NVIDIA Launches ComputeDomains Alongside Salesforce AI Acquisition

The artificial intelligence sector is currently experiencing a dynamic period marked by both rapid technological advancements and significant market speculation. NVIDIA, a key player in AI infrastructure, continues to drive innovation. On November 10, 2025, NVIDIA introduced ComputeDomains, a new Kubernetes tool designed to simplify secure and efficient GPU-to-GPU memory operations across multiple computers using multi-node NVLink, a technology tested on DGX GB200 systems. Concurrently, NVIDIA also unveiled Grove, a component within its Dynamo platform, which streamlines the management and scaling of complex AI inference systems on Kubernetes. Grove allows users to define entire inference systems as a single Custom Resource, automating scheduling, placement, and autoscaling across potentially tens of thousands of GPUs, supporting NVIDIA Blackwell systems and H200 GPUs through major cloud providers. Despite these innovations, veteran investor Mark Mobius issued a stark warning, predicting a possible 40 percent crash in AI stocks, labeling the current market a "classic bubble" driven by unsustainable excitement. This sentiment is echoed by tech analyst Scott Galloway, who, also on November 10, 2025, cautioned of a major market crash if issues at OpenAI persist. Galloway highlighted concerns about OpenAI's finances, with its CFO noting they are "burning through cash," and leadership instability involving CEO Sam Altman. He stressed that the high valuations of top tech companies like Nvidia and Microsoft are heavily reliant on generative AI's success, and a failure to deliver expected returns from AI trials could severely impact the global markets, especially given that the top 10 companies constitute 40 percent of the S&P 500. The discussion around an "AI bubble" also points to a fundamental imbalance: the rapid pace of AI software development versus the slower build-out of data center infrastructure. Huge investments, such as Microsoft's reported $100 billion for its Stargate supercomputer, face risks from uncertain demand and critical infrastructure challenges like power grid capacity, creating expensive bottlenecks. Maryland, for instance, aims to lead in AI and quantum technology but recognizes the necessity of also leading in clean energy and energy storage to power the vast electricity needs of AI data centers. Meanwhile, companies are actively integrating AI capabilities through acquisitions and partnerships. Salesforce recently agreed to acquire Spindle AI, an agentic analytics platform, to enhance its Agentforce analytics with AI, machine learning, and data modeling for predictive decision-making. Kaltura, a Nasdaq-listed company, acquired eSelf.ai for $27 million on November 10, 2025, integrating its virtual agent technology to offer human-like conversational AI avatars within video products. In Saudi Arabia, Workey partnered with Falak Investment Hub on November 10, 2025, to launch an AI-powered business solutions platform, offering smart digital bookings and support for startups. Beyond business, AI's influence is expanding into neuroscience, with NeuroAI rapidly growing as AI shapes brain theories and vice versa. However, the increasing regulation of AI also raises legal questions, with some arguing that government restrictions on AI use could potentially violate First Amendment free speech rights, depending on whether they are content-based.

Key Takeaways

• NVIDIA introduced ComputeDomains and Grove on November 10, 2025, to simplify multi-node GPU communication and scale complex AI inference on Kubernetes, supporting Blackwell and H200 GPUs.
• Veteran investor Mark Mobius predicts a potential 40 percent crash in AI stocks, calling the current market a "classic bubble."
• Scott Galloway warned on November 10, 2025, that a market crash could occur if OpenAI's financial and leadership issues persist, impacting top tech companies like Nvidia and Microsoft.
• Microsoft is reportedly investing $100 billion in its Stargate supercomputer, highlighting the massive infrastructure costs and risks associated with AI development.
• The rapid development of AI software is creating a bottleneck with the slower pace of data center and energy infrastructure build-out, as seen in Maryland's need for clean energy to power its AI ambitions.
• Salesforce is acquiring Spindle AI to enhance its Agentforce analytics with AI, machine learning, and data modeling for predictive business decisions.
• Kaltura acquired eSelf.ai for $27 million on November 10, 2025, to integrate virtual agent technology and offer human-like conversational AI avatars in its video solutions.
• Workey partnered with Falak Investment Hub on November 10, 2025, to launch an AI-powered business solutions platform in Saudi Arabia.
• Neuroscience is experiencing rapid growth in areas like NeuroAI, where artificial intelligence influences brain theories and research.
• Government restrictions on AI use are facing legal scrutiny, with arguments suggesting they could violate First Amendment free speech rights.

NVIDIA introduces ComputeDomains for multi-node NVLink

On November 10, 2025, NVIDIA announced a new Kubernetes tool called ComputeDomains. This tool simplifies how multiple GPUs communicate across different computers using multi-node NVLink. ComputeDomains ensure secure and efficient GPU-to-GPU memory operations. It works by automatically setting up and managing IMEX domains for each workload, making the system more flexible and secure. This technology has been tested on NVIDIA DGX GB200 systems and will support future multi-node NVLink architectures.

NVIDIA Grove simplifies complex AI on Kubernetes

NVIDIA introduced Grove, a new component within its Dynamo platform, on November 10, 2025. Grove helps manage and scale complex AI inference systems on Kubernetes, which often involve many different parts like prefill and decode. It allows users to define their entire inference system as a single Custom Resource, simplifying deployment. Grove handles scheduling, placement, and autoscaling across many GPUs, even tens of thousands. This tool makes it easier to run advanced AI models efficiently in large data centers.

NVIDIA Dynamo and Grove simplify AI inference at scale

NVIDIA Dynamo now helps simplify large-scale AI inference, especially with disaggregated serving, across GPU clusters. This method splits AI tasks like prefill and decode to run on different, optimized GPUs for better performance. Dynamo is now available through major cloud providers' Kubernetes services, supporting NVIDIA Blackwell systems and H200 GPUs. A new component called NVIDIA Grove, part of Dynamo, further simplifies managing these complex AI workloads on Kubernetes. Grove lets users easily define their inference needs, and it automatically handles the complex coordination and scaling of components.

Mark Mobius warns AI stocks face 40 percent crash

Veteran investor Mark Mobius issued a strong warning about the artificial intelligence stock market. He predicts a possible 40 percent crash, calling the current situation a classic bubble. Mobius believes that excitement around AI has pushed stock values too high, making them unsustainable. He advises investors to be careful and focus on a company's real value and manage risks. Mobius also sees good investment chances in emerging markets once the AI bubble eventually bursts.

Smart investing in the AI stock market bubble

The article suggests a thoughtful approach to the current AI stock market excitement. It advises investors to remember that AI outcomes are not all or nothing, and to consider the different stages of AI development. Diversifying investments across various AI sectors like hardware and software is crucial to reduce risk. Investors should focus on a company's real value and long-term potential, not just the hype. Understanding the risks, such as regulations and competition, and thinking long-term will help navigate this market effectively.

Salesforce buys Spindle AI for Agentforce analytics

Salesforce recently signed an agreement to acquire Spindle AI, an agentic analytics platform. This acquisition aims to boost the analytics capabilities of Salesforce's Agentforce. Spindle AI uses artificial intelligence, machine learning, and data modeling to help businesses predict outcomes and make better decisions. Its technology can simulate the effects of complex business choices, such as changes in pricing or sales strategies. The acquisition is expected to finalize in Salesforce's fourth fiscal quarter of 2026.

Neuroscience sees rapid growth in AI and brain studies

On November 10, 2025, a report highlighted several fast-growing areas in neuroscience. Experts pointed to computational neuroscience, systems neuroscience, neuroimmunology, and neuroimaging as key fields. The intersection of neuroscience and artificial intelligence, or NeuroAI, is also expanding quickly, with AI influencing brain theories and vice versa. Researchers are also making progress in understanding supporting brain cells like microglia and how fluids circulate in the central nervous system. New technologies for measuring the whole brain and using machine learning for behavior profiling are driving much of this growth.

Maryland needs clean energy to power AI and quantum leadership

Maryland aims to become a national leader in artificial intelligence and quantum technology. However, to achieve this, the state must also lead in clean energy and energy storage. AI data centers and quantum computers require vast amounts of electricity, and their energy needs will only grow. A strong, sustainable energy system is vital for economic resilience and attracting future industries. Maryland has already made good progress in clean energy, with significant solar power and offshore wind projects. The state needs to further invest in research and development for advanced energy storage and grid modernization to power its tech ambitions.

Workey partners with Falak for AI business platform

On November 10, 2025, Workey announced it closed an investment round with Falak Investment Hub. This partnership will help Workey launch its new AI-powered business solutions platform. The platform currently offers smart digital bookings for over 200 locations in Saudi Arabia, including offices, meeting rooms, and event spaces. It also provides support services for startups and foreign investors. This strategic investment is expected to speed up Workey's growth and strengthen its role in Saudi Arabia's business market.

Scott Galloway warns of market crash if OpenAI fails

On November 10, 2025, tech analyst Scott Galloway warned of a major market crash if problems at OpenAI continue. He believes the high values of top tech companies like Nvidia and Microsoft depend on generative AI's success. Galloway pointed to concerns about OpenAI's finances, including its CFO stating they are "burning through cash," and leadership issues involving CEO Sam Altman. If companies report that AI trials are not delivering expected returns, investor confidence could drop sharply. He stressed that if the top 10 companies, which make up 40 percent of the S&P 500, see their values cut, the impact on the global markets would be severe.

Understanding the AI bubble and its risks

On November 10, 2025, an article discussed a different way to view the AI market bubble. It explains that a bubble happens when a bet becomes too big, leading to more supply than demand. A key challenge is the difference between fast AI software development and the slow pace of building data centers. It is hard to predict future AI demand and supply due to complex supply chains and potential tech breakthroughs. Huge investments, like Microsoft's reported $100 billion for the Stargate supercomputer, face risks from uncertain demand and infrastructure issues such as power grid capacity. These slow-moving infrastructure elements can create expensive bottlenecks for rapid AI growth.

AI restrictions may violate free speech rights

On November 10, 2025, an article explored whether government rules on artificial intelligence are constitutional. It argues that many restrictions on AI use might violate the First Amendment because they control speech based on its content. The legality depends on whether these rules are about the "time, place, or manner" of speech or about the "viewpoint or subject-matter." Content-based speech rules are usually unconstitutional, with few exceptions. While some AI restrictions in schools might be allowed if they relate to teaching, broader government rules on AI use could face legal challenges.

Kaltura buys eSelf AI for 27 million dollars

On November 10, 2025, Kaltura acquired eSelf.ai for $27 million. eSelf.ai was co-founded by Alan Bekker, who previously created Snap's AI, and has a team of about 15 AI experts in natural language processing and computer vision. Kaltura, a Nasdaq-listed company with $180 million in revenue, offers cloud-based video solutions to over 800 enterprise clients. Kaltura plans to integrate eSelf.ai's virtual agent technology into its video products, allowing AI agents to listen, speak, and understand user screens in real time. This move helps Kaltura become a video-based customer and employee experience provider, offering human-like conversational AI avatars for various business uses.

Sources

• Enabling Multi-Node NVLink on Kubernetes for GB200 and Beyond
• Streamline Complex AI Inference on Kubernetes with NVIDIA Grove
• Think SMART: New NVIDIA Dynamo Integrations Simplify AI Inference at Data Center Scale
• Mark Mobius warns of 40% crash in AI stocks, calls it a classic bubble in the making
• A better way of thinking about the AI bubble
• Salesforce acquires Spindle AI to bolster analytics for Agentforce
• What are the fastest-growing areas in neuroscience?
• If Maryland wants to lead in AI and quantum, it must lead in clean energy, too
• Workey Closes Investment Round with Falak to Launch AI-Powered Business Solutions Platform
• Scott Galloway warns of ‘nowhere to hide’ in markets if the OpenAI story unravels
• A better way of thinking about the AI bubble 
• Are Government Restrictions on AI Constitutional?
• Kaltura acquires eSelf, founded by creator of Snap’s AI, in $27M deal