI job growth is surging by 10x in 2026, even as other tech companies implement mass layoffs. Research from Anthropic indicates that AI is enhancing productivity across various roles, though it raises concerns for early-career professionals. This trend highlights a significant shift in the labor market, where AI is redefining required skills rather than simply replacing jobs. AI

Anthropic is experiencing significant growth, reportedly achieving 10x year-over-year expansion and a valuation of $1-1.2 trillion, positioning it as a leading global company. This expansion occurs while other tech firms like Block, Coinbase, and Cloudflare have implemented substantial layoffs, often citing AI readiness as a factor. OpenAI has also been active, releasing a suite of GPT-5.5 models across various modalities and products, including GPT-5.5 Cyber for security applications. Additionally, OpenAI is evolving its Codex model into a long-running agent runtime and addressing safety concerns related to its AI development processes. AI

The European Union has reached a provisional agreement to simplify its AI Act, focusing on reducing compliance burdens for businesses. Key changes include a ban on 'nudifier' applications, which are designed to alter images to appear nude. This update aims to streamline the regulatory framework while addressing specific concerns about harmful AI applications. AI

Meta is constructing a $10 billion data center in a small parish in Louisiana, transforming the area into an AI hub. This significant investment is rapidly changing the local landscape and economy. The project is expected to bring substantial development and economic shifts to the region. AI

DeepSeek has released V3-0324, an open-source coding model that matches or surpasses leading models like GPT-4o and Claude 3.5 Sonnet in coding performance. This Mixture-of-Experts model, with 671 billion total parameters and 37 billion active parameters, offers significant cost savings for inference. The model supports a 128K token context window and is available via an OpenAI-compatible API, making it easy for developers to integrate. AI

The United States is investigating allegations of chip smuggling involving Nvidia hardware and Thailand. A company linked to Thailand's national AI initiative, Siam.AI, is suspected of facilitating the illicit transfer of at least $2.5 billion worth of advanced Nvidia chips to China. Siam.AI, which later became Thailand's first official Nvidia Cloud Partner, has connections to a prominent political family in Thailand. AI

Micron has begun shipping its new 6600 ION SSD, boasting a massive 245 terabyte capacity. This release marks the highest-capacity commercially available SSD on the market. The drive's immense storage capability is highlighted as being difficult to fully comprehend. AI

Top researchers are departing OpenAI in 2026, citing concerns about the alignment crisis and the potential for artificial general intelligence (AGI) to surpass human control. This exodus is fueling the creation of independent alignment labs focused on mitigating these risks. OpenAI's own hiring of an 'AI Preparedness Officer' for $555,000 further highlights the internal anxieties surrounding the safety of advanced AI. AI

Intel's stock has experienced a remarkable 490% increase over the past year, driven by Wall Street's optimistic outlook on the company's turnaround. Despite this significant rally, some analysts suggest the stock's performance may be outpacing the company's actual operational progress. The chipmaker faces a challenging path ahead as it works to regain its footing in the competitive tech landscape. AI

AI company Sierra secured $950 million in funding, reaching a valuation of $15 billion. This funding round was among the largest of the week, alongside significant investments in other tech sectors. The investment highlights continued strong financial backing for AI ventures. AI

IBM CEO Arvind Krishna stated that quantum advantage is now closer than ever, highlighting quantum computing as a major theme at IBM Think. Several early adopters, including Cleveland Clinic, Boeing, and Allstate, shared their experiences using the technology at the conference. The advancements discussed have implications for enterprise AI applications. AI

Datacenter developers are increasingly choosing to build in unincorporated county land, bypassing city limits and their associated regulations. This strategic move allows them to avoid city council approvals, municipal zoning votes, and urban land-use reviews. The shift is driven by the desire for fewer municipal constraints, larger parcels of land, and space for essential infrastructure like substations and backup power. AI

Chinese AI company Deepseek is reportedly planning a record-breaking funding round of up to $7.35 billion. Concurrently, Core Automation, a startup founded only six weeks ago by an ex-OpenAI researcher, is aiming for a $4 billion valuation. These significant financial moves highlight continued strong investor interest in the AI sector. AI

Anthropic is collaborating with TrendAI to leverage AI for autonomous vulnerability discovery in Claude Opus 4.7. Separately, Google is enhancing its Search AI Overviews with five new features, including improved source attribution, and is expanding its Google Health Coach globally, powered by DeepMind and Gemini. OpenAI's ChatGPT is also seeing advancements, with efforts to refine its image generation capabilities. AI

SoftBank is reportedly planning to reduce its loan to OpenAI by 40%, signaling potential concerns about the AI company's valuation. This move suggests that lenders are becoming more skeptical of private assessments of OpenAI's worth. The reduction is expected to occur in 2026, reflecting a broader trend of questioning the high valuations of some AI firms. AI

The enterprise AI market is experiencing a significant acceleration in 2026, driven by a surge in acquisitions and strategic alliances among major tech companies and startups. This trend is reshaping the landscape of the burgeoning $1 trillion market. Reports from TechCrunch and Caphatch highlight this rapid transformation, with a particular focus on how companies like OpenAI and Cohere are leveraging AI to redefine their strategic value. AI

Sony is exploring generative AI's potential in video production through a collaboration with Bandai Namco Holdings. The company is also integrating AI tools into its PlayStation studios to enhance game development, including a facial animation tool called Mockingbird that reportedly speeds up production by a thousandfold. Despite these advancements, Sony's recent earnings report showed a significant 46% drop in PlayStation 5 sales, highlighting the company's push towards AI as a strategic business transformation. AI

Baidu's DuMate agent has achieved top rankings on two key benchmarks, PinchBench and DeepResearch Bench. On PinchBench, which evaluates multi-step reasoning and tool use in real-world scenarios, DuMate secured the top two positions, surpassing models from Anthropic and OpenAI. The agent's success is attributed to its end-to-end collaborative Harness architecture, which intelligently handles tasks locally or in the cloud and optimizes context assembly. DuMate also led the DeepResearch Bench, designed for complex research tasks, showcasing its advanced information retrieval and analysis capabilities. AI

The Middle East was poised to become a major hub for data center development, attracting significant global investment due to cheap energy, available capital, and government support. Nations like the UAE, Qatar, and Saudi Arabia had launched national AI strategies and were seen as prime locations for building AI infrastructure. However, recent geopolitical escalations, including drone and missile attacks on data centers in the UAE and Bahrain, have introduced significant uncertainty. AI

Synera, a company focused on agentic AI for engineering, has appointed Ubaldo Rodriguez as its first Chief Revenue Officer. This move is intended to accelerate the company's expansion into the United States market. The appointment signals a strategic push for growth and increased market presence. AI

Major technology firms are proposing unprecedented investments in SK hynix's memory chip production, including funding new fabrication lines and purchasing expensive EUV machines. These offers aim to secure supply amid severe AI-driven shortages, where available capacity is reportedly at zero. SK hynix is cautiously considering these proposals, which differ significantly from traditional memory industry practices. AI

Baidu is reportedly planning a dual stock market listing in Asia, focusing on its artificial intelligence chip business. This move aims to capitalize on the growing demand for AI hardware and potentially raise significant capital. The company's strategic direction indicates a strong commitment to advancing its AI chip technology and expanding its market presence. AI

AI-powered children's toys are rapidly proliferating with minimal regulation, raising concerns among consumer groups and researchers. These toys, ranging from plush companions to interactive robots, have been found to discuss inappropriate topics like sex, drugs, and violence, with some even spouting political talking points. Beyond content issues, studies are exploring the potential negative impacts on children's social and language development, as the toys' conversational abilities and one-to-one interaction models may hinder crucial developmental milestones. AI

Google has unveiled the Fitbit Air, a minimalist, screenless health tracker designed for continuous wear. This small sensor, which fits into interchangeable bands, focuses on core health metrics like heart rate, blood oxygen, and sleep tracking, powered by AI through the new Google Health Coach. The device aims to be unobtrusive, particularly for sleep monitoring, and can sync data with a Pixel Watch. It is priced at $99.99 and will be available starting May 26th. AI

Kunluncore, an AI chip spinoff from Baidu, has officially filed for an IPO on Shanghai's STAR Market, alongside a concurrent filing for a Hong Kong listing on January 1st. The company announced its P800 GPU cluster, featuring 32,000 GPUs, as China's first self-developed cluster capable of handling trillion-parameter AI models. This move aims to gauge investor interest in domestically produced AI hardware. AI

OpenAI has introduced three new real-time voice models that leverage GPT-5 level reasoning to significantly reduce speech translation costs by over 80%. These advancements aim to provide highly accurate multilingual live interpretation. The company's substantial investment in AI infrastructure is also noted, contributing to a decade-low free cash flow for Big Tech. AI

Anthropic's valuation has reportedly surged past OpenAI's, reaching nearly $1 trillion in secondary market deals. This significant increase is attributed to high demand for its Claude AI models and strong investor confidence in its enterprise applications. Meanwhile, a separate report highlights issues with age verification systems, including those used by OpenAI, which are failing adults by imposing childlike restrictions. AI

Anthropic has unveiled a new technology called Natural Language Autoencoder (NLA) designed to translate the internal 'thoughts' of AI models into human-readable language. This development aims to provide greater insight into AI decision-making processes. Separately, Google is reportedly testing an AI agent named Remy, which could function as a 24/7 personal assistant powered by Gemini. Additionally, Anthropic and SpaceX have announced a partnership, signaling an intensifying race in AI development. AI

Tencent's Hy3 Preview model has achieved the top position on the weekly rankings of OpenRouter, just two weeks after its release. Separately, Alibaba's Qwen3.6 model now supports native MTP, a feature for which Google recently added support to its Gemma4 model. AI

Rongxin Zhiyuan, an AI infrastructure company founded by Tsinghua University alumni, has secured hundreds of millions of yuan in an angel funding round. The company is developing its novel AGC architecture, which positions GPUs as the core processing units and CPUs as peripheral controllers. This shift aims to significantly increase the GPU-to-CPU ratio, enhance memory consistency, and enable hot-swappable GPU fault tolerance for improved efficiency and reduced downtime in AI computing systems. AI

Nvidia has released validated blueprints for AI data centers, detailing configurations for 4-node to 128-node clusters. These designs, named RTX PRO, HGX, and NVL72, are intended for advanced applications like agentic AI and handling trillion-parameter models. Separately, a study indicates a significant drop in accuracy for multimodal LLMs like GPT-4.1 when applied to real-world medical cases compared to benchmark performance. AI

Pit, a new AI startup founded by the co-founders of the European scooter company Voi, has secured $16 million in seed funding led by Andreessen Horowitz (a16z). The Stockholm-based company aims to provide an "AI product team as a service" for enterprises, developing custom software to automate internal business processes. Pit differentiates itself with its Pit Studio and Pit Cloud offerings, focusing on governance, certifications, and auditability for enterprise clients. AI

SpaceX has leased its massive Colossus 1 data center, equipped with 220,000 Nvidia GPUs, to Anthropic for an estimated $3-4 billion annually. This deal, which includes a clause allowing SpaceX to reclaim compute if Anthropic's AI harms humanity, strategically positions SpaceX as a major AI infrastructure provider ahead of its IPO. The agreement highlights compute capacity as a critical bottleneck in the AI industry, forcing even AI safety leaders like Anthropic to secure necessary resources from any available vendor, regardless of past public criticism. AI

The Gemini 3.1 Flash-Lite model is now generally available, having transitioned from its preview status. This release follows a March preview of the model, with no apparent changes noted since then. The announcement was made on May 7, 2026, and is part of ongoing discussions about AI developments. AI