Google unveiled Gemini 4 Argon, its first frontier model in seven months, with initial access restricted to trusted cybersecurity partners. The model matches GPT-6 Astra in benchmarks but faces internal employee skepticism over performance.
Anthropic is seeking to go public as soon as mid-November at a valuation of $25-26 billion, according to multiple reports. The AI model maker pushed back earlier plans and will hold a pre-IPO investor day on October 14. Reuters obtained exclusive details from the confidential S-1 filing.
Palo Alto Networks unveiled a new cybersecurity service for businesses using advanced AI models from Anthropic and OpenAI to help identify security vulnerabilities across corporate networks around the clock.
Alibaba Group Holding introduced what it calls China's most powerful AI chip, the Zhenwu M890, which triples the performance of its predecessor. The company also announced plans to train an AI model with up to 10 trillion parameters, driving shares up roughly 5%.
More than 100 academics, evaluators, and AI leaders signed a public letter demanding truly independent third-party safety evaluators at companies like OpenAI and Anthropic, with transparent conditions for their work.
Data center operator Crusoe has secured a multiyear cloud capacity agreement with AI search company Perplexity AI, providing access to Nvidia GB300 processors to support the startup's model training and inference workloads.
Chinese AI company DeepSeek launched its V4.1-Flash model on September 10, 2026, featuring a 552-billion-parameter architecture that cuts memory requirements by 75% and offers inference pricing as low as $0.003 per million tokens. The model claims to match or exceed performance of competitors like GPT-6 Astra, Claude Opus 5, and GPT-5.6 Sol on various benchmarks while operating at a fraction of the cost.
Chinese AI company DeepSeek launched its V4.1-Flash model on September 10, 2026, featuring a 552-billion-parameter architecture that cuts memory requirements by 75% and offers inference pricing as low as $0.003 per million tokens. The model claims to match or exceed performance of competitors like GPT-6 Astra, Claude Opus 5, and GPT-5.6 Sol on various benchmarks while operating at a fraction of the cost.
DeepSeek has engaged CITIC Securities as an underwriter to prepare for an initial public offering on Shanghai's STAR Market, signaling the AI startup is advancing its domestic listing plans. The Hangzhou-based firm aims to begin the IPO process this year.
H Company has launched NeoMME, a family of 260M and 800M parameter single-tower multimodal encoders designed for visual document retrieval. The models represent a departure from existing approaches that repurpose generative vision-language models as encoders.
H Company has launched NeoMME, a family of 260M and 800M parameter single-tower multimodal encoders designed for visual document retrieval. The models represent a departure from existing approaches that repurpose generative vision-language models as encoders.
Thinking Machines, the AI lab founded in early 2025 by former OpenAI CTO Mira Murati, is negotiating to raise at least $1 billion in funding at a $40 billion-plus pre-money valuation, with Accel reportedly leading the round.
Google DeepMind and Google Research launched WeatherNext 3, an AI weather forecasting model that learns directly from real-time satellite data rather than physics simulations. The model delivers hourly forecasts at 5-kilometer resolution—five times more detailed than its predecessor—and is designed to improve predictions for renewable energy production, precipitation, and localized weather events.
Nvidia has agreed to acquire open-source AI platform Hugging Face for $12.93 billion, marking one of the chipmaker's largest acquisitions as it expands beyond hardware into AI model distribution and software. Hugging Face CEO approached Nvidia's Jensen Huang weeks ahead of the deal announcement.
OpenAI announced the release of GPT-6, which it describes as its most advanced AI model yet, to select customers. The rollout comes amid growing concerns over AI safety following recent security incidents and increased regulatory pressure.
Broadcom CEO Hock Tan revealed during the company's Q3 2026 earnings call that Anthropic is the company's largest XPU customer. Broadcom reported $16.7 billion in AI semiconductor revenue and forecasts AI revenue reaching $230 billion by fiscal 2028.
Nvidia announced NVHBM, a new memory architecture that relocates the HBM memory controller from the AI accelerator die into the memory stack itself, freeing up to 25% more compute silicon per chip. The move addresses AI infrastructure bottlenecks as trillion-parameter models become mainstream, while Nvidia expands its NVLink licensing strategy amid growing custom chip competition.
Google Research has released TimesFM-3, a 330 million parameter foundation model designed for zero-shot forecasting of multiple related time series simultaneously in a single forward pass. The model represents an advancement over previous TimesFM checkpoints through version 2.5.
China's Zhipu AI confirmed that its viral Ox Alpha model, now revealed as GLM-5.3-Flash, is a 320B-parameter mixture-of-experts AI system that ran entirely on 100,000 domestically produced Chinese chips during launch. The geopolitically significant claim about Chinese chip capabilities remains unverified by independent sources.
AI startup Anthropic has agreed to rent $45 billion in cloud computing power from Nscale's West Virginia data center campus as it prepares for a major initial public offering. The deal with the British AI infrastructure company, founded in 2024, represents a massive expansion of Anthropic's computing capacity.
Chinese AI startup Moonshot AI is in talks with Microsoft, Amazon, and Google to host its K3 model on their cloud platforms through revenue-sharing agreements. This would mark the first time a Chinese AI company's model appears on major US cloud infrastructure.
Chinese AI startup DeepSeek is close to completing a pre-IPO funding round at a $74 billion valuation, with revenues reaching $70.7 million through July 2026, representing a tenfold increase. The company is targeting a 2027 public listing amid reports that Chinese hackers are integrating its open-source AI models into cyberattack operations.
A new AI model called Ox Alpha is gaining significant developer adoption after appearing online for free, though the identity of its creator remains unknown. Multiple outlets report developers are impressed with its capabilities.
New research reveals that current AI world models like Sora and Genie fail to predict human actions because they only simulate physics while ignoring beliefs, intentions, and feelings. A "Mental World Modeling" framework that incorporates these mental variables shows weaker language models can outperform stronger models that lack this capability.
New research reveals that current AI world models like Sora and Genie fail to predict human actions because they only simulate physics while ignoring beliefs, intentions, and feelings. A "Mental World Modeling" framework that incorporates these mental variables shows weaker language models can outperform stronger models that lack this capability.
Chinese AI company DeepSeek released an experimental multimodal language model with visual comprehension capabilities, claiming its performance approaches that of Anthropic's advanced Opus 4.8 model. According to DeepSeek's published benchmarks, the new V4 Flash Vision model outperforms Opus 4.8 on three out of several tested metrics.
Meta Platforms has emerged as one of Microsoft's biggest AI customers, joining a roster that includes Adobe, ByteDance (TikTok's parent company), and Perplexity. The development highlights how demand for AI technology remains concentrated among major tech companies, even as Meta simultaneously develops its own open-weight AI models.
IBM announced a strategic partnership with OpenAI to embed frontier models like GPT-5.6 into IBM Consulting's AI delivery platform, helping enterprises deploy AI at scale across core business operations. The announcement comes alongside a separate $240M deal between IBM Cloud and Together AI for Nvidia infrastructure deployment.
A new Nature Medicine paper describes a system of integrated multiscale AI foundation models that could simulate how potential drugs or genetic changes move through molecules, cells, and eventually people. The digital organism aims to accelerate biological and medical research by modeling complex biological processes in silico.
Coinbase, Block, BitGo and dozens of other cryptocurrency companies signed a Bitcoin Policy Institute letter calling on AI labs to provide qualified security researchers with access to advanced AI models, arguing that current safety guardrails block legitimate defenders while attackers face no such restrictions.
Jeff Dean, Google's longtime chief AI architect and a pioneering figure in the company's AI development, is leaving to start his own AI venture that is already in talks for a $10 billion valuation. His departure comes amid a major AI organizational reshuffle at Google involving DeepMind and the Gemini project.
Tencent Holdings nearly tripled its capital expenditure in Q2 2026 to accelerate AI investments, reporting a 176% increase in capex that pushed free cash flow negative. The spending surge comes as usage of the company's Hy3 AI model increased more than 68 times compared to its predecessor, with Tencent planning to release a larger Hy4 model.
Pathway has secured funding at a $500 million valuation, backed by Id4 Ventures, TQ Ventures, Red Bridge Ventures, Kadmos Capital, WS Investment Co, and Databricks chief AI scientist Jonathan Frankle. The company is challenging AI reasoning cost curves with its 150-million-parameter model.
AT&T and other major corporations are increasingly adopting open-weight AI models over proprietary alternatives to reduce costs, with AT&T processing 45 billion AI tokens daily. The shift comes as American tech firms respond to competition from cheaper Chinese open AI models that rival offerings from Anthropic and OpenAI.
ByteDance has established a new top-level department focused on AI data and safety, while simultaneously training a massive AI model with up to 10 trillion parameters that could rival Anthropic's Mythos system in scale.
Junyang Lin, who left Alibaba in March after leading its flagship AI model development, has founded Shanghai-based startup Pragmatik Labs to build next-generation AI agents. The company has secured backing from Tencent Holdings and Chinese venture capital firms.
Novo Nordisk and Amazon Web Services have established a strategic partnership featuring a co-innovation hub in London where AWS engineers work alongside Novo Nordisk R&D teams. The collaboration deploys agentic AI, biological foundation models, and Amazon Bio Discovery platform to accelerate drug discovery workflows including target identification and therapy design.
River AI, founded by xAI co-founder Igor Babuschkin just two months ago, has secured $1.1 billion in a seed/Series A funding round led by General Catalyst and AMP PBC, with participation from Nvidia and AMD. The startup aims to make customized AI models faster, cheaper, and more accessible to developers and companies.
Meta announced the release of Muse Glimmer, a new open-source AI model designed to run on laptops, and plans to release an open-weight version of its most advanced Muse Spark 1.2 model. Mark Zuckerberg published a 6,500-word essay advocating for American open-source AI development as a counter to Chinese rivals.
Palantir CEO Alex Karp has launched a public attack on major AI model developers, accusing them of attempting to make businesses dependent on their platforms and warning about broader strategic implications for U.S. competitiveness against China in the AI race.
Chinese AI companies are confronting a severe shortage of high-quality training data for developing next-generation language models, with Chinese accounting for only 1.3% of global web content compared to 49% for English. Beijing plans to create national datasets by 2028 to address the bottleneck.
TikTok parent company ByteDance is training a massive AI model with up to 10 trillion parameters, approximately three times larger than Moonshot's Kimi K3, in an effort to compete with Anthropic's advanced Mythos system. The development intensifies the global competition in large-scale artificial intelligence.
Amazon is redesigning its AI data center campus in rural Indiana, consolidating multiple facilities into a single powerful cluster while deprioritizing most of its Nova models to focus engineering talent and computing resources on developing next-generation frontier AI systems.
Researchers have developed an artificial intelligence foundation model that analyzes routine polysomnography data to identify patients' long-term health risks, with high-risk patients showing twice the mortality risk over five years compared to low-risk patients. The study, published in Nature Communications, demonstrates that sleep studies contain far more predictive health information than currently utilized by clinicians.
Moonshot AI secured access to approximately 20,000 Nvidia Hopper generation chips through a computing agreement with Alibaba's cloud services to train its Kimi K3 AI model, demonstrating China's continued access to advanced US chip technology despite export restrictions.
Military researchers in China have utilized outputs from leading American artificial intelligence models developed by OpenAI and Anthropic to train domestic AI systems aimed at advancing China's defense capabilities, according to an exclusive Reuters report.
Researchers have developed a neuroimaging AI foundation model trained on routine clinical CT and MRI scans that achieves state-of-the-art diagnostic performance, highlighting a shift toward using real-world health system data over controlled trial data—a trend that regulatory frameworks have not yet addressed.
Thinking Machines has released Inkling-Small, an open-weights Mixture-of-Experts transformer model with 276 billion parameters that achieves performance comparable to its larger predecessor Inkling while being one-quarter the size.
Beijing-based AI startup Moonshot AI has secured $3.5 billion in funding at a $35 billion valuation, far exceeding its initial $1-2 billion target. The round follows the company's release of its K3 model, which approached the performance of frontier models from OpenAI and Anthropic.
Anthropic has launched Claude Opus 5, a new AI model that approaches the intelligence capabilities of its Claude Fable 5 model while being offered at half the price. The model is designed for coding and knowledge work tasks.
Chinese AI company Moonshot AI has released the full weights of its Kimi K3 model, making the 2.8 trillion parameter system freely downloadable. The release raises questions about deployment practicality and trust in Chinese open-weight models despite the technical milestone.
NVIDIA has formed a long-term strategic partnership with Safe Superintelligence Inc. (SSI), the AI safety startup founded by former OpenAI co-founder Ilya Sutskever. The deal includes an investment in SSI and provides the company access to NVIDIA's Vera Rubin computing systems, expanding SSI's compute capacity by an order of magnitude.
Multiple leaked reports indicate both Anthropic's Fable 5.1 and OpenAI's GPT-6 are targeting August 2026 releases, with OpenAI CEO Sam Altman reportedly briefing the Trump administration on GPT-6's advanced capabilities including autonomous research and penetration testing.
The launch of Moonshot AI's Kimi model has intensified divisions in Silicon Valley over American AI competitiveness and whether to support open-source AI or restrict Chinese open-source models. Chinese models from startups like DeepSeek and Moonshot AI are gaining U.S. adoption due to their affordability and efficiency.
Meta is upgrading its Meta AI assistant with new autonomous task automation capabilities, including calendar integration, daily briefings, and recurring task management. The features, powered by the Muse Spark 1.1 model, are being rolled out in select markets.
Etched, an AI chip startup founded by three Harvard dropouts in 2022, has closed a $300 million Series C funding round led by Sequoia Capital, more than doubling its valuation to $10.3 billion. The company specializes in chips optimized for transformer-based AI models.
AMD CEO Lisa Su announced that the Helios rack-scale AI infrastructure system was in full production. AMD claims Helios delivers up to 15% more peak theoretical FP4 compute than Nvidia's Vera Rubin NVL72, based on AMD calculations. The system integrates 72 Instinct MI455X GPUs, 18 EPYC “Venice” CPUs and AMD Pensando networking for large-scale AI training and inference.
Advanced Micro Devices and Anthropic have announced a major partnership involving up to $5 billion in investment and the deployment of up to 2 gigawatts of AMD's upcoming Instinct MI450 series GPUs, marking AMD's largest AI infrastructure deal and a significant challenge to Nvidia's dominance in the AI chip market.
South Korea's Samsung Electronics is negotiating to invest between hundreds of millions and up to €1 billion in French AI company Mistral AI as part of a funding round that would value the European startup at approximately €20 billion ($22.81 billion), according to Financial Times reports citing sources.
Alphabet's Google released three new versions of its Gemini AI model on July 21, 2026—Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, and Gemini 3.5 Flash Cyber—while providing no timeline for its flagship Gemini 3.5 Pro model, which has missed its June launch target by several weeks.
Microsoft and French AI startup Mistral AI announced a multibillion-dollar deal on Tuesday that will see Microsoft fund Mistral's computing infrastructure and data center expansion in Europe, while expanding distribution of Mistral's AI technology through Microsoft's cloud platform.
The University of Tennessee Research Foundation filed a patent infringement lawsuit against Anthropic in Delaware federal court, alleging that the company’s AI systems infringe two patents covering neural-network and neuroscience-inspired machine-learning technology.
Alibaba Group previewed its new Qwen3.8 artificial intelligence model, claiming it ranks second only to Anthropic's Claude Fable 5 in performance. The company plans to make the model open-weight soon, allowing developers to download and modify it.
The Trump administration is implementing new controls over access to frontier AI models through executive action, fundamentally shifting power dynamics from tech companies to the federal government in determining who can use advanced AI systems.
Meta Platforms is reportedly in early-stage negotiations to lease computing infrastructure to Anthropic in a potential deal worth up to $10 billion over two years, which would provide the AI startup with data center capacity to run its Claude AI models.
Moonshot AI unveiled its Kimi K3 model, described as the world's largest open-weights AI system, which the company claims matches or exceeds OpenAI and Anthropic's leading models on some benchmarks. The release triggered market volatility in AI and semiconductor stocks as investors drew comparisons to last year's 'DeepSeek moment.'
Released July 15, 2026 by Thinking Machines Lab, Inkling is a 975-billion parameter open-weights Mixture-of-Experts model with 41 billion active parameters, a 1-million-token context window, and native multimodal input support across text, image, audio, and video.
