Praruh is targeting a January launch for its NeuralOps AI platform, which manages IT infrastructure including networks and data-centre services. The company expects the platform to boost profit after tax by 5% to 10% in 2027-28 and aims to build a Make-in-India alternative to imported software. Sales will begin as outright purchases or SaaS subscriptions.
Saksoft is accepting short-term revenue pain to pursue an AI-first strategy, offering free proof-of-concepts and reduced team sizes to customers. Q2-FY27 operating income came in at Rs 248.612 crore, down 3.8% year on year, while H1-FY27 operating income fell 2.0% to Rs 497.231 crore. Management expects H2-FY27 performance to improve over the first half.
Workmates has brought six AI accelerators into live production across IT and business functions. The IT-side products include IntelliOps, CloudTrustMate, and Enterprise AI Edge, while VoxoRA, VerID, and VoyaGE serve business applications. H1 FY27 recurring revenue reached Rs 82.97 crore, up 43% year on year and now representing 87% of total revenue.
Chinese AI lab Moonshot, maker of the Kimi model, has reportedly explored European data centres to access Nvidia Blackwell GPUs that are banned from direct import to China. Talks involved renting clusters of 8,000 or more GPUs hosted in Europe, with Moonshot described as very aggressive in seeking compute power. The U.S. House passed the Remote Access Security Act to close such cloud-based loopholes, and the White House has accused Moonshot of accessing AI capabilities through Thailand.
Omnissa expanded its Sovereign Solution to six new markets, including Austria, Germany, Spain, the Czech Republic, Slovakia, Sweden, and South Korea. CEO Amit Singh announced the expansion in Amsterdam. The platform helps organizations manage large language models and AI agents within their own jurisdiction, keeping data and operations local. Windows Server management is now available on OSS, and seven more markets are expected soon.
Laza Medical raised $160 million from GE HealthCare and other investors to develop AI-powered robotics for catheterization labs. The technology uses AI to analyze images and data during procedures, helping doctors improve decision-making and patient outcomes. GE HealthCare said the investment will support faster development and commercialization.
Qonic launched Qonic AI, embedding agents directly into its cloud-based BIM modeling platform. Users describe tasks in plain language through the Ask Qonic chatbot, and the AI finds, checks, and updates elements. The tool can complete missing properties, build reports, edit geometry, and model buildings from prompts. Users can choose models like Sonnet, Opus, or Astra, and an MCP server lets them connect external assistants such as Claude or Codex. Purchased tokens never expire and start from €50 for the free plan.
Valour, a subsidiary of DeFi Technologies, launched eight new digital asset exchange traded products on Sweden's Spotlight Stock Market. The products cover assets including Venice, Canton Network, Raydium, Morpho, Maple Finance, Pump.fun, Dogwifhat, and Bonk. Each ETP carries a management fee of 1.9% and allows investors to track digital assets through traditional brokerage channels.
Los Angeles County authorities are warning residents about AI voice-cloning phone scams, where criminals imitate family voices and create convincing messages or videos to request money. The Sheriff's Department recommends pausing to verify unexpected requests and calling the person directly using a known number rather than sending money based on a familiar voice or appearance.
New research argues that distributed learning should become a core method for post-training large language models. Current approaches like RLHF and DPO assume preference data must be centralized, but the authors call this assumption unsustainable. Distributed methods could better support privacy, representational equity, and scalability, though technical challenges remain.
AI agent autonomy is advancing but remains limited in real-world use, according to a feature report. Some IT leaders allow agents to handle low-risk tasks like processing refunds or support tickets without human approval, but most say full autonomy is not yet safe for mission-critical work. Organizations are using policies, API gateways, and credential controls to keep agent behavior within acceptable bounds.
Professors John Cairney and Rick Burton argue that AI's growing accessibility may shorten the competitive advantage it offers sports organizations. Teams are spending heavily on AI for player performance, ticket pricing, and fan engagement, but once one team adopts a tool, others can follow quickly. They say human skills like intuition, leadership, and decision-making under uncertainty will become even more valuable as AI takes on more work.
Key Takeaways
- Praruh plans a January launch for its NeuralOps AI platform, expecting a 5% to 10% boost in profit after tax for 2027-28.
- Saksoft is accepting near-term revenue loss to pursue an AI-first strategy, with H1-FY27 operating income down 2.0% year on year.
- Workmates has six AI accelerators live in production, with H1 FY27 recurring revenue up 43% to Rs 82.97 crore.
- Moonshot AI explored European data centres to access Nvidia Blackwell GPUs banned from direct import to China, seeking clusters of 8,000 or more GPUs.
- Omnissa expanded its Sovereign Solution to six new markets under CEO Amit Singh, keeping AI resources within local jurisdictions.
- Laza Medical raised $160 million from GE HealthCare to develop AI-powered robotics for catheterization labs.
- Qonic AI embeds agents into its BIM modeling platform and supports external assistants like Claude and Codex via an MCP server.
- Valour launched eight digital asset ETPs in Sweden, each with a 1.9% management fee.
- Los Angeles County warns about AI voice-cloning phone scams and urges residents to verify unexpected requests directly.
- Research advocates distributed learning for post-training large language models to support privacy and scalability.
Praruh plans January launch for NeuralOps AI platform
Praruh plans to launch its AI-enabled NeuralOps platform in January and begin outright and SaaS sales in January or February. The platform manages customer IT infrastructure including networks and data-centre services. Praruh expects NeuralOps to increase profit after tax by at least 5% to 10% in 2027-28. The company aims to build a Make-in-India alternative to imported software.
Saksoft accepts short-term revenue loss to push AI
Saksoft is offering free proof-of-concepts and reduced team sizes to customers as part of its AI-first strategy, accepting near-term revenue cannibalisation. Q2-FY27 operating income was Rs 248.612 crore, flat sequentially but down 3.8% year on year. H1-FY27 operating income was Rs 497.231 crore, down 2.0% year on year, with PAT at Rs 61.410 crore. Management expects H2-FY27 performance to improve over H1-FY27.
Workmates launches six AI products in live production
Workmates reports six AI accelerators are now live in production across IT and business functions. The IT products include IntelliOps, CloudTrustMate and Enterprise AI Edge, while VoxoRA, VerID and VoyaGE serve business applications. H1 FY27 recurring revenue reached Rs 82.97 crore, up 43% year on year and representing 87% of total revenue. Production AI engagements reportedly take four to 12 weeks.
Chinese AI lab Moonshot eyes European data centres for Nvidia chips
Chinese AI lab Moonshot, maker of the Kimi model, has reportedly looked at European data centres to access Nvidia Blackwell GPUs banned from direct import to China. Talks involved renting clusters of 8,000 or more Blackwell GPUs hosted in Europe. Moonshot was primarily interested in inferencing capacity and also asked about training clusters. The U.S. House passed the Remote Access Security Act to ban such cloud-based access loopholes.
Moonshot AI pursues Nvidia Blackwell access through Europe
Moonshot AI, a Chinese AI lab founded in 2023, has discussed leasing Nvidia Blackwell GPU capacity at European facilities through neoclouds and data centre operators. Four sources familiar with the matter said talks involved clusters of 8,000 or more GPUs. Moonshot was described as very aggressive in seeking compute power. The White House accused Moonshot of accessing AI capabilities via Thailand, and U.S. export controls restrict Nvidia's most advanced chips from China.
Distributed learning now essential for training large AI models
A new vision paper says distributed learning should become a core method for post-training large language models. Current approaches like RLHF and DPO assume preference data must be centralized, but the paper calls this assumption unjustified and unsustainable. The authors argue that distributed methods support privacy, representational equity, and scalability. They also outline technical challenges and research directions to make distributed post-training practical.
AI agent autonomy is still a work in progress
A new feature report says AI agent autonomy is advancing but remains limited in real-world use. Some IT leaders allow agents to handle low-risk tasks like processing refunds or support tickets without human approval. However, most leaders say full autonomy is not yet safe for mission-critical or sensitive work. They are using policies, API gateways, and credential controls to keep agent behavior within acceptable limits.
Valour launches eight new digital asset ETPs in Sweden
DeFi Technologies' subsidiary Valour launched eight new digital asset exchange traded products on the Spotlight Stock Market in Sweden. The products began trading on November 5, 2026, and cover assets like Venice, Canton Network, Raydium, Morpho, Maple Finance, Pump.fun, Dogwifhat, and Bonk. Each ETP has a management fee of 1.9 percent and lets investors track the underlying digital asset through traditional brokerage channels. The launch expands European access to a range of blockchain and AI-related tokens.
Los Angeles County warns about AI voice-cloning scams
Los Angeles County authorities are warning residents about phone scams that use AI voice cloning. Scammers imitate the voices of family members and create convincing messages or videos to request money. The Sheriff's Department advises people to pause and verify unexpected requests. It recommends calling the person directly using a known number and never sending money just because a voice or appearance looks familiar.
Laza Medical raises $160M for AI-powered cath lab robotics
Laza Medical, a California-based medtech startup, raised $160 million from GE HealthCare and other investors. The company is developing AI-powered robotics for the cath lab to improve catheterization procedures. The technology uses AI to analyze images and data, helping doctors make better decisions and improve patient outcomes. GE HealthCare said the investment supports faster development and commercialization of the technology.
Omnissa expands sovereign AI security to six new markets
Omnissa expanded its Sovereign Solution (OSS) to six new markets in Europe and Asia, including Austria, Germany, Spain, the Czech Republic, Slovakia, Sweden, and South Korea. OSS is already available in India and seven more markets are expected soon. The platform helps organizations manage AI resources like large language models and AI agents within their own jurisdiction. It focuses on keeping customer data and service operations local. Amit Singh, CEO of Omnissa, announced the expansion in Amsterdam. Windows Server management is now available on OSS.
Qonic AI brings agents directly into BIM modeling
Qonic launched Qonic AI, a set of agents built into its cloud-based BIM modeling platform. Users can describe tasks in plain language through the Ask Qonic chatbot, and the AI finds, checks, and updates elements. The tool can complete missing properties, build reports, edit geometry, and model buildings from prompts. Users can choose AI models like Sonnet, Opus, or Astra. Qonic also introduced an MCP server that lets users connect their own AI assistant, such as Claude or Codex, to the platform. Purchased tokens never expire and start from €50 for the free plan.
AI may make sports advantages shorter than teams think
Professors John Cairney and Rick Burton argue that AI is becoming more accessible, which may shorten the competitive advantage it offers sports organizations. Teams are spending heavily on AI for player performance, ticket pricing, and fan engagement. However, once one team adopts a tool, others can follow quickly. The authors say human skills like intuition, leadership, and decision-making under uncertainty will become even more valuable. They believe the real question is which human capabilities grow in importance as AI takes on more work. Their column appeared on Yahoo Sports.
Sources
- Praruh plans January NeuralOps launch with outright, SaaS sales
- Saksoft accepts near-term revenue cannibalisation in AI push
- Workmates has six AI accelerators live in production across IT and business
- Chinese AI lab Moonshot reportedly looked at European data centers to access Nvidia chips, report claims — Kimi-maker primarily interested in inferencing capacity, but also asked about training clusters
- Chinese AI lab Moonshot turned to Europe for Nvidia Blackwell access amid U.S. China curbs, sources tell CNBC
- Vision Paper: Why Distributed LLM Post-Training and Decentralized Alignment Are Now Essential
- AI agent autonomy remains an elusive work in progress
- DeFi Technologies' Subsidiary Valour Launches Eight New Digital Asset ETPs on Spotlight Stock Market, Expanding Access to AI, Institutional Blockchain, DeFi and Community Tokens
- Los Angeles County warns residents about AI voice-cloning scams
- Startup bringing AI-powered robotics to the cath lab raises $160M from GE HealthCare, other investors
- Omnissa expands sovereign AI security push as enterprises confront agentic threats
- Qonic AI launches inside BIM platform
- In an AI World, Sport’s Competitive Advantage May Be Its Oldest One
Comments
Please log in to post a comment.