AI-tonomy Summit at Plug & Play in Sunnyvale. Great Panel with A listers from Silicon Valley. linkedin.com/posts/cphelan_a…

HaaS is the new SaaS. 🚀 Had a blast speaking at Ai Tonomy summit to a room full of brilliant AI researchers, founders and investors! When we talk about Physical AI, building the tech is only half the battle, the other half is nailing the business model and GTM strategy. Here are the 3 hardware lessons: Upfit, Don't Build from Scratch: SaaS succeeds by owning a specific software layer, not building the OS. In Physical AI, the winning play is upfitting, which means retrofitting existing, industry-standard machinery with an autonomous "brain" rather than trying to manufacture heavy hardware from the ground up. The Physical Data Flywheel is Your Real Moat: You can’t train enterprise-ready robotics models just by scraping videos off the internet. Your true moat is closed-loop operational data. Every hour your hardware spends handling messy, real-world edge cases in the field creates a proprietary barrier to entry no one can copy. Sell Outcomes, Not Boxes: Traditional hardware relies on low-margin, transactional CapEx sales. True HaaS shifts the model to recurring subscriptions or hourly productivity fees. This turns a risky hardware purchase for the customer into a software-like recurring revenue stream for you. Physical AI is completely rewriting the startup playbook. Thanks to everyone who came out! What’s your take, will HaaS fully replace traditional hardware sales in the enterprise? Many thanks to Brett Liu, Engineer/CFA , Wickey Wang, CISA, Six Sigma Green Belt and team for the invite! Great to share stage with @ashishkakran , Ankur Saxena , @TheWilsonLin, @PaulFangBayArea #AI #Robotics #PhysicalAI #VentureCapital #Startups #HaaS

Last week I royally fumbled an opportunity to pitch Bloomberry on the main stage at AI-tonomy in front of 500 founders, operators, and investors. Because I literally wasn’t in the room when they called my name. I went to my car to rehearse the pitch again while someone else listened for the finalist announcements. I got a text: “better luck next time.” So I drove off. 30 minutes later, the event director texted me saying she was calling my name. I had actually advanced. The last few days have been a weird mix of frustration, embarrassment, and reflection. But there’s a lesson in it I needed: Never remove yourself from the room before the outcome is final. Especially if you know you belong there. Super painful experience. Valuable one. And now back to building.

One topic came up in every conversation at the AI-tonomy Summit last week: cost. Not accuracy, not latency. Those problems are mostly solved. The question everyone's asking now is how to make these models cheaper to run. That’s what we’re trying to accomplish with Pioneer. More on that soon.

Every person at AI-tonomy Summit asked us the same question: LLM costs are too high, are there any alternatives we can use to reduce them? Last week, we showed how Pioneer can use small models to improve model accuracy on tasks and reduce costs across workflows without sacrificing performance. More on that soon. Here are a few highlights from the event.

Getting an agent to execute is solved. Getting one you can trust with real work is not. That was the conversation all day at the AI-tonomy Summit, where our founder @panda_liyin showed AdaL live to a room of builders, researchers, and founders at @PlugandPlayTC Thanks for having us!

12 Startup Events: West Coast San Jose, CA Tech Fest 2026 – Silicon Valley (AI Networking), June 7th, 2026 luma.com/sftechfest26 Cupertino, CA Apple WWDC26 (Worldwide Developers Conference), June 8–12th, 2026 developer.apple.com/wwdc26 San Francisco / Bay Area, CA The AI Rabbit Hole 2026 – AI Bubble Tea Party, June 5th, 2026 luma.com/AIRabbitHole2026 AI-tonomy Summit: Models & Agents, June 5th, 2026 luma.com/cvonarf6 CommunityKit at WWDC26 – Keynote Watch Party, Meetups & Indie Fair, June 8–12th, 2026 luma.com/848w7ld2 Seattle, WA AI Con USA 2026 – AI & Machine Learning Conference, June 7–12th, 2026 aiconusa.techwell.com AI Tinkerers Seattle: Dev Tools Track, June 8th, 2026 seattle.aitinkerers.org Deep Tech Week Seattle – DeepTech Founder Show & Tell, June 9th, 2026 deep-tech-week.com/seattle-2… Seattle DTW Investor Summit ft. Boost VC, UNION Labs & Roadster Capital, June 11th, 2026 deep-tech-week.com/seattle-2… Denver, CO CVPR 2026 – Computer Vision & Pattern Recognition Conference, June 5–7th, 2026 cvpr.thecvf.com Las Vegas, NV Info-Tech LIVE 2026 – Agentic IT: From Hype to Value, June 9–11th, 2026 infotech.com/events/las-vega… Berkeley, CA TechCrunch Sessions: AI 2026, June 5th, 2026 techcrunch.com/events/tc-ses…

Just heard an NVIDIA researcher talk about Nemotron / efficient agentic LMs at AI Tonomy Summit. A few quick takeaways: 1/ Agent models may not be scaled by AR only. AR = strong left-to-right language prior Diffusion = lookahead planning / parallel token decoding Self-speculation = faster draft verify loop 2/ This is why Nemotron-Labs-Diffusion is interesting. It tries to unify AR / diffusion / self-speculation in one model, instead of treating them as competing paths. 3/ For agents, this matters a lot. Agent work is usually long-loop: plan / draft / verify / revise / recover Sequential decoding gets expensive very fast. 4/ From a research angle, the next bottleneck may be less about “bigger model” and more about: decoding efficiency / throughput / inference cost / training objective / long-horizon reliability 5/ My takeaway: Efficient agentic AI needs models that are smart, fast, reliable, and economically scalable.

Inference isn't the end of deployment, it's where continual learning begins. Our co-founder, @ash_csx, will be at AI-tonomy Summit demoing Pioneer and joining a panel alongside companies like @cerebras, @FeatherlessAI, and @DeepInfra, discussing how to best approach model inference. Make sure to stop by both of these talks to hear more about what we've been working on.

We’re excited to announce the panel “Unique Funding and GTM for Agents” at AI-tonomy Summit 2026: Models & Agents. As AI agents move from demos to real products, founders face a new set of questions: how to fund, position, distribute, and monetize agentic applications in a market that is still rapidly evolving. Joining the conversation: 🟠 Grace Gong — Founder, Smart Venture Media 🟠 Ankur Saxena — Investment Director, @TDKVentures 🟠 Ashish Kakran — Managing Partner, Seligman Ventures 🟠 Wilson Lin — AI Investment Lead, @DCGco 🟠 Paul Fang — President, @BFC_Global Looking forward to a practical discussion on the funding paths, go-to-market strategies, and ecosystem channels that can help agent startups move from early traction to scale.

Would love to invite founders, researchers, engineers, and investors in AI to join us at AI-tonomy Summit: Models & Agents on June 5 at Plug and Play in Silicon Valley. The morning research forum will feature top AI researchers discussing AI agents, AI4Science, verified reasoning, continual learning, and physical / multimodal intelligence. We’ll also have around 100 VC firms, investors, and strategic partners joining. Sharing complimentary passes for founders and investors here — come say hi if you’ll be there :) Founder pass: luma.com/cvonarf6?coupon=AIT… Investor pass: luma.com/cvonarf6?coupon=AIT…

We’re excited to announce the panel “Enterprise Agent Adoption: Agent Runtime Environment” at AI-tonomy Summit 2026: Models & Agents. As enterprises move from AI pilots to production-grade agentic systems, the runtime environment becomes critical — enabling agents to execute safely, connect with enterprise tools, manage workflows, and operate reliably at scale. Joining the conversation: 🟠 Andi Partovi — Co-founder & CTO, @veris_ai 🟠 Li Yin — CEO, Sylph.ai 🟠 Jean Malha — Generative AI Specialist, @AWS 🟠 Deepak Prasad — Vice President, @QualcommVenture 🟠 Jonathan Chan — Ex-CISO 🟠 Nnamdi Iregbulem — Partner, @lightspeedvp 🟠 Leo Tchourakovi — Member of Technical Staff, @FactoryAI Looking forward to a practical discussion on the infrastructure, security, and operational layers needed to deploy enterprise AI agents in the real world.

We’re excited to announce the panel “Modern AI Infrastructure: Physical AI & World Models” at AI-tonomy Summit 2026: Models & Agents. As AI expands beyond digital environments, physical AI and world models are becoming key foundations for systems that can perceive, reason, simulate, and act in the real world. Joining the conversation: 🟠 Enoch Zhu — CEO, @Orbifold_AI 🟠 Fan-Yun Sun — CEO, Moonlake AI 🟠 Gayathri Radhakrishnan — Partner, Hitachi Ventures 🟠 Misha Obolonskyi — Principal, @CorticalVC 🟠 Viren Tellis — Co-founder & CEO, @Uthana_Inc 🟠 Christoph Kohstall — Investor, @1x_tech Looking forward to a thoughtful discussion on the infrastructure, models, and investment landscape powering the next wave of physical AI.

We’re excited to announce the panel “Modern AI Infrastructure: Data Infrastructure for Agents” at AI-tonomy Summit 2026: Models & Agents. As AI agents move from demos to production, data infrastructure becomes a critical foundation — enabling agents to retrieve context, reason over complex information, interact with enterprise systems, and deliver reliable outcomes at scale. Joining the conversation: 🟠 Etienne Dilocker — CTO & Co-founder, @weaviate_io 🟠 Deepak Mishra — Co-founder, @GraphonAI 🟠 Christopher Di Gioia — Head of AI/Data, @PGIM 🟠 Madison Faulkner — Partner, @NEA 🟠 Andy Triedman — Partner, @Theoryvc 🟠 Kyle McNulty — Investor, iqt Looking forward to a deep discussion on the data, infrastructure, and investment layers powering the next generation of agentic systems.

We’re excited to announce the panel “Enterprise Agent Adoption & Agent Security” at AI-tonomy Summit 2026: Models & Agents. As enterprises move from AI experimentation to real agentic deployment, security becomes a core requirement — from identity and access control to workflow governance, risk management, and trust in autonomous systems. Joining the conversation: 🟠 Ankur Shah — Co-founder & CEO, @straikerai 🟠 Snehal Antani — Co-founder & CEO, @Horizon3ai 🟠 Dušan Vuksanovic — CEO, @Swisscom 🟠 Ken Elefant — Partner, @SorensonCap 🟠 Morgan Kyauk — Managing Director, @nightdragon 🟠 Dan Hatfield — Managing Director, @Stifel Looking forward to a timely discussion on what it takes to bring AI agents safely, reliably, and responsibly into enterprise environments.

We’re excited to announce another incredible panel at AI-tonomy Summit 2026: Models & Agents. As AI systems become more capable, the infrastructure behind them is becoming just as important as the models themselves — from inference and compute efficiency to deployment, scaling, and the economics of serving next-generation AI applications. Joining the conversation: 🟠 Nikola Borisov — Co-founder & CEO, @DeepInfra 🟠 Eugene Cheah — CEO, @FeatherlessAI 🟠 Ash Lewis — CEO, @fastinoAI 🟠 Qi Jin — EVP, @cerebras 🟠 Denise Teng — Partner, @GradientVC 🟠 Will Horyn — Principal, @GeodesicCap Looking forward to a thoughtful discussion on the infrastructure, investment, and deployment layers powering the next wave of AI systems.

AI-tonomy Summit 2026: Where Top Silicon Valley AI Researchers Define the Future of Models & Agents June 5, 2026 · Plug and Play Tech Center · Silicon Valley Artificial intelligence is shifting from static models to autonomous systems. AI-tonomy Summit: Models & Agents is a one-day gathering focused on the next generation of AI infrastructure, agents, and real-world deployment — bringing together researchers, engineers, founders, investors, and operators building at the frontier of AI. Hosted in Silicon Valley, the summit will convene 500 attendees across the AI ecosystem, with discussions centered around how models evolve into systems that can reason, act, collaborate, and operate in production. Agenda 10:00 AM – 12:00 PM AI Researcher Forum A technical session featuring researchers from leading AI labs and universities discussing frontier directions in autonomous AI systems. Topics include: Autonomous AI Scientists Agentic RL & Reasoning Self-Evolving Agents Multimodal & Physical Intelligence 12:00 PM – 1:00 PM Networking Lunch An open networking session connecting researchers, founders, engineers, and investors across the AI ecosystem. 1:00 PM – 5:00 PM Panels · Demos · Founder Sessions The afternoon program features high-density sessions across both Model and Agent tracks, focused on real-world AI deployment and company building. Topics include: Model Compute Optimization World Models Runtime Infrastructure Agent Agent Security Agent Data & Memory Production Systems Alongside: Live Product Demos Founder Showcases Lightning Pitches 5:00 PM – 7:00 PM Cocktail Reception An evening reception designed for deeper conversations and ecosystem connections across the AI community. Featured Speakers @lecong Associate Professor · Stanford University @BretWaters Lecturer · Stanford University @maddiehfaulkner Partner · New Enterprise Associates (NEA) Andy Triedman Partner · Theory Ventures Ankur Shah CEO & Co-founder · Straiker @bobvanluijt CEO & Co-Founder · Weaviate @g_tweets Partner · Hitachi Ventures @kelefant Partner · Sorenson Capital Qi Jin Executive Vice President, Software · Cerebras Pengtao Xie Associate Professor · UC San Diego @snehalantani Co-Founder & CEO · Horizon3.ai More Speakers @andiPartovi CTO · Veris AI Asit Sohoo AI Product Lead · Iron Mountain @ashishkakran Managing Partner · Seligman Ventures @BrettLIu7 Investment Lead · Sixty Degree Capital Christopher Di Gioia Vice President · PGIM @CKohstall AI Scientist · 1X Dan Hatfield Managing Director · Stifel Bank @danmoskowitz Partner · Third Point Ventures Clark Zhang Co-Founder & CTO· Graphon AI Deepak Prasad Vice President · Qualcomm Ventures @DeniseTWong Partner · Gradient Ventures Dušan Vuksanovic Executive · Swisscom Ventures @Orbifold_AI Co-Founder & CEO · Orbifold AI @etiennedi CTO & Co-Founder · Weaviate Erran Li Head of AI • AWS Eugene Cheah CEO & Co-Founder · Featherless AI Fan-Yun Sun Co-Founder · Moonlake AI Grace Gong Founder · Smart Venture Media Holly Hon SVP, AI & Operation Jea Malta Senior AI Architect · AWS Jonathan Chan Former CIO · Episource Li Yin CEO · Sylph AI Morgan Kyauk Managing Director · NightDragon Nikola Borisov Co-Founder & CEO · deepinfra Nnamdi Iregbulem Partner · Lightspeed Venture Pramanya Guda Developer & Community Lead · Founders Bay Sophie Ren Founder · ChatSlide Wilson Lin AI Investment Lead · DCG Viren Tellis Co-Founder & CEO · uthana Yaron Schneider CTO & Co-Founder · Diagrid How To Register Comment “AI-tonomy” below to get the exclusive registration links. Looking forward to seeing you there on June 5. — Dr. Paul Fang Founder, Bay Area Founders Club

We’re excited to announce the panelists for “Physical & Multimodal Intelligence” at AI-tonomy Summit 2026: Models & Agents. As AI moves beyond text and software-only environments, physical and multimodal intelligence will be critical for agents to perceive, reason, and act in the real world. Joining the conversation: 🟠 Chongkai Gao — AI Researcher, Stanford University @Stanford 🟠 Haotian Cui — Senior AI Scientist, PhD, Xaira 🟠 Tiange Xiang — AI Researcher, Massachusetts Institute of Technology @MIT Looking forward to an insightful discussion on the future of multimodal systems, embodied intelligence, and AI agents that can understand and interact with the physical world.

We’re excited to announce the panelists for “Continual Learning & Self-Evolving Agents” at AI-tonomy Summit 2026: Models & Agents. As AI agents move into dynamic real-world environments, continual learning becomes essential — enabling agents to adapt, improve, and evolve from ongoing interactions and feedback. Joining the conversation: 🟠 Hanchen Li — AI Researcher, UC Berkeley @UCBerkeley 🟠 Pengtao Xie — Associate Professor, PhD, UC San Diego @UCSanDiego 🟠 Xiaolong Jin — AI Researcher, Microsoft @Microsoft 🟠 Zhe Zhao — AI Researcher, Stanford University @Stanford 🟠 Fan Nie — Research Assistant, Stanford University Looking forward to a thoughtful discussion on how agents can learn continuously, evolve safely, and become more capable over time.