Back
Table of Contents
US Series A activity this week was dominated by a single outsized deal. Starcloud pulled in $170M to build data centers in orbit, accounting for more than three-quarters of the $222M raised across four rounds. The remaining three deals split $52M between fintech, cybersecurity, and payments. Compared to the $29.5M raised across four US seed deals this week, Series A capital was nearly 8x larger, while late-stage rounds dwarfed both with $939M across seven US Series B+ deals.
Two themes stand out. First, the AI infrastructure bottleneck is pushing founders toward unconventional solutions, from space-based compute to identity governance for AI coding agents. Second, investor appetite for agentic AI applications remains strong, with Variance's autonomous fraud investigation platform drawing backing from cybersecurity-specialist VCs. The deals reflect a market where the biggest checks go to companies tackling physical constraints, while smaller rounds reward sharp vertical focus.
1. Starcloud Raises $170M for Orbital Data Centers
Deal Overview
- Stage: Series A
- Sector: Deep Tech / Space Infrastructure
- Geography: Seattle, United States
- Round Size: $170M
- Valuation: $1.1B (unicorn)
Investor Profile
Benchmark led the round alongside EQT Ventures, Macquarie Capital, NFX, Y Combinator, and 776 Ventures. Benchmark's involvement is notable given its track record backing category-defining infrastructure companies. Macquarie Capital brings deep energy and infrastructure expertise, signaling that institutional capital sees orbital compute as real infrastructure, not science fiction.
Company and Leadership
Starcloud was founded in January 2024 by Philip Johnston (CEO, ex-McKinsey, previously co-founded Opontia), Adi Oltean (CTO, ex-SpaceX and Microsoft Azure), and Ezra Feilden (ex-Airbus Defence and Space). Originally called Lumen Orbit, the company joined Y Combinator's Summer 2024 cohort and reached unicorn status in 17 months post-YC, making it the fastest YC startup to hit that milestone.
Problem and Opportunity
AI training and inference demand is outpacing terrestrial power supply. Data center operators face multi-year permitting delays and grid constraints. New facilities compete for scarce electricity with residential and industrial users. Starcloud's thesis is simple: space has unlimited solar power, natural cooling, and no land-use conflicts.
Product and Technology
The company builds orbital data centers powered by 24/7 solar energy in low Earth orbit, claiming 90% lower electricity costs than ground-based facilities. Starcloud 1 launched in November 2025 carrying an Nvidia H100 GPU as a proof of concept. Starcloud 2 will follow later in 2026 with multiple GPUs including Nvidia Blackwell chips, an AWS server blade, and a bitcoin mining computer. The third-generation Starcloud 3 is a 200-kilowatt, 3-ton spacecraft designed for SpaceX Starship deployment, targeting cost-competitive compute at roughly $0.05/kWh. The long-term plan calls for 88,000 data center satellites.
Use of Proceeds and Vision
Funds will go toward building and launching Starcloud 2 and Starcloud 3, expanding the team, and advancing toward the full constellation. The company positions itself as the solution to AI's power bottleneck, arguing that orbital infrastructure is the only path that scales without competing for terrestrial resources.
Market Context
BIS Research projects the in-orbit data center market at $1.77B by 2029 and $39.09B by 2035. Competitors include Aetherflux (raising at a $2B valuation), Google's Project Suncatcher, Aethero, and SpaceX, which has applied for a 1 million satellite compute constellation. The space is heating up fast.
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. Variance Raises $21.5M for AI Fraud Investigation
Deal Overview
- Stage: Series A
- Sector: Fintech / Risk & Compliance
- Geography: San Francisco, United States
- Round Size: $21.5M
- Total Raised: $26M
Investor Profile
Ten Eleven Ventures led, joined by 645 Ventures, Y Combinator, Urban Innovation Fund, and Okta Ventures. Ten Eleven is a cybersecurity-specialist fund, and their lead signals conviction that agentic investigation represents a genuine shift beyond traditional rule-based detection. Okta Ventures' participation ties into the identity verification angle of the platform.
Company and Leadership
Variance was founded in 2023 by Karine Mellata (CEO) and Michael Lin. Both came from Apple's fraud engineering and algorithmic risk team. The company originally operated under the name Intrinsic before rebranding.
Problem and Opportunity
Legacy rule-based fraud and compliance systems produce false positive rates above 95%. Human investigators spend days or weeks reviewing cases that could be resolved in minutes with proper data synthesis. Financial crime compliance is a $29B market in 2026, growing to nearly $70B by 2034, and most of that spend goes to manual processes that don't scale.
Product and Technology
Variance built an agentic AI platform that autonomously executes fraud, risk, compliance, and KYC/AML investigation workflows. It processes over 70 million context signals per day and executes roughly 300,000 automated enforcement actions. The data access layer connects to 150+ global business registries, sanctions lists, court dockets, adverse media sources, and identity verification platforms across the surface, deep, and dark web. AI agents reason over fragmented data and return fully auditable decisions. At one Fortune 500 client, the platform achieved human-level precision on KYC investigations, fully replacing manual review queues.
Use of Proceeds and Vision
The capital will deepen investment in the investigative AI infrastructure and expand work with financial institutions and Fortune 500 companies. Variance's core bet is that the industry will shift from detection and alerting to autonomous investigation, and they want to own that transition.
Market Context
The AML market alone is projected to grow from $4.4B to $22.7B by 2035. Competitors include Featurespace, Feedzai, Unit21, Sardine, Alloy, and legacy players like NICE Actimize and Oracle. The opportunity lies in replacing a workflow that's still overwhelmingly manual at most institutions.
3. BlueFlag Security Raises $16.5M for Developer Identity Security
Deal Overview
- Stage: Series A
- Sector: Cybersecurity
- Geography: San Francisco, United States
- Round Size: $16.5M (Series A), $28M total raised
- Revenue Growth: 300% YoY
Investor Profile
Maverick Ventures and Ten Eleven Ventures co-led. Maverick also led the $11.5M seed round, doubling down after seeing 300% revenue growth and a 5x increase in Fortune 500 customers during 2025. Ten Eleven's presence in both this deal and Variance's round shows the firm's active pace in security-adjacent AI this week.
Company and Leadership
BlueFlag Security was founded in 2024 by Raj Mallempati (CEO) and Ken Schneider. The company emerged from stealth in March 2024 and was named an IDC Innovator for Software Development Life-Cycle Identity and Access in 2025.
Problem and Opportunity
Most application security tools scan code and artifacts. They miss the question of who has access to what within the software development lifecycle, and what they're doing with it. As AI coding agents like Copilot, Cursor, and Devin become standard tools, the attack surface expands to include non-human identities that traditional AppSec doesn't track.
Product and Technology
The platform has three pillars: Developer Identity Security (behavioral baselines, anomaly detection, overprivilege scoring), Developer Tool Posture Management (CI/CD and repository security), and Open-Source Software Risk Management. New capabilities include AI Agent Governance, which tracks AI coding agents operating in the SDLC, and Developer Behavioral Risk Analysis for detecting shadow AI usage and scoring AI contributions against approval workflows.
Use of Proceeds and Vision
BlueFlag will accelerate platform development and expand across the US and EMEA, focusing on regulated industries and organizations scaling AI-driven development. Their thesis is that governing non-human identities in the SDLC will become a requirement, not a nice-to-have, as AI agents write and deploy more production code.
Market Context
The company operates within the $30B+ cybersecurity software market. Adjacent players include Cycode, Apiiro, Ox Security, Legit Security, and Cider (now part of Palo Alto Networks). The ASPM category is consolidating, but BlueFlag's identity-first approach carves a distinct lane.
4. Chexy Raises $14M for Canadian Rental Payments
Deal Overview
- Stage: Series A
- Sector: Fintech / Payments
- Geography: Toronto, Canada
- Round Size: $14M CAD
- Key Metrics: 200K+ users, $1B+ annualized payment volume
Investor Profile
Khosla Ventures led the round, with participation from returning investors Crossbeam and Venrex, plus Air Canada as a strategic investor. Khosla's involvement brings significant Silicon Valley credibility to a Canada-focused company. Air Canada's investment ties directly to the Aeroplan loyalty integration, which gives Chexy distribution access to over 10 million program members.
Company and Leadership
Chexy was founded in 2022 by Lizaveta Akhvledziani (CEO), Abtine Monavvari (CPO), and Ben Gigone (CTO). The team came through the Antler incubator and has grown the platform to over $1B in annualized payment volume.
Problem and Opportunity
Canada has roughly 15 million renters. Rent, taxes, childcare, and insurance are recurring expenses that don't accept credit cards, meaning consumers miss out on rewards they'd earn on other spending. Bilt Rewards proved this model at a $3.1B valuation in the US, but it doesn't operate in Canada. That gap is Chexy's entire business.
Product and Technology
Chexy lets Canadians pay rent and other recurring bills using their credit cards. The platform handles the payment rails: it charges the user's credit card and sends a bank transfer to the landlord or payee. The fee is 1.75%, the lowest in the market compared to 2-3% charged by competitors. Users earn their normal credit card rewards, whether that's cashback, Aeroplan miles, or Amex points. Rent payments are automatically reported to Equifax for credit building. No landlord participation is required.
Use of Proceeds and Vision
Chexy plans to expand across Canada, grow its partner ecosystem, and extend beyond rent into business payments and other household expenses. The company has explicitly ruled out US expansion, choosing instead to dominate a single market. Their vision is to become the financial hub for Canadian households where no essential payment goes unrewarded.
Market Context
Competitors in Canada include Plastiq, RentMoola, and Get Digs. The rewards-on-rent model is proven by Bilt's success in the US. Credit card interchange economics make it viable at high volume, and Chexy's $1B+ annualized volume suggests it's already past that threshold. Users have earned over $35M in rewards since launch, which drives strong word-of-mouth growth.
Lessons for Founders
- Physical constraints create the biggest moats. Starcloud's $170M round shows that investors will write large checks for companies tackling tangible bottlenecks. When your solution requires orbital spacecraft, copycats can't spin up in a weekend.
- Proven models in new geographies still work. Chexy didn't invent rewards-on-rent. Bilt proved it in the US at a $3.1B valuation. Chexy's bet was that Canada's 15M renters deserved the same product, and Khosla agreed. If a model works elsewhere and your market lacks it, that's a real opportunity.
- Deep domain teams attract specialist capital. Variance's Apple fraud engineering pedigree drew Ten Eleven (cybersecurity-specialist VC). BlueFlag's identity-first thesis drew the same firm. Specialist investors bring more than money: they bring customers, credibility, and pattern recognition in your exact category.
- Time your product to a shifting attack surface. BlueFlag built AI Agent Governance before most companies even have policies for AI coding tools. Being ready when the market realizes it has a problem is worth more than being right three years too early.
- Pick your market and own it. Chexy's refusal to expand to the US isn't a limitation. It's a strategy. Depth in one market, with 200K users and $1B in volume, beats a shallow presence in two.
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