Obot AI raises $35M for MCP adoption in enterprises

Enterprises are adopting AI agents this year, and MCP is the key enabling technology. At Obot AI, we’re building the foundation to make that possible, and we’ve raised $35 million in Seed funding, co-led by Mayfield and Nexus Venture Partners. We’re scaling up our engineering and product teams to advance development of our Obot MCP Gateway and Nanobot Agent Framework. We are showing that enterprises don’t need to choose between innovation and safety. They can have both, and they can deploy MCP at scale today! Special thanks to Kyt Dotson at SiliconANGLE for breaking the story. Read more in Kyt’s article, linked in the comments below.

🚀 Big news in DeepTech! Munich-based startup Yasp has raised €4.2 million to accelerate AI Model Optimisation with its breakthrough Agentic AI Compiler. AI development is facing a critical bottleneck: the slow, resource-heavy process of training and deploying models. Yasp is solving this with compiler-level innovation that speeds up training and inference while reducing costs and avoiding vendor lock-in. Backed by Capnamic, Start-up BW Innovation Fonds, MBG BW, and experienced angels, Yasp is preparing for a Q4 2025 launch. With global AI investments projected to reach $1.8 trillion by 2030, scalable optimisation tools are becoming mission-critical for enterprises worldwide. This funding round signals growing investor interest in infrastructure-level AI solutions. By bridging software innovation with hardware efficiency, Yasp is positioning itself at the core of the next wave of AI transformation. 👉 What’s your take—will compiler-driven AI Model Optimisation change the way we build and deploy AI at scale? www.tycoonstory.com/ #aimodeloptimisation #deeptech #fundingnews #artificialintelligence #startupecosystem #machinelearning #innovation #techfunding yasp

AI agents aren’t just another layer of automation—they represent a new paradigm in software engineering. Instead of simple Q&A, they plan, reason, and act across multi-step workflows. This is a fundamental shift: 🔹 From tools → to ecosystems of collaborating agents 🔹 From tasks → to goals 🔹 From assistance → to autonomous orchestration For startups, this opens a rare window: the chance to build defensible products powered by proprietary data, create novel user experiences, and scale responsibly with built-in governance. The future of AI isn’t just large models—it’s agentic systems that reshape productivity and decision-making. 🚀 #AI #Innovation #AgenticAI #Startups

💡 From prototype to product — ventureLAB is where hardtech and AI scale. Through CIT’s DC & TAP programs, Ontario SMEs can access ventureLAB’s AI Compute Lab, commercialization hub, and advanced prototyping tools to accelerate go-to-market. Whether you're building AI models or validating hardware, ventureLAB is helping companies lead globally from Ontario. 🎥 Watch how ventureLAB is powering the next generation of innovators. #CITInnovation #AI #Hardtech #Commercialization #DCProgram #TAPProgram #OntarioSMEs #DevSites Laura Clark

As AI infrastructure companies climb the stack, startups can’t out-compute or out-scale them. Instead, they can compete where scale doesn’t matter: by building moats through scarcity. Defensibility comes from domains of data that are private, trusted, and hard to replicate. Marc Andrusko and Alastair (Alex) Rampell argue that the edge in AI belongs to those with permission, built on trust and access.

It’s great to see Andreessen Horowitz highlight data silos as the next critical entry point for AI defensibility. At LingoAI, Solid, and MetaLife Social, we go one step further — turning these silos into decentralized, interconnected, user & community-owned data infrastructures built on trust, sovereignty, and transparency. By returning data ownership to individuals and local communities, we enable a new model of inclusive AI and wealth redistribution, especially across the Global South. This is where the real moat, and perhaps the greatest investment opportunity begins #DecentralizedAI #DataSovereignty #AIForHumanity Una Wang Omar Larios Brian J. Esposito Henry Wang Lindsey DeWitt Prat, PhD Gloriana Monko (PhD) Chia Hock Lai Victoria Wymark (苹果) Lucky Uwakwe Jane Lo Roy Ka-Wei Lee

Data dignity becomes central for regenerative economics and the future of human-centric AI. Data ownership mandates digital personhood - built together with AI Reps. How long must the game go on before people start seeing this? Sigh

🔒 The Moat Has Moved 🔒 The real shift in AI isn’t about bigger models or scale. It’s about access to specialized data, the kind that’s still locked behind ownership, rights, and governance barriers. As Andreessen Horowitz wrote in “Fruits of the Walled Garden,” the new moat in AI is built on permissioned, specialized data, not open datasets anyone can scrape. The edge will belong to those who can unlock access with AI-first design, transparency, and trust. The rarest advantage in AI isn’t in the infrastructure anymore. It’s access to specialized data that can be leveraged by specific domains and verticals. Raily #contentEconomy #GenAI

How do you build defensibility when innovation and change move this fast? Bet on trust. As Charlie Munger said, trust is one of the greatest economic forces on earth. Domain experts, operators, and specialists hold this trust with their customers and partners. It’s a moat. It’s a differentiator. Munger’s “seamless web of deserved trust” is built on integrity, consistency, and time — the kind of trust that compounds quietly until it becomes second nature. But it can’t grow in toxic or manipulative environments where fear replaces honesty and values decay. Protecting your ability to trust, and to be trusted, sometimes means walking away. Real trust only forms where integrity holds, even when it costs you. https://lnkd.in/dtjX6gh4 #kmjp