Series B activity: July 2026
- 01AI Infrastructure Is the Dominant Investment Category
- 02Deep Tech / Hard Science Is Attracting Mega Rounds
- 03Energy Transition & Climate Tech Remain Active
- 04AI Security Is Emerging as a Distinct Vertical
- 05Supply Chain Automation Is Drawing Repeat Attention
1. Key Themes
AI Infrastructure Is the Dominant Investment Category
The largest deals of the month cluster around AI platforms — from materials discovery to enterprise automation to local inference — signaling that institutional capital has moved decisively from "AI apps" to foundational AI infrastructure.
"CuspAI builds AI systems that help scientists discover new materials faster... models that propose candidates, rank them for real-world constraints, and hand off to lab partners for validation." "Ollama makes it easy to run large language models locally—on a developer laptop, a workstation, or on-prem servers—without sending prompts and data to a hosted API."
Deep Tech / Hard Science Is Attracting Mega Rounds
Three of the five largest deals (CuspAI at $450M, Meshy at $400M, Quaise at $134M) are in physics-heavy or computationally intensive domains — materials science, geothermal drilling, and 3D generation — suggesting top-tier VCs are willing to commit large capital to long-cycle, high-barrier bets.
"A $450M Series B led by Kleiner Perkins and NEA gives CuspAI the runway to scale compute, data partnerships, and a bigger wet-lab validation loop." "Quaise is developing a new kind of drilling system aimed at making geothermal power viable almost anywhere, not just in rare 'good geology' locations."
Energy Transition & Climate Tech Remain Active
Multiple deals span the clean energy and sustainability spectrum — geothermal (Quaise), alternative proteins (The Protein Brewery), regenerative agritech (Rize), and circular economy analytics (Greyparrot) — indicating sustained LP appetite for climate-adjacent bets.
"The core idea is to reach much hotter rock deeper underground, turning geothermal into a more predictable baseload energy source that can plug into existing grid infrastructure."
AI Security Is Emerging as a Distinct Vertical
Two sizeable security-focused deals — Glow ($100M, endpoint security) and Onyx Security ($113M, AI governance) — reflect enterprise anxiety about the risks that come with rapid AI adoption.
"Onyx Security: Holistic platform for secure AI adoption and management." "Glow: AI-powered endpoint security platform preventing risky software."
Supply Chain Automation Is Drawing Repeat Attention
Multiple deals target procurement and supply chain intelligence — Auger ($50M, real-time supply chain insights), Freehand ($75M, AI agents for enterprise procurement), and Senra ($65M, manufacturing/assembly software) — pointing to a durable enterprise spending theme.
"Freehand: AI agents managing enterprise supply-chain spend and procurement." "Auger: AI-powered tool for real-time insights in supply chain management."
2. Contrarian Perspectives
Geothermal Is Becoming a Procurement-Level Conversation, Not a Science Project
The consensus view has long been that geothermal is niche and geology-dependent. The $134M Quaise raise challenges this, positioning geothermal as a scalable, baseload clean energy source suitable for broad grid integration.
"Geothermal is moving from niche to mainstream procurement conversations, with utilities and developers looking for firm clean power that complements wind and solar." The significance: unlike wind and solar, geothermal provides firm power — a capability gap the market increasingly needs to fill.
Local/Offline AI Inference Is Becoming Enterprise Standard, Not an Edge Case
The prevailing assumption is that hosted API models (OpenAI, Anthropic, etc.) are the default enterprise AI deployment pattern. Ollama's $65M raise suggests the opposite trajectory for regulated industries.
"That's becoming a default requirement in regulated environments and inside enterprises that want AI features but can't accept data leakage, latency, or unpredictable inference costs." "Smaller, faster models plus better tooling are turning offline inference from a novelty into a practical deployment pattern."
Materials Discovery Is Now an Engineering Workflow, Not Basic Research
The common assumption places AI-driven materials discovery in the realm of long-horizon R&D. CuspAI's positioning and $450M raise suggest it is being productized as a near-term industrial tool.
"The company's momentum is in turning 'materials discovery' into an engineering workflow: models that propose candidates, rank them for real-world constraints, and hand off to lab partners for validation." "The pull for better materials has become a near-term product problem, not a distant R&D wish."
3. Companies Identified
| Company | Description | Why Mentioned | Key Quote |
|---|---|---|---|
| CuspAI | AI-driven materials discovery platform | Largest deal of the month ($450M); named a top-3 most interesting deal | "Governments and industry pushing hard on energy storage, semiconductors, and industrial decarbonization." |
| Meshy | AI text-to-3D model generation platform | Second-largest deal ($400M); signals massive capital flowing into generative 3D | Listed among top deals by size |
| Simile | AI platform for automating communication and collaboration | Third-largest deal ($200M from Greenoaks) | "AI-driven platform for automating communication and collaboration." |
| Augustus | AI-native stablecoin clearing and banking for institutions | Large fintech/crypto deal ($180M from Tiger Global) | "AI-native stablecoin clearing and banking platform for institutions." |
| Kraken Technology | Autonomous maritime platforms for defense and industry | Large defense tech deal ($175M) | "Autonomous maritime platforms for defense, security, and industry." |
| Quaise | Novel geothermal drilling systems | Named top-3 most interesting deal; $134M raise | "Turning geothermal into a more predictable baseload energy source." |
| Ollama | Local LLM inference platform | Named top-3 most interesting deal; $65M raise | "Local-first tailwind: smaller, faster models plus better tooling." |
| Onyx Security | Secure AI adoption and management platform | $113M raise; reflects enterprise AI governance urgency | "Holistic platform for secure AI adoption and management." |
| Glow | AI-powered endpoint security | $100M raise backed by Sequoia, Greenoaks, Cyberstarts | "AI-powered endpoint security platform preventing risky software." |
| Twelve Labs | Video intelligence platform for large video libraries | $100M raise; niche but scalable AI infrastructure | "Video intelligence platform for searching and analyzing large video libraries." |
| Lyzr AI | Enterprise agent framework with self-learning AI | $100M raise; privacy-focused agentic AI | "Enterprise agent framework for self-learning AI and data privacy." |
| Dwelly | AI-driven lettings and property management | $95M from EQT and General Catalyst; notable proptech bet | "AI-driven lettings and property management platform." |
| Venus Aerospace | Hypersonic transportation for global travel | $91M raise; bold deep tech bet | "Innovative hypersonic transportation for global travel." |
| Freehand | AI agents for enterprise procurement | $75M from Battery Ventures; supply chain AI | "AI agents managing enterprise supply-chain spend and procurement." |
| Senra | Manufacturing software for assembly line automation | $65M from Lowercarbon Capital | "Manufacturing software for configuring and automating assembly lines." |
| Databento | Simplified financial market data access | $97M from NEA; picks-and-shovels fintech | "Simplified access to financial market data with flexible pricing." |
| Qolab | High-quality qubit production using semiconductor processes | $54.2M; quantum hardware play backed by UC Investments | "High-quality qubit production using advanced semiconductor processes." |
| Cashea | BNPL for underserved Latin American markets | $60M; emerging market fintech | "Buy Now, Pay Later for underserved Latin American markets." |
| Greyparrot | AI waste recognition and circular economy analytics | Notable ESG/AI intersection | "AI waste recognition and circular economy analytics." |
| Monumental | Robotic on-site construction automation | $32M from Khosla Ventures | "Automated on-site construction using robotics and software." |
| Risk Ledger | Vendor risk management for supply chain security | $32.1M; supply chain security | "Vendor risk management platform for secure supply chain ecosystems." |
| Epitel | Wearable EEG monitoring for epilepsy | $26M; medtech wearables | "Wearable EEG monitoring for epilepsy care." |
| Hilo | Continuous blood pressure monitoring | $19M; health wearables | "Continuous blood pressure monitoring health technology." |
| T2 | Level 4 autonomous driving for logistics trucks | $30.7M; Japan-based, broad corporate syndicate | "Autonomous driving systems and logistics services with Level 4 trucks." |
4. People Identified
| Person | Description | Why Mentioned | Quote |
|---|---|---|---|
| Aaron Harris | Co-author, PostRound; formerly YC Group Partner | Author of the newsletter | Byline credit |
| Jacob Dennis | Co-author, PostRound | Author of the newsletter | Byline credit |
Note: The article does not profile specific founders or executives beyond deal-level data.
5. Operating Insights
Position "Local-First AI" as a Compliance and Cost Feature, Not a Technical Limitation
Enterprises in regulated industries are increasingly requiring on-premise AI deployment. Founders building AI products should offer local inference as a first-class option — not an afterthought — framing it around data sovereignty, latency, and predictable cost rather than capability constraints.
"That's becoming a default requirement in regulated environments and inside enterprises that want AI features but can't accept data leakage, latency, or unpredictable inference costs."
Validate Via Workflow Integration, Not Standalone Tools
CuspAI's framing reveals a winning enterprise GTM pattern: don't sell a research tool, sell an engineering workflow with clear handoffs (propose → rank → validate). Products that embed into existing team processes — rather than requiring new standalone behavior — earn stickier adoption.
"Models that propose candidates, rank them for real-world constraints, and hand off to lab partners for validation."
Hardware-Intensive Startups Should Anchor Narratives to Grid/Procurement Urgency
Quaise's ability to raise $134M for a capital-intensive drilling company reflects a narrative shift: when a hard-tech solution maps to an active procurement need (utilities buying firm clean power), fundraising becomes easier. Founders in energy and industrial tech should anchor pitches to buyer urgency, not technology elegance.
"Utilities and developers looking for firm clean power that complements wind and solar."
6. Overlooked Insights
The Mean-Median Gap Signals a Barbell Market Structure
The average deal size ($80M) is significantly larger than the median ($55M), meaning a handful of outsized rounds (CuspAI at $450M, Meshy at $400M, Simile at $200M) are pulling the average up substantially. The typical Series B in this dataset is a more modest $55M — the megadeals are rare outliers, not the norm.
"The average deal size was $80.0 million while the median was $55.0 million."
Pediatric Health is Quietly Building Critical Mass
Two pediatric-focused companies raised in the same month — Pediatrica Health Group ($28M, pediatric healthcare services) and Handspring Health ($19M, pediatric mental health) — suggesting a quiet but growing investor thesis around underserved pediatric care infrastructure, particularly mental health.
"Handspring Health: Pediatric mental health care platform." "Pediatrica Health Group: Pediatric healthcare services group."