AI News Archive: August 10, 2026 — Part 20
Sourced from 500+ daily AI sources, scored by relevance.
- FirstFT: Wall Street giants partner with Nvidia on $500bn AI funding package
Also in today’s newsletter: yen sinks as effect of US-Japan intervention fades and China’s ‘Ice Silk Road’
- Nvidia to team with Wall Street on US$500 billion package for AI infrastructure projects
A group of US investment giants are partnering with Nvidia on US$500 billion in funding for AI infrastructure projects, the Financial Times reported. Apollo Global Management, Blackstone, BlackRock’s Global Infrastructure Partners, Brookfield Asset Management, Goldman Sachs and KKR are among the firms in talks with Nvidia on a deal to invest in the AI buildout, the Financial Times reported, citing unidentified sources. The deal may be announced as soon as Monday, the Times said. The named firms...
- Chart: PE consortium weighs $500 billion AI infrastructure financing for Nvidia
Chart: PE consortium weighs $500 billion AI infrastructure financing for Nvidia PitchBook
- Nvidia, Wall Street giants partner to raise $500-billion for AI infrastructure
Nvidia signed MoUs with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR
- Nvidia partners with lenders to finance AI infrastructure
The chip giant joined six major Wall Street firms to lend more than $500 billion.
- Top Wall Street Firms Reportedly Partnering With Nvidia for $500 Billion AI Investment
Get ready to hear the phrase "circular dealmaking" a lot more, again.
- OpenAI Buys Back $7 Billion of Employee Shares in Tender Offer
OpenAI has completed a deal to help employees sell roughly $7 billion worth of shares in the company ahead of a possible Wall Street debut, according to a person familiar with the matter.
- Corma raised $60M from Sequoia to build the defensive AI that cybersecurity is missing
In hundreds of simulations modelled on Fortune 500 companies with dozens of security tools, Corma tested leading AI models including GPT and Claude. First, the models attacked the simulated organisations and planted persistent threats. Then the same models were asked to defend and find what they had planted. The attackers succeeded in 88% of simulations. […] This story continues at The Next Web
- Meta to open source its most powerful AI model as it takes swipe at OpenAI, Anthropic
Meta launched Muse Glimmer and plans to release Muse Spark 1.2 weights as Zuckerberg pushes for U.S. leadership in open AI.
- CleanQuote.ai
Quote cleaning jobs from customer photos.
- Corma launches with $60M in funding for defensive cybersecurity AI
Defensive cybersecurity startup Corma Labs Ltd. today announced it has raised $60 million in seed funding to build a foundation model purpose-built for security defense. Founded in 2025, Corma runs offices in Tel Aviv and San Francisco and describes itself as a frontier artificial intelligence lab working only on the defensive side of security. Its […] The post Corma launches with $60M in funding for defensive cybersecurity AI appeared first on SiliconANGLE .
- Oslo-based Visoid raises €2.2 million to scale its AI visualisation platform for architects
Visoid, an Oslo-based AI visualisation platform for architects and industry professionals, has secured €2.16 million ($2.5 million) in fresh funding to support the company’s next stage of growth. The round was led by Skyfall Ventures, with participation from existing investors StartupLab, OBOS and Antler, alongside new investors Farvatn and byFounders Angel Collective. “Architects have been […] The post Oslo-based Visoid raises €2.2 million to scale its AI visualisation platform for architects appeared first on EU-Startups .
- China’s Moore Threads plans Hong Kong listing
The company said the move would support its global expansion and help it attract research, development, and management talent.
- London-based edge AI infrastructure company Edgify raises €7.7 million to combat retail losses
Edgify, a London-based edge MLOps platform for physical retail providing the AI infrastructure to protect against in-store loss, has raised €7.7 million ($9 million) in Series A+ funding to accelerate the rollout of its platform. The round was backed by Rank Ventures and Mangrove Capital Partners, bringing the company’s total funding to €21.6 million ($25 […] The post London-based edge AI infrastructure company Edgify raises €7.7 million to combat retail losses appeared first on EU-Startups .
- As AI-led attacks multiply, OpenAI launches a new cyber model
OpenAI is expanding its AI cybersecurity defense program Daybreak, and rolling out a new cyber-trained AI model with it.
- OpenAI introduces a new cyber model amid fears of AI cyberattacks
OpenAI is introducing a more cyber-permissive version of GPT-5.6 Sol to vetted defenders as it prepares companies for autonomous cyberattacks . Why it matters: The move comes just days after OpenAI said it was delaying the release of its forthcoming model, Astra, after it reached critical hacking abilities during safety testing. The big picture: OpenAI is unveiling GPT-5.6-Cyber while also expanding Daybreak , its program that gives cybersecurity defenders access to the company's cyber models and other tools. Many cyber defenders have been experiencing high refusal rates across frontier AI models as the labs try to balance giving defenders the tools they need, while not accidentally leaking those abilities to malicious hackers. Under the new program, Daybreak will have two tiers: Daybreak Blue, which includes access to GPT-5.6 Sol without its system-level cyber guardrails; and Daybreak Red, which offers access to GPT-5.6-Cyber to validate exploits and do more advanced vulnerability research. OpenAI is also expanding how program members can use its tools, allowing companies like Accenture, IBM, CrowdStrike, Cisco and Palo Alto Networks to incorporate the models into security products, managed services and work with customers. Zoom in: During testing, GPT-5.6-Cyber responded to 95% of requests tied to advanced cybersecurity work, including prompts related to exploit-chain development, authentication bypass and privilege escalation. GPT-5.6-Sol only responded to 1.5% of requests, and the version of the model that defenders get through Daybreak Blue responded to just 2%. Yes, but: Unlike Astra, GPT-5.6-Cyber only reached the "High" cyber capability threshold under OpenAI's Preparedness Framework , the company said. State of play : OpenAI's new model comes as it continues to investigate how its tools hacked Hugging Face . Last week at the Black Hat cybersecurity conference, two OpenAI employees said the agents created a message board where they left information about vulnerabilities they found that ultimately helped them break into Hugging Face. What's next: Both OpenAI and Anthropic have been rolling out tools designed to help defenders find vulnerabilities and secure their code — and they're each eying ways to expand on cyber product offerings. Go deeper: Anthropic and OpenAI are now cybersecurity's kingmakers This story is developing.
- OpenAI launches GPT-5.6-Cyber to help defenders find vulnerabilities before attackers do
OpenAI aims to give cybersecurity defenders a head start: The new GPT-5.6-Cyber model answers up to 98.5 percent of security queries that would otherwise be blocked and has already uncovered two previously unknown Chrome vulnerabilities. According to OpenAI, the window of opportunity for defenders is shrinking. Access requires identity verification. The article OpenAI launches GPT-5.6-Cyber to help defenders find vulnerabilities before attackers do appeared first on The Decoder .
- OpenAI flags possible critical cybersecurity risk in upcoming model Astra
Preliminary evaluations over the past several days, along with outside expert assessments, indicated Astra may be capable of performing increasingly sophisticated cyber tasks autonomously
- OpenAI Pauses Work on AI Model Over Serious Cybersecurity Risks
OpenAI Pauses Work on AI Model Over Serious Cybersecurity Risks PCMag UK
- Meta’s new Glimmer AI model offers a hint at Zuckerberg’s personal intelligence vision
Meta’s new open-weight Muse Glimmer model offers a glimpse of Mark Zuckerberg’s personal superintelligence vision, as well as the emerging divide between AI users can own and access.
- Meta launches new AI model as Zuckerberg champions open-weight push
Meta launches new AI model as Zuckerberg champions open-weight push Reuters
- Meta's latest model advances Zuckerberg's vision for personal AI assistants
Every weekday, the Investing Club releases the Homestretch; an actionable afternoon update just in time for the last hour of trading.
- Meta Unveils an Open Version of Its Most Powerful A.I. Model
The release of Muse Glimmer, a model that can be freely downloaded and modified, is likely to intensify a debate over whether A.I. should be restricted.
- What smart people are saying about Meta's open-weight model Muse Glimmer and Mark Zuckerberg's AI manifesto
What smart people are saying about Meta's open-weight model Muse Glimmer and Mark Zuckerberg's AI manifesto Business Insider
- Meta launches Muse Glimmer, an open-weight model that can run on a laptop
Meta launches Muse Glimmer, an open-weight model that can run on a laptop Business Insider
- Zuckerberg pushes ‘superintelligent’ AI for all as Meta drops open-source model
Meta CEO presents utopian vision of AI in 6,000-word essay amid Silicon Valley debate over government regulation Mark Zuckerberg published a lengthy essay on Monday detailing his views on artificial intelligence and announced several plans for how Meta would develop the technology in the future. The CEO’s essay went online the same day as Meta released a new, open-source AI model that seeks to rival Anthropic and OpenAI’s products called Muse Glimmer. Over the course of more than 6,000 words in a post titled “The Future is for Everyone,” Zuckerberg addressed a range of topics related to AI that included datacenters, government regulation, cybersecurity, the creation of bioweapons, labor market disruption, surveillance powers and more. The essay presented a utopian vision of AI as a personalized “superintelligence” – using the word 60 times. Continue reading...
- Meta launches new AI model as Zuckerberg champions open-weight push
Meta CEO Mark Zuckerberg urged the US to reduce policy barriers around open-weight AI as the company unveiled a new model designed to run agentic tasks on personal computers
- Meta AI Releases Muse Glimmer: A 30B Open-Weights Agentic Model That Runs on One Consumer GPU
Meta AI Releases Muse Glimmer: A 30B Open-Weights Agentic Model That Runs on One Consumer GPU MarkTechPost
- Meta Reverses Course with Open-Weight Muse Glimmer
This shift responds to enterprise demand for tighter data control and local infrastructure systems, in contrast to Meta's recent focus on closed models such as Muse Spark 1.
- Meta is back with Muse Glimmer: local, agentic, multimodal, and open source
Meta is back with Muse Glimmer: local, agentic, multimodal, and open source
- Muse Glimmer from Meta Superintelligence Labs is now available
Meta's Muse Glimmer, a 30B multimodal model under Apache 2.0, supports local coding agents and image input via Ollama's MLX engine.
- Muse Glimmer: Meta says its 30B newcomer beats Qwen3.6 on local jobs
Distilled from Muse Spark to let consumer-grade hardware run agentic tasks, Meta is pitching Glimmer as a drop-in replacement that lets you go local.
- Meta launches new AI model as CEO Mark Zuckerberg champions open-weight push
Meta launches new AI model as CEO Mark Zuckerberg champions open-weight push The Straits Times
- Meta launches new AI model as Zuckerberg lays out vision for open-weight AI
Meta launches new AI model as Zuckerberg lays out vision for open-weight AI CBC
- Meta's 'open source' Muse Glimmer model can run on a single computer
Meta has released a new slimmed down 'open source' AI model that's light enough to run on a single computer.
- Meta releases open-source Muse Glimmer model with 30B parameters
Meta Platforms Inc. today released Muse Glimmer, an open-source language model that can run on personal computers. The company also published a lengthy essay penned by Chief Executive Officer Mark Zuckerberg. The document discusses the risks of artificial intelligence, open-source model regulations and several related topics. Muse Glimmer features 30 billion parameters, which means that it would […] The post Meta releases open-source Muse Glimmer model with 30B parameters appeared first on SiliconANGLE .
- Meta, Mark Zuckerberg launch new Muse Glimmer AI model
Meta, Mark Zuckerberg launch new Muse Glimmer AI model USA Today
- Meta launches Muse Glimmer open source AI model for devices
Meta launches Muse Glimmer open source AI model for devices USA Today
- Meta, Mark Zuckerberg launch new Muse Glimmer AI model
Meta, Mark Zuckerberg launch new Muse Glimmer AI model azcentral.com and The Arizona Republic
- Meta Stock Climbs On Open-Source Model Launch; Zuckerberg Criticizes AI Rivals
Meta stock climbed after the company's Muse Glimmer release and an essay on the future of AI from Mark Zuckerberg. The post Meta Stock Climbs On Open-Source Model Launch; Zuckerberg Criticizes AI Rivals appeared first on Investor's Business Daily .
🤖 ModelsAug 10, 2026https://www.investors.com/news/technology/meta-stock-muse-glimmer-zuckerberg-ai/ - Meta’s New Open-Weight Model Can Run AI Agents on Your Laptop
Meta Glimmer is a bit smaller than the company’s closed-weight Muse Spark models.
- Meta’s latest AI model wants to live on your PC
Meta’s Muse Glimmer is a 30B model that can run on your laptop.
- Zuckerberg manifesto pushes an open-source approach on AI as Meta releases its latest model
Meta Platforms on Monday announced the release of a new artificial intelligence model that's open for other software developers to use as CEO Mark Zuckerberg warned of the risks if control of advanced AI is concentrated under a select few companies, institutions or governments.
🤖 ModelsAug 10, 2026https://techxplore.com/news/2026-08-zuckerberg-manifesto-source-approach-ai.html - KGCaRe: Explainable Complex Conditional Question Answering using Automatic Knowledge Graph Construction and Context Retrieval with LLMs
Answering complex conditional questions using Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) remains a challenge, particularly in domain-specific contexts where general-purpose LLMs and RAG tend to underperform. We hypothesize that augmenting RAG with unstructured and structur...
- Defining Decentralization: An Ontological Perspective
Decentralization as a concept in computer science has existed for over half a century. Despite its fundamental role across domains such as security, distributed computing, artificial intelligence, cloud infrastructures, and Internet of Things (IoT) architectures, there remains no universally accepte...
- SR-OPSD: Self-Referenced On-Policy Self-Distillation
On-policy self-distillation (OPSD) converts feedback into dense token-level supervision on trajectories generated by the policy to be optimized, providing a useful complement to reinforcement learning with sparse outcome rewards. However, the self-teacher policy used in OPSD is typically a stop-grad...
- Rethinking Factor Sharing in Federated LoRA: A Rank-Aware Adaptive Approach
Low-rank adaptation (LoRA) represents large language model (LLM) updates with two compact matrix factors, i.e., $A$ and $B$, providing an efficient way to fine-tune large models in federated learning paradigm. Inspired by the asymmetric roles of the LoRA factors, we study whether $A$ should be share...
- ColluSkill: Adversarial Cross-Skill Composition for Evading Agent Skill Scanners
Agent skills are emerging as an important attack surface in LLM-based agent systems. Through an empirical study of existing skill scanners, we find that current defenses mainly inspect individual skills, leaving risks from cross-skill composition insufficiently examined. This creates a practical bli...
- How Do Large Language Models Judge Social Attraction? Evidence from Theory-Grounded Persona Ratings Across Multiple LLMs and Humans
Large language models (LLMs) are increasingly used to perform subjective evaluations traditionally made by humans, yet their validity as social judges remains unclear. This paper examines whether LLMs can assess social attraction from theory-grounded persona profiles constructed from ten psychologic...
- Matryoshka Language Model Suites
Training a language model suite classically requires training each model separately and serving them independently. We improve both training and inference efficiency by stacking sub-models of increasing size into a single nested architecture trained end-to-end. This Matryoshka training framework red...