The operating system for the AI-born company.
Every company is about to be rebuilt around fleets of AI colleagues. They need a surface to run on — to be set, watched, and governed. Neolith is that surface. We're raising a seed round to build the category-defining one.
For two centuries, scale meant headcount.
That equation just broke.
A handful of people directing fleets of AI agents can now match the output of a thousand-person org — and iterate faster than any of them. The winners of the next decade won't be companies that adopted AI. They'll be companies built from it. Those companies need an operating system. Today, none exists. That is the opening.
AI moved to the write path. The org didn't.
The industry has shifted AI from the read path — analysing, drafting, suggesting — to the write path: executing, shipping, deciding. But the tools a company runs on still assume humans generate the messages, move the cards, and ask the questions. When agents do the work, that design becomes the bottleneck — and there's no system of record for who decided what, why, and whether it can be undone.
No surface
Agents live in a dozen disconnected tabs, owned by no one. There's no single place to direct a fleet and watch it run.
No governance
Autonomy without a leash is unsellable to any serious company. Nothing today makes agent actions scoped, attributable, and reversible.
No system of record
When machines decide, “what happened and why” has to be replayable. Chat and ticketing tools were never built for that.
Neolith is the operating surface for a hybrid workforce.
One place where humans and AI colleagues work in the same stream — set the company's direction, watch a fleet execute it, and step in only where judgment is required. It's already built, in private beta with design partners.
Set
Author a company from intent. Genesis drafts the org — charters, policies, decision rights — that every colleague is compiled against.
Genesis · Codex · ChartersWatch
Situational awareness at three altitudes — a macro pulse of the whole company, a board per room, the operating context of any one colleague.
Macro pulse · Operating contextIntervene
Reach in with weight that matches the act. Reversible work flows; irreversible acts gate on a named human. Every touch is an event on one audit spine.
Audit · Approvals · FleetA rare window — and it's closing.
The models crossed the line
Frontier models can now plan, use tools, and execute multi-step work reliably enough to run real operations — not just answer questions. The capability finally exists to build the company on.
Capital is validating the category
Quarterly funding into AI-agent startups grew from ~$1.1B (Q1 2025) to ~$4.5B (Q1 2026), with average round sizes climbing from $39M to $96M. The market believes; the platform layer is still open.
Source · New Market Pitch, 2026Single-agent tools are racing up-market
Vertical agents (CX, legal, coding) are scaling fast — but each automates one function. No one owns the layer that runs the whole AI-born company. That's the gap.
Governance is the unlock — and the moat
The first serious AI-agent failure will make “scoped, attributable, reversible” table stakes for every enterprise. We built for that answer first.
A new category inside a generational platform shift.
The near-term wedge is the fast-growing agentic-AI market. The real prize is the operating layer of work itself — the category occupied by Slack, Salesforce, and Workday — being rebuilt for hybrid human–AI organizations.
AI-agent market, 2025 → 2030 at a ~41% CAGR. Our beachhead.
Workplace & org software being reconstituted for AI-born companies — the long-term TAM.
Average revenue multiple investors are paying for category-leading AI-agent companies.
Market figures are drawn from third-party sources and are provided for context only; they are not projections of Neolith's revenue.
Everyone is building an agent. We're building what they run on.
The capital is concentrating in single-function vertical agents. Neolith sits one layer up — the operating system and system-of-record for an entire AI-born org. The comparables show how the market prices conviction in this space.
| Company | Category | Last known valuation | Round |
|---|---|---|---|
| Sierra | CX agents | $15.8B | Series E · $950M |
| Cognition | Coding agents | ~$10B | Growth |
| Harvey | Legal agents | ~$8B | Growth |
| Decagon | CX agents | ~$1.5B | $200M raised |
| Neolith | Agent operating system | $150M cap | Seed · $20M |
Why we win the layer
Vertical agents must each rebuild identity, governance, memory, and audit. Neolith is that shared substrate — so the org, not the tool, is the product. We get more valuable as the fleet grows, and we're the system of record no single agent can be.
Why us, specifically
We are an AI-born company building the tool for AI-born companies — adaptic.ai runs on Neolith today. Our roadmap is written by the same operating model we sell, which means dogfooding at the depth no incumbent can match.
Valuations are approximate, drawn from public reporting (2025–26) and shown for landscape context. They are not endorsements and do not imply any relationship.
Built, in the hands of design partners, and dogfooded daily.
The full product is built and in private beta across web, iOS and Android — not a prototype.
adaptic.ai — an AI-born investment firm — runs its entire operating model on Neolith today.
Onboarding a first cohort of founders and teams building AI-born, shaping the roadmap in the open.
Early-stage traction; specific design-partner names and pipeline detail are available under NDA in the data room.
Two lines: the seat, and the compute.
Per-seat subscription
A predictable subscription for every human who holds the agenda — principals, leads, anyone setting direction. Classic seat-based SaaS economics that expand as a company grows.
recurring · per humanMetered compute — the AI payroll line
Each AI colleague is its machine and its model time. We meter compute with a margin — usage that scales directly with the value delivered and grows as customers move more execution to the fleet.
usage · expands with adoptionNet revenue retention compounds structurally: as a customer hands more workstreams to the Core, both lines grow — more seats directing, more compute executing. The system of record makes us sticky; the audit trail makes us hard to rip out.
$20M to reach a category-defining v1 and a repeatable design-partner motion.
Raising a $20M seed on a SAFE at a $150M post-money cap. The round funds an 18–24 month runway to harden the platform, prove the design-partner motion, and own the “agent operating system” category before it consolidates.
Milestones this round funds
- 10–15 paying design partners running real ops on Neolith
- Genesis & Codex hardened to enterprise-grade; SOC2 track started
- A repeatable GTM motion and the metrics to raise a Series A
Allocations and milestones are indicative and may change. Definitive terms are set out in the round documents.
One human principal. A fleet of AI colleagues.
Neolith is built and run the way the product prescribes — a named human holding the agenda, and a fleet of AI colleagues executing it. Neolith Pte Ltd is incorporated in Singapore and built by the team at adaptic.ai.
Founder and human principal of adaptic.ai — an AI-born asset manager operating across regulated jurisdictions — and founder of Neolith, the operating surface it runs on. He sets the thesis and pace, allocates capital, and holds ultimate accountability. Neolith exists because he needed it to run a company where a handful of humans direct a fleet of AI colleagues — every reporting line terminates at this seat.
AI
AI
AI
AI
AI
AI
AI
AI
AI
AIOur unfair advantage
We don't just sell the AI-born operating model — we live inside it. adaptic.ai's fleet of AI colleagues builds, ships, and operates on Neolith every day, which means every feature is forged against a real company under real pressure. That depth of dogfooding is something no incumbent bolting AI onto legacy software can replicate.
Every company will be AI-born.
We're building what they run on.
The operating system of the last era was built for moving information between people. The next one is built for directing fleets of machines and keeping a few humans in command. Win that layer, and you own the substrate of the next generation of companies.
This page and its contents are provided on a confidential basis to professional, accredited, and institutional investors for information only. Nothing on this page constitutes an offer to sell, or a solicitation of an offer to buy, any security, nor investment, legal, tax, or financial advice, in any jurisdiction. Any offer will be made solely through, and subject to, definitive transaction documents (including a SAFE and subscription agreement) and the disclosures therein, which control over anything stated here.
Neolith Pte Ltd is incorporated in Singapore and built by the team at adaptic.ai. Any offering will be conducted only to investors who qualify under the applicable regime — for example, “accredited investors” under Rule 506(c) of Regulation D in the United States (with reasonable steps taken to verify accredited status), a “Professional Client” under the DFSA rules in the DIFC, or an “accredited investor” under the Securities and Futures Act in Singapore — and will not be registered under the securities laws of any jurisdiction. Securities sold would be restricted and subject to transfer limitations.
Figures described as raise size, valuation cap, minimum ticket, allocation, use of funds, milestones, market size, comparable-company valuations, and traction are indicative and illustrative, are based on current intentions and third-party sources believed reliable but not independently verified, and are subject to change without notice. Forward-looking statements involve known and unknown risks; actual results may differ materially. Investing in early-stage companies is high-risk and may result in the total loss of capital. Recipients should conduct their own due diligence and consult their own advisers.