Watch Tech Series · Startup Funding Watch | September 28, 2026
A $3.36B neocloud note, two $400M security rounds, and a wave of AI data and agent financings — where startup capital moved this week, and what it says about the market.
Startup capital this week didn't chase the smartest model. It chased the compute, data, and security that let AI run in production.
Across 13 tracked rounds worth roughly $5.9 billion, the money clustered in three places: AI infrastructure (Nscale, Verda), the data that trains and validates models (Snorkel AI, Micro1), and the security and agent software that has to be trusted before enterprises deploy any of it (Island, Cyera, Ema, Ando). The core question for this edition: where is startup capital moving, and what does it reveal about the technology market? It is a companion to our recent The Term Sheet article.
1. Funding Snapshot: The Week's Rounds at a Glance
Source: Crunchbase News / Company Announcements
Crunchbase's tally for September 19–25 was led by two $400 million cybersecurity rounds, while the largest single financing of the week sat outside that list: Nscale's $3.36 billion in pre-IPO convertible notes, reported September 25. Weekend deal flow was light, so this table reflects the announcements that landed through September 25 plus company disclosures earlier in the week.
| Company | Round | Amount | Valuation | Lead / Key Investors | Theme |
|---|---|---|---|---|---|
| Nscale | Pre-IPO convertible notes | $3.36B | IPO target up to ~$35B (reported) | Third Point; Nvidia ($1B, Nov.) | AI infrastructure |
| Island | Series F | $400M | $6.4B | Evolution Equity Partners | Enterprise security |
| Cyera | Series G extension | $400M | Not disclosed | Evolution Equity; Goldman Sachs Growth Equity | Data security |
| Snorkel AI | Series E | $350M | $3.2B (Crunchbase); Reuters reported $3.5B | Insight Partners; S32 | AI training data |
| Enveda | Series E | $311M | Not disclosed | Catalio Capital | AI drug discovery |
| Precision Neuroscience | Series D | $250M | Not disclosed | Pershing Square; Ackman Oxman Institute | Neurotech |
| Hubble Network | Series C | $200M | $1.6B | Smith Point Capital | Space connectivity |
| Verda | Series B + additional | $189M | >$1B (Bloomberg-attributed) | Emergence Capital | European AI cloud |
| Rightway | Series E | $155M | Not disclosed | Francisco Partners | Health benefits / AI |
| Numeral | Series C | $100M | Not disclosed | Insight Partners | AI tax compliance |
| Micro1 | Reported new round | >$100M (reported) | ~$4B (Forbes, sources) | Not disclosed | AI training data |
| Ema | Series B | $77M | Undisclosed; >4x its 2024 valuation | Creaegis; Accel, Section 32, Prosus | Enterprise AI agents |
| Ando | Pre-seed / seed | $20M | Not disclosed | Accel; Index Ventures; Emergence Capital | Agent-native software |
What It Means: The tracked rounds total roughly $5.9 billion (CODEW-calculated from the disclosed figures above), and one convertible note, Nscale's, accounts for about 57% of it. Strip that out and the week looks like what it is: a broad run of $100–$400 million rounds, most of them tied to AI.
The CODEW Lens: One financing can distort a weekly total. Read the median round, not the headline sum.
2. AI Startup Funding: Compute, Data, and Agents
Source: TechCrunch / Verda / Crunchbase News
This week's AI financings sorted into three layers. At the bottom, compute: Nscale's notes and Verda's $189 million round, which its Helsinki-based team says will scale a full-stack AI cloud and raise compute capacity, lifting total funding above $450 million. In the middle, data: Snorkel AI's $350 million Series E and Micro1's reported $4 billion valuation, both selling training data and evaluation environments to frontier labs. At the top, agents: Ema's $77 million Series B for AI agents that automate HR, IT, and finance work, and Ando's $20 million for a team-messaging product built for humans and agents together.
The prior week set the same pattern at a larger size: Temporal raised $550 million at a $12.55 billion valuation for infrastructure that runs long-lived AI agents, Factory raised $200 million at $5 billion for AI software-development tools, and Arcee AI closed $150 million at a $1 billion pre-money valuation to build open-weight models enterprises can run themselves.
What It Means: Investors are funding the parts of the AI stack that sell to every model builder and every enterprise buyer, not just the model builders themselves. That is a bet on durable demand for picks and shovels, and it reflects how much of AI's cost now sits outside the model.
The CODEW Lens: The AI funding story has moved from "who builds the best model" to "who supplies everyone building one."
3. Mega-Rounds & Valuations: Four Deals Read Closely
Source: Crunchbase News / TechCrunch / Reuters
Nscale — $3.36B convertible notes. The British AI neocloud raised the notes before a planned NYSE listing (ticker NSCL) that could value it at up to $35 billion, with Third Point leading and Nvidia committing $1 billion in November. Its previous priced round was a $2 billion Series C in March at $14.6 billion, per Crunchbase. Its IPO filing discloses statements of work of up to $43.8 billion with Microsoft and $44.6 billion with Anthropic, subject to delivery and financing conditions. The reporting reviewed did not detail the use of proceeds beyond funding that contracted buildout. This is debt-like capital, not a conventional venture round, and it should be compared that way.
Island — $400M Series F at $6.4B. The Dallas-based enterprise-browser and secure-workspace company more than doubled its 2024 valuation, in a round led by Evolution Equity Partners. Use of proceeds was not detailed in the reporting reviewed.
Snorkel AI — $350M Series E. Insight Partners and S32 led, valuing the seven-year-old company at $3.2 billion per Crunchbase (Reuters reported $3.5 billion, so treat the exact figure as reported). Crunchbase says Snorkel now exceeds $375 million in annual recurring revenue, which implies a valuation of roughly 8.5 times ARR (CODEW-calculated), a more useful quality signal than the round size.
Verda — $189M, Series B plus additional investment. Emergence Capital led an oversubscribed round with MUFG Innovation Partners, Supermicro, and Finnish pension insurer Varma participating, making Verda Europe's newest AI-cloud unicorn. The stated use: accelerate product development across compute capacity and platform services for frontier AI teams.
What It Means: Size alone says little. Island's valuation doubled, Snorkel's is anchored to real revenue, Verda's is backed by hardware and pension money, and Nscale's rests on contracted demand from two customers. Those are four different kinds of evidence, and only some of them are revenue.
The CODEW Lens: Ask what each valuation is standing on: revenue, contracts, or momentum. They are not equally sturdy.
4. Venture Capital Activity: Specialists, Crossovers, and Strategics
Source: Crunchbase News
Investor behavior was as informative as the deals. Evolution Equity Partners led both $400 million security rounds, Island and Cyera, in the same week, which marks a specialist conviction bet on enterprise security rather than a generalist chasing AI. Third Point led Nscale's notes, a crossover-style move that fits a company heading to public markets. Nvidia's $1 billion commitment to Nscale is strategic: a chip supplier financing a customer of its own hardware. Supermicro's participation in Verda follows the same logic.
Repeat backers also stood out. Emergence Capital appeared in Ando, Verda, and Arcee AI within about two weeks, and Insight Partners led or co-led Snorkel AI and Numeral in the same week. Underneath it all, Crunchbase counts at least 114 Series A rounds of $100 million or more globally in 2026, collectively about $33 billion, with over 70% going to AI companies, and over 40% of all seed and Series A dollars now flowing into rounds of $100 million or more.
What It Means: Capital is concentrating in fewer, larger checks from investors with a clear thesis, and strategic money from chipmakers and hardware vendors is now a routine part of infrastructure financings. Early-stage rounds increasingly look like growth rounds.
The CODEW Lens: When the supplier funds the customer, part of the demand signal is circular. Read strategic rounds with that in mind.
5. Emerging Funding Themes
Source: CODEW Analysis
AI infrastructure and neoclouds. Nscale, Verda, Temporal, and Crusoe's $3 billion-plus raise (made official the prior week) show compute and orchestration still drawing the largest checks.
The agent trust stack. Island and Cyera describe their products as securing work done by both people and AI agents, and Ando and Ema build the agents themselves. Security and agent software are being financed together.
The AI data supply chain. Snorkel and Micro1 sell human and synthetic training data to frontier labs. Together they turn data preparation into its own venture category.
Owned models. Arcee AI's open-weight pitch, aimed at organizations that want to run models inside their own environments, shows a counter-thesis to renting frontier APIs.
Physical and deep tech. Hubble Network, Precision Neuroscience, Enveda, and Impulse Space show capital is not exclusively flowing to AI software.
What It Means: The themes interlock. Every dollar spent on agents raises the value of the security, data, and compute beneath them, which is why these categories are being funded in the same weeks.
The CODEW Lens: Themes that fund each other are more durable than themes that merely share a buzzword.
6. What the Funding Signals
Source: Crunchbase News / CODEW Analysis
Investor appetite is intact and selective. Crunchbase estimates AI took roughly 77% of all venture and growth capital in the first half of 2026, about $394 billion. Appetite is strong for companies tied to AI's cost centers and weaker for anything that cannot show that link.
Concentration is the risk. The median late-stage round has roughly doubled since 2020 to about $100 million, while Crunchbase notes that deal counts remain below prior peaks. More money is going to fewer companies, and the largest private valuations now approach $1 trillion; Crunchbase recorded Anthropic's Series H at a $965 billion post-money valuation earlier this year.
Infrastructure demand is contractual, not speculative. Nscale's statements of work with Microsoft and Anthropic show buyers committing capital years ahead. Valuation step-ups follow: Island more than doubled, and Ema's valuation rose more than fourfold since 2024, per TechCrunch.
Enterprise adoption is the underlying bet. Security, data, and agent-software rounds only make sense if enterprises keep moving agents into production, and the competition for strategic technology is now visible in who writes the checks: chipmakers, hardware vendors, sovereign-linked funds, and pensions.
What It Means: The funding market is pricing in a long AI buildout, and pricing it unevenly. The winners of the next quarter's fundraising are likely to be companies that can point to contracts or recurring revenue, not just a category.
The CODEW Lens: This cycle rewards proof of demand. Momentum alone raises less than it did a year ago.
7. Companies to Watch
- Nscale: Its NYSE listing will test whether public investors will price an AI neocloud whose demand is concentrated in a few contracts.
- Snorkel AI: More than $375 million in ARR makes it one of the rare data-supply startups with revenue to match its valuation.
- Cyera: A repeat mega-round in about a year suggests data-security discovery is becoming a prerequisite for agent rollouts.
- Island: A valuation that more than doubled since 2024 makes it a bellwether for enterprise-browser and secure-workspace demand.
- Verda: Europe's newest AI-cloud unicorn is a test of regional, sovereignty-minded compute alternatives.
- Micro1: A reported $4 billion valuation for AI training data, if confirmed, would put it among the most highly valued data-supply startups.
- Ando: A $20 million bet that team messaging gets rebuilt around AI agents instead of humans.
- Arcee AI: Open-weight models for enterprises that want to own their AI stack, funded at a $1 billion pre-money valuation.
Sources
→ Crunchbase News — The Week's 10 Biggest Funding Rounds: Cybersecurity, AI and Health Take the Lead (Sept. 25)
→ Crunchbase News — Large Rounds for AI Infrastructure, Space Tech and Investment Management Lead (Sept. 18)
→ Crunchbase News — Jumbo-Sized Series A Rounds Are on the Rise
→ TechCrunch — Ahead of U.S. IPO, Nscale secures $3.36B in convertible financing
→ TechCrunch — Ema raises $77M as AI starts eating into enterprise software and services
→ TechCrunch — Ando wants to take on Slack with a messaging app for humans and agents
→ Verda — Verda raises $189M to accelerate its full-stack AI cloud
→ Fortune — Arcee AI trained four models for $20 million. Now it's worth $1 billion
→ Forbes — Micro1 raise at a reported $4 billion valuation
→ Reuters — Snorkel AI valued at $3.5 billion
→ Wall Street Journal — Cyera raises $400 million in extension round
The CODEW Stat
13 tracked rounds · ~$5.9B disclosed · 57% in one convertible note · 70%+ of $100M+ Series A rounds to AI The tracked rounds (CODEW-calculated from disclosed figures; reported amounts included at face value) show a market that is well funded and increasingly concentrated. Excluding Nscale's notes, the remaining twelve rounds average roughly $210 million, with security, AI data, and agent software leading. The year-to-date Crunchbase figure that more than 70% of $100 million-plus Series A rounds went to AI companies confirms the same skew at the earliest stage. The takeaway for founders and investors is the same: capital is available, but it is flowing to companies that sit on AI's cost and trust layers and can show demand for them.
Editorial Note
Startup Funding Watch tracks where startup capital is moving — funding rounds, valuations, venture capital activity, strategic investments, and the funding themes that reveal where the technology market is heading.
Educational content only. Not investment or business advice. Analysis is based on company announcements and reporting from Crunchbase News, TechCrunch, Reuters, the Wall Street Journal, Fortune, and Forbes cited above. Funding amounts and valuations are as reported by those sources; where sources differ (for example, Snorkel AI's valuation) the discrepancy is noted. Metrics referenced are labeled as reported, calculated, or CODEW-derived. Some products referenced may be affiliate partners — see our Affiliate Disclosure for full details. Coverage, data sources, and methodologies can change as the intelligence platform evolves.
Reviewed by Erwin Castro
on
Monday, September 28, 2026
Rating:
No comments: