🚀 The Startup Khan Newsletter
Today’s big picture: Manus returns to independent ownership with fresh funding. Nous Research reaches a $1.5B valuation. And a $1.8B biology initiative shows AI expanding beyond software.
🧠Top Story: Manus Raises More Than $500M
AI-agent startup Manus has raised more than $500 million following the collapse of its reported $2B-plus acquisition by Meta.
The round was co-led by Boyu Capital and IDG Capital, with participation from Tencent, Sequoia China and ZhenFund.
Chinese regulators ordered the Meta transaction unwound in April amid scrutiny of cross-border AI investments.
Now Manus is back as an independent company.
Why founders should care
Manus builds general-purpose AI agents that execute research, automation and multistep digital workflows.
Its story highlights three realities:
Distribution matters. Great agents still need paying customers.
Geopolitics matters. Regulatory barriers can derail even attractive acquisitions.
Independence matters. Strong customer demand can give startups alternatives to selling.
👉 Founder takeaway: The next AI-agent winners won’t simply execute tasks. They’ll own customer relationships, trust and recurring revenue.
💰 Funding Watch: Nous Research Hits $1.5B
Open-source AI startup Nous Research raised $90 million in Series B funding, reaching a reported $1.5 billion valuation.
Investors include Robot Ventures, Nvidia, Union Square Ventures, Menlo Ventures and Samsung.
Its open-source Hermes Agent is gaining developer adoption, while the company is expanding into enterprise deployments.
The business model:
Open source → Developer adoption → Enterprise deployment → Recurring revenue
Enterprise customers pay for capabilities such as security, privacy, permissions and organizational controls.
👉 Founder lesson: Open source can build distribution. Enterprise infrastructure can turn that distribution into revenue.
The challenge is making those paid capabilities valuable enough to retain customers.
🧬 AI Science: $1.8B for Biological Intelligence
Biohub, U.S. government agencies and technology partners have announced a $1.8 billion coordinated initiative supporting AI models of biology.
This is not a venture-capital round. It combines research funding, infrastructure, datasets and scientific resources.
The goal is to develop better data for AI systems that can predict biological processes.
Why does it matter?
Generic internet data isn’t enough to teach AI complex science.
Specialized models need:
Experimental data → Expert validation → Evaluation systems → Scientific infrastructure
👉 Founder signal: The next AI opportunities may emerge from building the specialized data and tools that make advanced models useful in science, healthcare and engineering.
🔴 Founder Opportunity Radar
🚨 October 8 — Jasiri Growth Accelerator
Applications close today for eligible ventures in Kenya and Rwanda. Selected companies may receive up to $75,000 in combined funding and support, subject to program terms.
October 15 — BitRoot AI Accelerator
Cairo-based accelerator targeting AI-native startups across Egypt and MENA. Potential funding of up to $250,000 for qualifying companies.
Deadline: 11:59 p.m. Cairo time.
November 18 — Techstars 2027
Applications remain open for several accelerator programs, including London, New York City, Chicago and Anywhere.
âš¡ Founder Framework: The Agent Independence Test
Before building an AI-agent startup, answer four questions.
CONTROL — Can customers retain their data and workflows when switching models?
DISTRIBUTION — Can you acquire customers without depending on one platform?
ECONOMICS — Does recurring revenue comfortably exceed model, infrastructure and support costs?
DEFENSIBILITY — Does your product become harder to replace as customers add integrations and knowledge?
If the answer is no, you may have a useful feature rather than a durable company.
Quick win: Audit your startup against these four questions. Identify the weakest answer and make it your next strategic priority.
🧩 Final Thought
Manus and Nous Research represent two paths toward independent AI-agent businesses.
One has secured major financing after a failed acquisition.
The other is turning open-source adoption into an enterprise business.
Both point toward the same lesson:
AI capabilities will improve. Model costs may fall. Competition will intensify.
The winners will need more than impressive agents.
They’ll need distribution, customer trust, sustainable economics and defensibility.
The agent performs the work.
The business owns the relationship.
— The Startup Khan
Your 5-minute startup edge — funding news, tools, and founder playbooks.
​
​