US Seed Weekly Funding Roundup (Mar 17-24, 2026): $15.0M Raised Across 2 Deals

Table of Contents

Two seed-stage startups pulled in a combined $15 million this week, both betting on the agentic AI wave but from opposite ends. Manifold is building the security layer to keep autonomous AI agents in check, while Obin AI is deploying those agents inside financial institutions. The common thread: as AI agents move from demos to production, the infrastructure around them is where the real money flows.

These seed rounds sit alongside a much larger US funding week. Five Series A deals totaled $275.5 million, and eight late-stage rounds brought in $760 million. The contrast is telling. Seed capital is targeting narrow, technically deep problems where founder credibility matters more than revenue, while later-stage dollars chase companies already proving unit economics at scale.

1. Manifold Raises $8M For AI Agent Security

Deal Overview

• Stage: Seed
• Sector: Cybersecurity
• Geography: San Diego, United States
• Round Size: $8M
• Valuation: Not disclosed

Investor Profile

Costanoa Ventures led the round with participation from Cherry Ventures, Rain Capital, and Modern Technical Fund. Costanoa has a strong enterprise infrastructure portfolio and called endpoint agent security "the next major layer of enterprise infrastructure." Cherry Ventures, based in Berlin, brings European enterprise go-to-market expertise that could help Manifold's international expansion down the line.

Company and Leadership

Manifold was founded in 2025 by Neal Swaelens, Oleksandr Yaremchuk, and Michael McKenna. Swaelens and Yaremchuk previously co-founded Laiyer AI, creators of LLM Guard, the most widely adopted open-source LLM firewall. Laiyer AI was acquired by Protect AI in January 2024, which Palo Alto Networks then bought. The advisory board includes Joe Sullivan, former Uber CSO, and Vijay Bolina, former Google DeepMind CISO.

Problem and Opportunity

Roughly 85% of developer environments now run coding agents like Copilot, Claude Code, or Cursor. These agents don't just suggest text. They call APIs, read and write files, and execute commands across production systems. Existing AI security tools focus on the prompt and output layer, which is the wrong abstraction for multi-step autonomous behavior. There's a genuine blind spot: no one is watching what these agents actually do at runtime.

Product and Technology

Manifold's Agentic AI Detection and Response (AIDR) platform operates at the endpoint runtime layer. It provides full visibility into every tool called, system accessed, and action taken by AI agents. The platform maps each agent's connections to MCP servers, databases, and external systems in real time. Anomaly detection compares observed behavior against established baselines, and all interactions are recorded for forensics. Deployment is agentless and takes days, not months.

Use of Proceeds and Vision

The $8 million will go toward building out runtime monitoring capabilities, expanding anomaly detection models, and making early sales and customer success hires. Manifold positions itself as the EDR equivalent for the agentic era. Just as CrowdStrike became essential for traditional endpoints, Manifold wants to become the default security layer as AI agents spread from developers to every knowledge worker.

Market Context

The EDR market, the closest analogue, is a multi-billion dollar category dominated by CrowdStrike, SentinelOne, and Microsoft Defender. First-generation AI security vendors like Prompt Security, Aim Security, and Lakera focus on inference-time guardrails rather than runtime agent behavior. The longer-term competitive risk comes from CrowdStrike and Palo Alto Networks building native AIDR capabilities, but the category is new enough that Manifold has a real head start.

Case Studies

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

Read their stories →

2. Obin AI Raises $7M For Agentic Finance Workflows

Deal Overview

• Stage: Seed
• Sector: Fintech / Enterprise AI
• Geography: New York, United States
• Round Size: $7M
• Valuation: Not disclosed

Investor Profile

Motive Partners led the seed round. Motive is a specialist fintech private equity and venture firm, which makes their backing a strong signal of domain credibility. Angel investors include Dr. Fei-Fei Li, the Stanford AI pioneer, and Lukasz Kaiser, co-author of the original Transformer paper. That combination of financial services expertise and AI research pedigree is hard to assemble.

Company and Leadership

Obin AI emerged from stealth in March 2026. CEO Apoorv Saxena was formerly Head of AI at JPMorgan and ran multiple Google Cloud AI products including Translate and Contact Center AI. He also led AI-driven value creation across $100 billion or more in AUM. CTO Dr. Valliappa "Lak" Lakshmanan served as Operating Executive for Data and AI at Silver Lake, spent years at Google, and has written seven books on AI. The founding team already has the trust of financial institutions representing over $1 trillion in AUM.

Problem and Opportunity

Financial institutions spend billions each year on compliance, monitoring, and underwriting workflows that are still heavily manual. A single $500 billion private credit fund processes over 50,000 loan notices annually. These tasks require institutional knowledge, regulatory precision, and full audit trails. Generic AI tools don't understand the governance requirements, and hiring more analysts doesn't scale.

Product and Technology

Obin builds a production-ready AI workforce that runs defined workflows end-to-end inside a firm's existing controls and audit boundaries. The platform encodes institution-specific logic and produces traceable, inspectable outputs. Its open architecture means institutions retain full ownership of models, data, and IP. The system embeds historical context into the agent layer, enabling reasoning across multi-decade datasets and complex financial documents. In production, it automated the processing of 50,000+ annual loan notices for a major private credit fund, saving 70 to 90% of the time while improving accuracy.

Use of Proceeds and Vision

The funding will expand workflow coverage, deepen financial-domain AI capabilities, and scale enterprise deployments from pilot to production. Obin's goal is to become the enterprise-grade agentic workforce for global financial institutions, letting them standardize and accelerate core workflows without growing headcount or weakening governance.

Market Context

The enterprise AI market for financial services is one of the fastest-growing segments. Competitors range from generic platforms like Palantir AIP and C3.ai to financial-specific tools like Eigen Technologies and Kensho (owned by S&P). Large banks also run sizable internal AI teams. Obin differentiates with its open architecture, which removes the data lock-in concern that blocks most enterprise AI adoption in regulated industries. Motive Partners' backing validates the financial services focus in a way that a generalist VC round wouldn't.

Lessons For Founders

• Founder-market fit still wins at seed. Both teams have direct, verifiable experience in their target markets. Manifold's founders built the most adopted open-source LLM security tool. Obin's CEO ran AI at JPMorgan. At seed, investors are buying the team's right to build in a space, and domain credibility compresses sales cycles with enterprise buyers.
• New infrastructure categories emerge around behavioral shifts, not technology launches. AI agents aren't a future possibility. They're running in 85% of developer environments right now. Manifold and Obin aren't selling AI hype; they're solving problems that already exist at scale.
• Open architecture removes the biggest blocker in regulated industries. Obin's decision to let institutions keep full ownership of their models, data, and IP isn't a feature. It's a go-to-market strategy. In financial services and healthcare, data sovereignty concerns kill more deals than pricing ever does.
• Strategic angels can be worth more than the check. Fei-Fei Li and the Transformer paper co-author backing a fintech seed round sends a signal that no marketing budget can buy. Joe Sullivan (former Uber CSO) on Manifold's advisory board opens doors that cold outreach never will. Choose angels who de-risk your next fundraise and your first ten enterprise deals simultaneously.