--- url: 'https://qubit.capital/blog/us-series-a-weekly-funding-roundup-week-5-june-2026' title: 'US Series A Weekly Funding Roundup (Jun 22-29, 2026): $657.0M Raised Across 6 Deals' author: name: Mayur Toshniwal url: 'https://qubit.capital/blog/author/mayur' date: '2026-06-29T05:17:27+05:30' modified: '2026-06-29T11:25:39+05:30' type: post summary: 'Six US startups raised $657.0M in Series A funding (Jun 22-29, 2026), led by General Intuition''s $320M. AI took five of six deals. Full breakdown inside.' categories: - Weekly Funding Roundup image: 'https://qubit.capital/wp-content/uploads/2026/06/featured-us-series-a-67957.webp' published: true --- # US Series A Weekly Funding Roundup (Jun 22-29, 2026): $657.0M Raised Across 6 Deals Six US startups raised a combined $657.0M in Series A funding between June 22 and 29, 2026. AI ran the table this week. Five of the six rounds went to companies building AI models, compute infrastructure, or agent governance, and the sixth applies AI to senior care. The round sizes split hard. General Intuition’s $320M was nearly half the week’s total on its own, while the smallest deal, Hera’s senior-care round, closed at $27M. Capital kept moving across stages too. US seed startups raised $65.0M over two deals the same week, and later-stage US companies pulled in $226.0M across three Series B and beyond rounds. Khosla Ventures appeared in four of the six Series A deals, a sign of how concentrated AI conviction has become among the top funds. Weekly Funding Roundup JUN 22-29, 2026 $657M TOTAL RAISED 6DEALS CLOSED MixedSTAGE $110MAVG DEAL SIZE USTOP REGION BY STAGE Series A$577M88% combined seed and Series A$80M12% BY SECTOR General IntuitionAI (world models, physical AI)$320M RunpodAI infrastructure (GPU cloud)$100M Scaled CognitionAI (enterprise agents)$100M Sail ResearchAI infrastructure (inference)$80M RunlayerCybersecurity (agent governance)$30M HeraHealthtech (senior care)$27M Table of Contents - [1. General Intuition Raises $320M For AI Agents Trained On Gameplay](#1-general-intuition-raises-$320m-for-ai-agents-trained-on-gameplay) [Deal Overview](#deal-overview) - [Investor Profile](#investor-profile) - [Company and Leadership](#company-and-leadership) - [Problem and Opportunity](#problem-and-opportunity) - [Product and Technology](#product-and-technology) - [Use of Proceeds and Vision](#use-of-proceeds-and-vision) - [Market Context](#market-context) - [2. Runpod Raises $100M For Its AI Developer Cloud](#2-runpod-raises-$100m-for-its-ai-developer-cloud) [Deal Overview](#deal-overview-1) - [Investor Profile](#investor-profile-1) - [Company and Leadership](#company-and-leadership-1) - [Problem and Opportunity](#problem-and-opportunity-1) - [Product and Technology](#product-and-technology-1) - [Use of Proceeds and Vision](#use-of-proceeds-and-vision-1) - [Market Context](#market-context-1) - [3. Scaled Cognition Raises $100M For Reliable Enterprise AI](#3-scaled-cognition-raises-$100m-for-reliable-enterprise-ai) [Deal Overview](#deal-overview-2) - [Investor Profile](#investor-profile-2) - [Company and Leadership](#company-and-leadership-2) - [Problem and Opportunity](#problem-and-opportunity-2) - [Product and Technology](#product-and-technology-2) - [Use of Proceeds and Vision](#use-of-proceeds-and-vision-2) - [Market Context](#market-context-2) - [4. Sail Research Raises $80M For Long-Horizon Agent Inference](#4-sail-research-raises-$80m-for-long-horizon-agent-inference) [Deal Overview](#deal-overview-3) - [Investor Profile](#investor-profile-3) - [Company and Leadership](#company-and-leadership-3) - [Problem and Opportunity](#problem-and-opportunity-3) - [Product and Technology](#product-and-technology-3) - [Use of Proceeds and Vision](#use-of-proceeds-and-vision-3) - [Market Context](#market-context-3) - [5. Runlayer Raises $30M To Secure Corporate AI Agents](#5-runlayer-raises-$30m-to-secure-corporate-ai-agents) [Deal Overview](#deal-overview-4) - [Investor Profile](#investor-profile-4) - [Company and Leadership](#company-and-leadership-4) - [Problem and Opportunity](#problem-and-opportunity-4) - [Product and Technology](#product-and-technology-4) - [Use of Proceeds and Vision](#use-of-proceeds-and-vision-4) - [Market Context](#market-context-4) - [6. Hera Raises $27M For AI Senior-Care Coordination](#6-hera-raises-$27m-for-ai-senior-care-coordination) [Deal Overview](#deal-overview-5) - [Investor Profile](#investor-profile-5) - [Company and Leadership](#company-and-leadership-5) - [Problem and Opportunity](#problem-and-opportunity-5) - [Product and Technology](#product-and-technology-5) - [Use of Proceeds and Vision](#use-of-proceeds-and-vision-5) - [Market Context](#market-context-5) - [Lessons For Founders](#lessons-for-founders) ## 1. General Intuition Raises $320M For AI Agents Trained On Gameplay ### Deal Overview - Stage: Series A - Sector: AI (world models, physical AI) - Geography: New York, United States - Round size: $320M, lifting total funding to $454M - Valuation: $2.3B post-money ### Investor Profile Khosla Ventures led the round, joined by General Catalyst, Jeff Bezos, Eric Schmidt, and Nico Rosberg. Khosla and General Catalyst also led the $134M seed in October 2025, so this is a doubling-down by backers who have watched the dataset grow. ### Company and Leadership [General Intuition](https://www.generalintuition.com) spun out of gaming-clip platform Medal.tv in October 2025. CEO Pim de Witte still runs Medal. He co-founded the lab with Eloi Alonso, Adam Jelley, and Vincent Micheli, who carry DeepMind and academic world-model experience. OpenAI reportedly offered around $500M for Medal. De Witte turned it down and built the lab on top of Medal’s data instead. ### Problem and Opportunity AI agents read text well but act poorly in physical space. They struggle to anticipate how an environment changes when they move through it. General Intuition is betting that gameplay, where millions of players make split-second decisions, is the closest available proxy for that kind of reasoning. ### Product and Technology The lab trains large action foundation models that pick optimal actions for a given scene, paired with world models that predict how the scene evolves. The data moat comes from Medal.tv: roughly 17M monthly users uploading on the order of 2 billion clips a year, each tagged with exact button presses and timing. De Witte argues rivals can only infer actions from raw video, which loses that signal. ### Use of Proceeds and Vision Most of the $320M goes to compute, with GPUs sourced through a CoreWeave partnership to pre-train the next model. A public API is targeted for the end of summer 2026. The company is a public-benefit corporation and frames itself as an ecosystem enabler, not a self-driving or robotics product company. ### Market Context General Intuition sits in the world-models and physical-AI space, which it pegs at roughly $1.5B in 2026 growing toward $15B by 2032. It competes with NVIDIA Cosmos, Fei-Fei Li’s World Labs, and Google DeepMind’s Genie 3. Its edge is action-labeled gameplay, since most rivals train on video or synthetic physics without real player inputs. ## 2. Runpod Raises $100M For Its AI Developer Cloud ### Deal Overview - Stage: Series A - Sector: AI infrastructure (GPU cloud) - Geography: Newark, New Jersey, United States - Round size: $100M - Valuation: $1B ### Investor Profile Summit Partners led the round. Runpod’s earlier $20M seed in May 2024 was co-led by the venture arms of Dell and Intel, with Nat Friedman and Julien Chaumond participating. The company has publicly rejected buyout offers above $500M to stay independent. ### Company and Leadership [Runpod](https://www.runpod.io) was founded in 2022 by Zhen Lu and Pardeep Singh, two former Comcast engineers who converted their crypto-mining GPU rigs into AI servers. Early traction came from a single Reddit post, and the team bootstrapped to profitability before raising outside money. ### Problem and Opportunity Most developers stitch together several tools to experiment, train, and serve models. Runpod’s thesis is that builders want one place to take an idea from first run to production traffic, without managing clusters. ### Product and Technology The platform spans the full AI lifecycle with self-serve access, per-second pricing, and no commitment minimums. Most developers reach their first running workload in under an hour. Its serverless product has handled more than 20 billion inference requests, with over 90% of deployments succeeding on the first try. Loyalty is high: 85% of developers who deploy come back to build more, and the base has passed one million developers. ### Use of Proceeds and Vision The capital funds platform and developer-experience work, larger engineering and developer-relations teams, and broader access for international users. Annual recurring revenue roughly doubled to about $240M from $120M in January 2026. ### Market Context The round lands during an intense compute crunch. Runpod competes with CoreWeave, Lambda Labs, and Together AI, and differentiates on developer-first self-serve access rather than raw scale. The market is shifting from training toward inference, where its serverless model fits well. ## 3. Scaled Cognition Raises $100M For Reliable Enterprise AI ### Deal Overview - Stage: Series A - Sector: AI (enterprise agents) - Geography: United States - Round size: $100M - Valuation: roughly $750M ### Investor Profile Khosla Ventures led, with participation from Genesys. Genesys is a major contact-center platform, so its check doubles as a distribution signal into enterprise customer-experience deployments. ### Company and Leadership [Scaled Cognition](https://www.scaledcognition.com) was founded in 2022 by CEO Dan Roth, former Microsoft VP of conversational AI, CTO Dan Klein, a UC Berkeley NLP professor, and CFO Damon Pender. The three previously built Semantic Machines, which Microsoft acquired in 2018. The company is already in production with Fortune 500 firms in finance, healthcare, telecom, and insurance. ### Problem and Opportunity Hallucinations block AI from high-stakes workflows where a wrong answer carries real consequences. Reliability, not raw capability, is what gates enterprise adoption in these settings. ### Product and Technology Its flagship is APT-1, an Agentic Pretrained Transformer built from scratch for customer experience. Rather than generating new data, APT predicts actions, retrieves existing records, verifies tasks before confirming them, and enforces enterprise policy in the architecture itself. The model was trained on synthetic agentic data and hardened through agent-to-agent self-play. Because it is smaller, faster, and available for self-hosted deployment, enterprises own it outright instead of depending on a third-party provider. ### Use of Proceeds and Vision The $100M accelerates research and commercialization and extends the company beyond customer service into broader finance and other high-stakes use cases. It positions itself as a vertical AI “neolab” building deterministic intelligence for customer experience. ### Market Context Scaled Cognition targets the enterprise AI agent market across finance, healthcare, insurance, and telecom. It contrasts its purpose-built model against vendors that wrap general-purpose LLMs and add guardrails afterward. ## 4. Sail Research Raises $80M For Long-Horizon Agent Inference ### Deal Overview - Stage: combined seed and Series A - Sector: AI infrastructure (inference) - Geography: United States - Round size: $80M - Valuation: $450M ### Investor Profile Sequoia led the seed and Kleiner Perkins led the Series A. Backers include Redpoint, Theory Ventures, Vine Ventures, CRV, A*, and Abstract Ventures, plus angels John Hennessy, Lip-Bu Tan, and Tri Dao. ### Company and Leadership [Sail Research](https://www.sailresearch.com) was founded in late 2025 by CEO Neil Movva and CTO Samin Menon, who met on their first day of freshman year at Stanford. Movva worked on iPhone computer-vision chips at Apple and later did GPU-level inference at Together AI. Menon is an Apple security-engineering veteran. The company came out of stealth in June 2026. ### Problem and Opportunity Today’s inference stack was built for a human waiting at a prompt. Autonomous agents work for hours or days and consume tokens at 50 to 500 times the rate of chat apps. Even as per-token prices fall, enterprise AI bills have tripled because agents burn so many tokens per task. ### Product and Technology Sail rebuilt the inference stack to maximize throughput across thousands of concurrent calls rather than minimize latency on one request. It sells an inference engine that it says cuts cost per token by up to 10x, with customers reporting 3x to 10x gains, plus “Sailboxes,” sandboxes that run for hours or days. Early users include Detail.dev running code-review agents over long sessions. ### Use of Proceeds and Vision The $80M funds the throughput-optimized engine and Sailbox infrastructure to serve agents spending billions of tokens on a single task, plus team and capacity growth. ### Market Context Goldman Sachs forecasts roughly a 24-fold increase in token consumption by 2030. Sail’s main contrast is Together AI and other latency-focused providers, along with frontier labs building their own inference. Its bet is that throughput, not latency, is the right axis for agentic workloads. ## 5. Runlayer Raises $30M To Secure Corporate AI Agents ### Deal Overview - Stage: Series A - Sector: Cybersecurity (agent governance) - Geography: United States - Round size: $30M, lifting total funding to about $42M ### Investor Profile Felicis led, with Khosla Ventures. Both backed the earlier $11-12M seed. Vinod Khosla reportedly wanted to buy every available dollar of the round, and Felicis preempted the deal before a full competitive process could run. ### Company and Leadership [Runlayer](https://www.runlayer.com) was founded in 2025 by Andrew Berman, a three-time founder. He co-founded Nanit, the baby monitor that passed $100M in revenue, and Vowel, a video conferencing tool Zapier acquired in 2023, and later ran AI at Zapier. ### Problem and Opportunity As employees delegate more work to swarms of agents, companies need a governance layer the way human workforces need identity and access management. Without it, agents connect to corporate systems with no unified oversight. ### Product and Technology Runlayer builds security and compliance infrastructure around the Model Context Protocol, the open spec for how agents connect to corporate data. It spans 5 to 20 AI clients including IDEs, chat apps, and Salesforce Agentforce, offering scoped identity, granular permissions, approval workflows, audit logs, shadow-AI discovery, and runtime risk controls. Early adopters include Gusto, dbt Labs, Instacart, Opendoor, Decagon, AngelList, and Lemonade. ### Use of Proceeds and Vision The $30M expands engineering and go-to-market teams as enterprises shift from treating AI as an experiment to making it core strategy. Runlayer frames itself as a neutral, model-agnostic control layer, a “Switzerland business” that stays independent from the platforms it governs. ### Market Context Gartner projects 40% of enterprise applications will include AI agents by the end of 2026, up from under 5% in 2025. The agentic AI security market is valued around $55B in 2026. Runlayer competes near Wiz, Palo Alto Networks, and Okta but differentiates on its neutral, MCP-native stance. ## 6. Hera Raises $27M For AI Senior-Care Coordination ### Deal Overview - Stage: Series A - Sector: Healthtech (senior care) - Geography: New York, United States - Round size: $27M ### Investor Profile Bain Capital Ventures led, with continued support from earlier backers Accel and IA Ventures, plus angels including the CFO of Mount Sinai. ### Company and Leadership [Hera](https://hellohera.com) was founded in 2025 and is based in New York. CEO Jenny Lee is a software engineer who worked at Google and Uber, and she co-founded the company with Connie Kang and CTO Myles Novick. Hera serves more than 1,000 families with 95% retention. ### Problem and Opportunity Adult children, often daughters, spend hours on hold with Medicaid and juggling appointments for aging parents. Older adults without adequate support face 34% higher hospitalization rates, which Bain estimates drives about $37 billion in avoidable costs each year. ### Product and Technology Hera pairs human care experts, licensed geriatric social workers and nurses it calls “Heroes,” with an AI assistant named Juno that handles the non-clinical work around an appointment: scheduling specialists, filling prescriptions, arranging equipment, organizing meals, and navigating Medicare and Medicaid. Juno encodes hyper-local knowledge that has never existed in structured form, such as which cardiologist has openings next week. Hera operates as its own clinic and bills Medicare’s Chronic Care Management codes, so roughly 90% of clients pay nothing out of pocket. ### Use of Proceeds and Vision The funding expands Hera into California, Florida, Maryland, Pennsylvania, and Massachusetts, with a goal of operating in 25-plus states by the end of 2026. It scales both the Heroes workforce and the Juno platform. ### Market Context The opening came from 2024 Medicare reforms that created billing codes for care-management services, a tailwind similar to the coverage expansion that mental-health startup Alma rode. No direct head-to-head rival was named, and Hera is pairing that regulatory shift with technology in a fragmented elder-care market. ## Lessons For Founders - Proprietary data you already own can become a fundraising story. General Intuition and Hera both turned an existing dataset, gameplay clips and local care knowledge, into a moat investors paid up for. - Reliability sells in regulated markets. Scaled Cognition raised $100M by attacking hallucinations head-on, because in finance and healthcare a wrong answer blocks adoption entirely. - Contrarian architecture can win when the workload changes. Sail bet on throughput over latency precisely because agents behave nothing like chat users. - Profitability and walking away from buyouts strengthen your hand. Runpod bootstrapped to profit and rejected $500M-plus offers before raising at a $1B valuation on its own terms. - Regulatory shifts open timing windows. Hera built its model around new Medicare billing codes, and Runlayer is racing the rollout of enterprise AI agents. Match the product to the rule change.