Startup Funding Watch · October 5, 2026
AI Megarounds, Infrastructure Capital and the New Venture Cycle — Where the Money Is Actually Going, and What It Reveals About the Next Generation of Important Startups
Venture Capital Is Moving Deeper Into the Stack — and Narrowing Who Can Raise
The headline numbers in this cycle are extraordinary and easy to misread. AI and infrastructure financing has reached a scale that would have seemed implausible two years ago, with major rounds across OpenAI, Anthropic, xAI, Databricks, and a widening set of infrastructure companies, according to Value Add Pulse tracking. But the more consequential story is not the size of the rounds — it is where inside the AI stack that capital is landing, and who is being excluded from it.
Three patterns define this week's funding activity. First, capital is moving deeper into infrastructure — from model companies toward the chips, memory, interconnects, data centers, networking and power that models require. Volantis raised $88 million specifically to connect AI chips to memory more efficiently, a textbook example of investors funding a named bottleneck rather than another general-purpose AI application, per Reuters. Second, security is becoming a second-order AI category — Armadin's $255.5 million Series B at a valuation above $2.5 billion shows investors are now treating AI-native cybersecurity as a standalone market, not a feature, per WSAU. Third, defense technology has moved into the venture mainstream, with European defense startup funding projected to quadruple to $10.5 billion in 2026, per Resilience Media.
But there is a countervailing story that most funding roundups ignore entirely. As mega-rounds concentrate among a shrinking set of perceived category leaders, founders outside the hottest AI categories are increasingly turning to revenue-funded growth and "seed-strapping" — smaller raises, longer runways, and deliberate avoidance of the dilution that comes with institutional capital, per Business Insider. The venture market is not loosening. It is bifurcating — and the gap between the two halves is widening.
Funding at a Glance
| Company | Amount | Sector | Valuation |
|---|---|---|---|
| Instinct (reported) | $1B | AI Agents | $10B (reported) |
| Armadin | $255.5M Series B | AI Cybersecurity | >$2.5B |
| EliseAI | $350M | Vertical AI | $4B |
| General Intuition | $220M | AI Reasoning | Undisclosed |
| PaleBlueDot AI | $200M | AI Infrastructure | Undisclosed |
| SiMa.ai | $150M Series C | Physical AI | ~$1.45B |
| Supabase | $150M | Developer Infrastructure | Undisclosed |
| Volantis | $88M | AI Memory Interconnect | Undisclosed |
| Metaview | $60M | AI Recruiting | Undisclosed |
Signal summary: The largest rounds cluster around three positions — AI infrastructure and compute, AI-native security, and vertical AI embedded in specific industries. Meanwhile, the mid-tier of the market is thinning: companies outside these categories are finding it harder to raise institutional capital at any size.
The Funding Stories That Matter
AI Continues to Dominate Venture Funding — But the Composition Is Changing Fast
What the data shows: AI remains the dominant destination for venture capital, and the concentration is accelerating. Mega-rounds are increasingly reserved for companies perceived as category leaders, while foundation-model companies absorb capital at a scale that has no historical precedent in venture. But the most instructive shift is internal to the AI category: capital is rotating out of generic application layers and into infrastructure, compute, and enabling technology.
What this financing tells us: The market has largely resolved the question of whether AI applications will be valuable — the open question is which layer of the AI stack captures durable margin. Investors appear to be betting that the answer is infrastructure, not applications, because infrastructure has defensible technical moats, long enterprise contracts, and demand that scales regardless of which application wins.
The concentration problem: As mega-rounds grow larger, they also grow fewer. A market where the top 20 AI companies absorb the majority of capital creates a structural problem for everyone else: the middle of the funding market — Series A and B rounds in the $10M–$50M range outside hot categories — is thinning.
Editorial question: Does capital concentration in AI infrastructure produce a healthier ecosystem, or does it starve the application layer that ultimately determines whether the infrastructure investment pays off?
AI Infrastructure Is Attracting Strategic Capital — and Targeting Named Bottlenecks
The pattern: Investors are no longer funding "AI infrastructure" as a broad theme. They are funding specific, identifiable bottlenecks — AI chips, memory and interconnects, data centers, AI cloud, networking, power, and compute infrastructure. The defining example: Volantis raised $88 million for technology that connects AI chips to memory chips more efficiently, per Reuters. This is not a platform play or a model play — it is a component play. And that is precisely why it attracted capital.
What this financing tells us: As AI models grow larger, the constraint shifts from raw compute to data movement. Moving data between memory and processors consumes disproportionate power and time. Any company that can reduce that cost has a defensible position, because the improvement compounds across every AI workload.
The strategic significance: Infrastructure investments like Volantis, SiMa.ai, and PaleBlueDot AI are the "picks and shovels" layer of the AI boom. They do not require consumer adoption or viral growth. They require becoming technically essential — and technical essentiality creates switching costs that application companies rarely achieve.
Editorial question: As AI infrastructure funding accelerates, does it create a durable layer of independent suppliers, or does it simply become an acquisition pipeline for hyperscalers and chipmakers?
AI Cybersecurity Is Becoming a Major Funding Category in Its Own Right
The pattern: A distinct funding category has emerged around protecting AI systems and using AI to defend them — AI-agent security, AI infrastructure security, autonomous threat detection, nonhuman identities, and AI governance. The defining example: Armadin raised a $255.5 million Series B at a valuation above $2.5 billion, per WSAU. A valuation of that scale for a security company would have been remarkable five years ago. Today it reflects a specific conviction: AI systems introduce an entirely new class of attack surface, and existing security tools were not designed for it.
What this financing tells us: Security is becoming a second-order AI investment category — a market that exists because AI exists, and that grows proportionally with AI adoption. This is structurally attractive to investors because demand is non-discretionary. Enterprises deploying AI agents cannot simply choose not to secure them.
The emerging sub-segment: Smaller companies are forming around specific niches. Rig Security raised $12 million specifically to secure AI agents, per Atom's funding tracker — a pure-play bet that agent security becomes its own market rather than a feature of broader platforms.
Editorial question: Does AI security consolidate around a few large platforms, or remain fragmented across specialized vendors addressing distinct threat surfaces?
Defense Technology Moves Further Into the Venture Mainstream
The pattern: Defense technology has shifted from a niche, ethically contested category into a mainstream venture market — autonomous defense, AI-enabled military systems, drones, counter-drone technology, intelligence software, and a fast-growing European defense startup ecosystem. The defining data point: European defense startup funding is projected to reach $10.5 billion in 2026 — roughly quadrupling, per Resilience Media.
What this financing tells us: Geopolitical priorities have created a major new venture category with characteristics that are structurally different from consumer or enterprise software. Procurement cycles are long, but contracts are large and sticky. Buyers are governments, which means revenue is not dependent on market timing. And demand is driven by policy rather than consumer preference — making it unusually durable.
The strategic shift: Defense technology is no longer a category that founders avoid or that institutional LPs resist. It has become a legitimate allocation, and European startups in particular are benefiting from a policy environment that actively encourages domestic defense capability.
Editorial question: Does defense technology become a permanent pillar of the venture market, or is the current surge a cyclical response to specific geopolitical conditions that could reverse?
Seed-Strapping and Revenue-Funded Growth: The Funding Story Nobody Is Covering
The pattern: Not every startup is chasing a $100M+ round — and increasingly, many are deliberately choosing not to. A growing cohort of founders is pursuing seed-strapping: raising small amounts of capital, funding growth from revenue, protecting founder control, extending runways, and avoiding the dilution that comes with larger institutional rounds. The defining data point: Reporting on Silicon Valley founders describes this as the "Bank of Best Friends" phenomenon — founders funding growth through personal networks and revenue rather than traditional VC, per Business Insider.
Why this matters: This is the most under-reported funding trend in the market, and it is arguably more important than the mega-rounds. If a meaningful share of high-quality founders is opting out of institutional capital entirely, then the venture industry's headline numbers overstate its actual grip on the next generation of companies. It also means the supply of fundable startups at Series A may be smaller than the capital available to deploy — a mismatch that could intensify competition for the companies that do choose to raise.
The founder calculus: A founder who raises $100M at a high valuation commits to a specific growth trajectory and a specific exit expectation. A founder who funds growth from revenue retains optionality — the ability to raise later on better terms, sell at a lower valuation if needed, or remain independent indefinitely. In a market where AI mega-rounds have made dilution expensive, that optionality has real value.
Editorial question: As mega-rounds concentrate capital in fewer companies, does seed-strapping become the default path for everyone else — and what does that mean for the venture industry's traditional middle?
Where Venture Capital Is Actually Flowing
The current funding cycle can be read as a shift in where in the stack capital is willing to go. Each successive layer is further from the end user and closer to a technical bottleneck — which is precisely why investors find it attractive.
| Layer | What Capital Funds | This Week's Example |
|---|---|---|
| Applications | End-user AI products | Metaview ($60M) |
| Vertical AI | AI embedded inside specific industries | EliseAI ($350M) |
| Agents & models | Autonomous systems and frontier capability | Instinct ($1B, reported), General Intuition ($220M) |
| Security | Protecting AI systems and agents | Armadin ($255.5M), Rig Security ($12M) |
| Developer infrastructure | Tools AI companies build on | Supabase ($150M) |
| Compute & silicon | Named hardware bottlenecks | Volantis ($88M), SiMa.ai ($150M), PaleBlueDot AI ($200M) |
The direction of travel is clear: capital is moving down the stack. Applications attract capital, but the largest and most strategic rounds are increasingly at the compute, silicon, and interconnect layers — where the technical moat is deepest, and the demand is least dependent on which application ultimately wins.
This has an uncomfortable implication for founders building at the application layer: the infrastructure they depend on is attracting capital faster than they are, which means their suppliers may capture more value than they do. That dynamic is not new in technology — but at this scale, in this cycle, it is unusually pronounced.
What the Funding Data Is Telling Us
- AI remains the dominant destination for venture capital. No other category comes close in either total dollars deployed or share of mega-rounds.
- Capital is moving deeper into the AI infrastructure stack. The largest strategic rounds are increasingly at the compute, memory, interconnect, and silicon layers rather than at the application layer.
- Security is becoming a major second-order AI investment category. Armadin's $2.5B+ valuation demonstrates that AI-native security is now treated as a standalone market, not a feature.
- Defense technology is becoming a much larger venture market. European defense funding projected at $10.5B in 2026 signals a structural shift, not a temporary spike.
- Mega-rounds are increasingly concentrated among perceived category leaders. Capital is consolidating into fewer companies, which narrows the field of who can raise institutional capital at scale.
- Founders outside the hottest AI categories are turning to revenue-funded growth. Seed-strapping and capital-efficient strategies are becoming the default path for companies that cannot or choose not to compete for mega-rounds.
- The next funding cycle may reward infrastructure, vertical AI, and bottleneck-solving companies. Generic AI wrappers face the hardest funding environment; companies solving specific, defensible technical or domain problems face the easiest.
What to Watch Next
- Infrastructure round sizes. Watch whether Volantis-scale component rounds ($80M–$100M) become routine, or whether the category consolidates into a smaller number of larger winners.
- AI security valuations. Monitor whether Armadin's $2.5B+ valuation holds as a benchmark, or whether it proves to be a peak that later rounds fail to justify.
- European defense funding. Track whether the projected $10.5B for 2026 is met, exceeded, or revised — and whether US defense startups see comparable acceleration.
- The Series A and B middle. Watch whether mid-stage funding outside hot AI categories recovers or continues to thin — the clearest signal of whether the market is genuinely healthy.
- Seed-strapping adoption. Look for data on how many high-quality founders are choosing not to raise institutional capital, and what that means for the supply of Series A-ready companies.
- Vertical AI durability. Monitor whether EliseAI's $4B valuation and similar vertical AI rounds produce follow-on financing, or whether the category consolidates around a few winners.
- Power and data center capital. Watch whether energy and power infrastructure becomes a distinct AI funding category, given that compute expansion is increasingly power-constrained.
Venture capital is not simply flowing into AI — it is flowing through AI and into the layers beneath it. The funding cycle is rewarding technical bottlenecks, defensive security categories, and policy-driven markets, while the traditional mid-stage funding market thins.
For founders: The funding question is no longer "Are we an AI company?" It is: What specific, defensible problem do we solve, and is it at a layer where capital is willing to go? Infrastructure bottlenecks, AI-native security, and vertical depth are attracting capital. Generic AI applications are not.
For investors: The most attractive positions in this cycle sit below the application layer. Compute, memory and interconnect, security, and developer infrastructure all benefit from AI adoption regardless of which application wins — and they carry technical moats that application companies struggle to build. But watch the concentration risk: capital is pooling in fewer companies, which narrows the field of future opportunities.
For enterprise buyers: The vendors funded today define your procurement options tomorrow. AI-native security companies like Armadin and Rig Security will shape how agents are governed in production. Infrastructure suppliers like Volantis and SiMa.ai will determine the cost and latency characteristics of the AI systems you deploy. And vertical AI platforms like EliseAI will increasingly displace general-purpose tools in regulated workflows.
For the venture industry: The rise of seed-strapping is a genuine threat to the traditional fundraising model. If the highest-quality founders increasingly choose revenue-funded growth over institutional capital, the industry's core assumption — that it has privileged access to the best companies — weakens. The firms that adapt will be the ones that offer value beyond capital.
The CODEW Research Flywheel
Startup Funding Watch is the capital-flow monitoring layer of The CODEW’s broader Startup Intelligence. Where the other products answer different questions:
| Product | Core Question |
|---|---|
| Startup Funding Watch | Who raised money and where capital is flowing |
| Startup Spotlight | Why an individual startup matters |
| Startup Database | Structured company-level intelligence |
| Company Profiles | Persistent company reference |
| The Term Sheet | Who gains strategic advantage from major financing |
| VC Intelligence | Broader venture-market structure and investor behavior |
Startup Funding Watch → Identify emerging funding patterns → Startup Database → Startup Spotlight → Company Profiles → VC Intelligence → The Term Sheet → AI Infrastructure Special Report → Semiconductor Intelligence → Cybersecurity Watch → Defense Technology Startups → Tech M&A Watch
The CODEW Stat
$10.5 billion — the projected total for European defense startup funding in 2026, roughly quadruple the prior level. It is the clearest single number illustrating how quickly a category can move from venture fringe to mainstream allocation when policy, geopolitics, and capital align. Defense technology was, until recently, a category many institutional investors would not touch. Today it is one of the fastest-growing venture markets in the world — and European startups, not American ones, are capturing the shift.
No comments:
Post a Comment