Back
Table of Contents
Four US seed-stage startups raised a combined $31.8M this week, with deals split evenly between AI software and deep tech hardware. The largest round went to Littlebird at $11M for its privacy-first AI assistant, while Applied Atomics pulled in $8.3M to build modular nuclear power plants for data centers. The week's seed activity ran alongside a much larger US Series A class that closed $320M across four deals and a US Series B+ cohort totaling $327M across five.
A theme runs through this week's seed deals: founders with deep domain credibility tackling problems that require years of specialized knowledge to even attempt. Two of the four companies are led by ex-SpaceX and ex-JPL engineers building physical infrastructure. The other two are repeat founders with prior exits north of $1B combined. Investors are clearly betting on founders who've already proven they can ship in hard domains.
1. Littlebird Raises $11M for Full-Context AI Assistant
Deal Overview
- Stage: Seed
- Sector: AI / Productivity
- Geography: San Francisco, CA
- Round size: $11M
- Valuation: Not disclosed
Investor Profile
Lotus Studio led the round, marking the company's first institutional capital. Angel investors include Lenny Rachitsky, the former Airbnb product lead whose newsletter reaches hundreds of thousands of product managers, and Scott Belsky, Adobe's Chief Strategy Officer and founder of Behance. The angel list signals strong product-community distribution potential from day one.
Company and Leadership
Littlebird was founded in 2024 by Alexander Green, Alap Shah, and Naman Shah. The company is building a native macOS application with Windows on the waitlist and mobile companion apps planned. It's the team's first venture-backed company, though the angel roster suggests deep ties across Silicon Valley's product and design communities.
Problem and Opportunity
Every AI assistant today starts every conversation from zero. Users spend minutes setting context before they can get useful answers. Littlebird's bet is that an assistant already aware of what you've been working on, across every app, can skip straight to being helpful. The friction they're targeting is real: knowledge workers lose significant time re-explaining their work to tools that should already know.
Product and Technology
Littlebird reads the structured content of every application on a user's screen in real time using text-based parsing, not screenshots. It transcribes meetings, builds a searchable memory of user activity, and grounds AI responses in actual work context. The text-based approach uses 90% less storage than screenshot alternatives like Microsoft Recall. Security specs are serious: AES-256 encryption at rest, TLS 1.3 in transit, SOC 2 certified, GDPR and CCPA compliant. Password managers and sensitive fields are automatically excluded. The company reports 84% of users save at least half a day per week.
Use of Proceeds and Vision
Funds will scale the platform, expand the engineering team, and push into enterprise adoption. The company is building SaaS, private VPC, and on-premises deployment options to serve organizations with strict data residency requirements. Their tagline captures the ambition: "What if your AI assistant was already in the room?"
Market Context
The desktop AI productivity space is heating up. Microsoft Recall generated significant controversy over privacy when announced, and Littlebird is positioning directly against that anxiety. The broader AI assistant market is projected to reach tens of billions in the coming years, but the privacy-first segment is still early. Littlebird's challenge will be expanding beyond macOS power users to a broader enterprise audience.
Startups like yours already closed their rounds with us.
Founders across every stage and industry. Here's what it took.
- Raised $7.6M for Swiipr Technologies
- Raised $0.5M for Ap Tack
- Raised €0.5M for Ivent Pro
2. Applied Atomics Raises $8.3M for Modular Nuclear Power Plants
Deal Overview
- Stage: Seed (oversubscribed)
- Sector: Deep Tech / Energy
- Geography: Los Angeles, CA
- Round size: $8.3M ($12M total in first 12 months)
- Valuation: Not disclosed
Investor Profile
Morpheus, Transition, and Alpaca VC led this oversubscribed round. The investor mix skews toward climate and deep tech specialists willing to back long regulatory timelines. For nuclear startups, the quality of the cap table matters as much as the capital since these companies need patient investors who understand 36-48 month licensing cycles.
Company and Leadership
Applied Atomics was founded in early 2024. CEO Ben Kellie is an ex-SpaceX engineer who led development of SpaceX's landing barges and later founded The Launch Company, which exited to Voyager Space in 2021. CTO Paul Keutelian was a SpaceX Responsible Engineer and co-founded Radiant Industries, which built the world's first portable 1.2 MWe microreactor. The VP-level team draws from Virgin Orbit, Relativity Space, ABL Space Systems, and Gravitics.
Problem and Opportunity
Data center electricity demand is on track to double from 460 TWh in 2022 to over 1,000 TWh by 2026. Hyperscalers like Microsoft, Google, Amazon, and Meta are all actively seeking carbon-free nuclear power but face a supply bottleneck. Goldman Sachs projects data center power demand could rise 160% by 2030. The gap between what's needed and what's available is enormous.
Product and Technology
Applied Atomics builds 100MW to 1GW modular nuclear power plants using proven light-water reactor architecture, Generation III small modular reactors. The design scales from one to ten modules per deployment. The company doesn't just design reactors. It builds, owns, and operates full plants deployed on-site at customer campuses. Each plant generates roughly $100M in recurring revenue through long-term power purchase agreements. They're using the NRC Part 50 licensing pathway, which carries a 36-48 month timeline.
Use of Proceeds and Vision
The capital will activate multiple engineering test stands, starting with a studio in LA's Arts District and a second facility breaking ground in 2026. The team is also advancing NRC licensing and accelerating customer siting discussions with hyperscalers. Their model borrows directly from the SpaceX playbook: vertical integration from design through operations, with each deployment feeding data back into the next.
Market Context
2026 is shaping up as a turning point for nuclear. Over 15 new reactors are under construction globally, and China's Linglong One SMR is scheduled for commercial operations in H1 2026. Nuclear co-location can reduce energy costs 15-30% by eliminating transmission infrastructure. Applied Atomics is competing for the same hyperscaler contracts as larger nuclear firms, but its SpaceX-style operational model could give it a speed advantage on deployment timelines.
3. Moda Raises $7.5M for AI Design Platform
Deal Overview
- Stage: Seed
- Sector: AI / Design
- Geography: New York, NY (SoHo)
- Round size: $7.5M
- Valuation: Not disclosed
Investor Profile
General Catalyst led, with Pear VC and WndrCo participating. General Catalyst's involvement is a strong signal given their portfolio depth in enterprise software. Their investment memo noted that "in the AI software era, design and brand judgment is a true moat," which tells you what thesis they're backing here.
Company and Leadership
Moda launched publicly on March 24, 2026. The founding team carries serious weight. CEO Anvisha Pai is an MIT alumna, ex-Dropbox PM, and co-founder of Dover, which raised $23M from YC, Founders Fund, and Tiger Global. COO Ravi Parikh co-founded Heap (valued at $960M) and Airplane (acquired by Airtable). CTO John Holliman was employee #1 at Dover and previously worked at Expanse, which Palo Alto Networks acquired. Combined exits across the founding team exceed $1.2B. The company already has 3,000+ active users from Google, McKinsey, Wix, and Samsara after one month of beta.
Problem and Opportunity
AI design tools generate plenty of output. Most of it looks generic. Moda calls this "AI Slop" and it's a real problem for companies that care about brand consistency. When a marketing team uses Gamma or Tome to create a deck, the result rarely matches the company's visual identity. Every output needs manual cleanup, which defeats the purpose of using AI in the first place.
Product and Technology
Moda runs a multi-agent system built on its Deep Agents framework. A Design Agent handles real-time creation, a Research Agent pulls content, and a Brand Kit Agent ingests and indexes visual assets. The system continuously scans websites, Google Drive, past slide decks, and brand files to build what the team calls a "living brand memory." The canvas is fully editable with layers, similar to Figma or Canva, and exports natively to Google Slides and PowerPoint. Think of it as Canva with a Cursor-style AI sidebar that actually knows your brand.
Use of Proceeds and Vision
Funds go toward product development, platform scaling, and team growth at the SoHo headquarters. The company is targeting enterprise and growth-stage teams where brand consistency directly affects revenue.
Market Context
Moda enters a market where Gamma just raised a $68M Series B at a $2.1B valuation with 70M+ users and over $100M in ARR. Tome, Beautiful.ai, and Canva's Magic Design feature are all active competitors. Moda's bet is that brand-aware output quality will matter more than volume as AI-generated content floods the market. Early traction from enterprise names like Google and McKinsey suggests the thesis has legs.
4. Airbase Raises $5M for Radio-Frequency Spectrum Coordination
Deal Overview
- Stage: Seed
- Sector: Deep Tech / Government
- Geography: New York, NY
- Round size: $5M
- Valuation: Not disclosed
Investor Profile
Andreessen Horowitz led, with Squadra Ventures and Founders You Should Know participating. a16z's American Dynamism practice has been actively backing defense and government tech startups, and Airbase fits squarely in that thesis. For a 7-person company, landing a16z as your seed lead is a strong validator.
Company and Leadership
Airbase was founded in 2025. CEO Ari Rosner previously worked at JPL on the Mars Perseverance and Europa Clipper missions before joining True Anomaly. CTO Millen Anand comes from Planet Labs and previously did RF payload engineering on Boeing geostationary satellites. The team is just seven people, but they've already secured a federal government contract and are negotiating their first Defense Department deal. Federal regulators are using the software today.
Problem and Opportunity
Radio-frequency spectrum is finite, and the process for licensing and coordinating it is stuck in the past. The FCC relies on manual processes, decades-old databases, and static PDFs. Getting spectrum allocated can take months or years. Meanwhile, connected devices, satellite constellations like Starlink and Amazon Kuiper, and 5G rollouts are creating exponential demand for bandwidth. The FCC has an AWS-3 auction deadline of June 23, 2026, adding urgency to modernization.
Product and Technology
Airbase's platform uses proprietary AI pipelines to ingest, connect, and translate decades of fragmented government records into unified, real-time data layers. The goal is to transform spectrum from a static utility into a software-defined resource where machines can negotiate bandwidth in milliseconds. The data moat is significant: organizing years of unstructured government records creates barriers that are hard to replicate. Federal regulators are already on the platform, creating institutional lock-in early.
Use of Proceeds and Vision
Most of the $5M will go toward hiring beyond the current 7-person team. The rest funds platform development and government and defense customer acquisition. The company describes spectrum as "the unseen highway that carries the data on which our economy and national security depend."
Market Context
The SATCOM market reached $27.6B in 2026. The 5G satellite communication market sits at $6.8B today and is projected to hit $44.4B by 2034, growing at 22.5% CAGR. Airbase is one of very few startups tackling spectrum coordination with modern software, which gives it a first-mover position in a space where government adoption cycles are long but contract values are high.
Lessons for Founders
- Domain credibility opens doors that pitch decks can't. Three of four companies this week are led by founders with direct operational experience in their target market. Applied Atomics' SpaceX veterans, Airbase's JPL and Planet Labs alumni, and Moda's $1.2B in combined exits all gave investors confidence to write checks at the seed stage for ambitious bets.
- Government and enterprise customers before fundraising changes the conversation. Airbase had a federal contract and regulators using its software before closing its seed. Moda had 3,000 users from Google and McKinsey within a month. Early revenue or adoption turns a fundraise from a bet on potential into a bet on trajectory.
- Privacy as a product feature, not a constraint. Littlebird's text-based approach to screen reading isn't just a privacy checkbox. It's a technical architecture that uses 90% less storage and enables better searchability than screenshot-based competitors. The best privacy-first products find ways to make the privacy constraint a performance advantage.
- Vertical integration still wins in hard-tech. Applied Atomics isn't selling reactor designs. It builds, owns, and operates entire power plants, generating $100M in recurring revenue per deployment. In capital-intensive industries, controlling the full stack creates defensibility that component sellers can't match.
- The best moats come from organizing messy data others won't touch. Airbase's entire competitive advantage comes from unifying decades of fragmented government spectrum records. The work is tedious, unglamorous, and creates a data asset that's nearly impossible to replicate. Founders should look for industries drowning in unstructured data that incumbents have no incentive to clean up.
Get your round closed. Not just pitched.
A structured fundraising process matched to your stage and investor fit.
- Fundraising narrative and structure that holds up
- Support from strategy through investor conversations
- Built around your stage, model, and timeline