Nicholas Melillo
Founding Partner
Strategy & operations
// About
An AI venture studio that keeps what it builds.
Cobalt Glacier conceives, builds, and operates AI-native B2B SaaS brands. Seven of them so far, all originated in-house. We don't spin them out and we don't hand them off. We run them.
// Why we exist
AI collapsed the cost of starting a software company, and almost every studio used that leverage to spin out faster. We used it to build companies worth keeping.
// Our thesis
Six things we believe in, and bet on.
01
Software is the best business model ever invented.
Near-zero marginal cost. Recurring revenue. Network effects. There has never been a better asset class to own, provided you actually own it long enough for the math to work.
02
We build from zero.
Every brand we run, we started. No auctions, no broker lists, no diligence on someone else's mess. Starting a company is now cheap enough and fast enough that originating a brand beats bidding for one, and it means we own the codebase, the data, and the roadmap from line one.
03
Operators beat investors.
We're not allocators dabbling in ops. We run brands. Every decision we make is informed by what it's like to ship product, hire engineers, and answer support tickets at 11pm.
04
AI compresses the cost of a new brand.
A product that took a team of fifteen and eighteen months to get to first revenue now takes a handful of people and a fraction of the time. That collapse in cost is the entire reason a studio makes sense in 2026 when it did not in 2016. We use the leverage to widen margins, not to subsidize a race to the bottom.
05
Distribution is the only durable moat left.
Code is commodity, and AI made it more so. Brand, channel, proprietary data, and trust are not. The brands in our portfolio win because customers actively choose them, not because the tech is impossible to copy.
06
Studio, not accelerator.
We don't take a batch and hope. We build the company ourselves: the idea, the first version, the first customers, the team. When we bring in an outside operator, they co-found a specific brand with a specific thesis, not a cohort seat and a demo day.
// How we're different
A studio, but not the kind that spins out.
Every model bends toward the incentive of its clock. Studios optimize for the spin-out, accelerators for demo day, funds for the vintage. We build the same way a studio does and then keep the company, which changes almost every decision that follows.
// them
Traditional venture studio
- Horizon
- Incubated inside, then spun out to raise
- Exit posture
- The studio's return comes from the spin-out and the eventual sale. Once the company raises institutionally, the clock starts.
The workshop is real, but the business model is exit-shaped. Studio equity gets diluted round after round, and the incentive is to hand the company off, not to run it for a decade.
// them
Accelerator
- Horizon
- You bring the company, they bring a batch
- Exit posture
- Demo day. The accelerator's job ends when the cohort does.
Twelve weeks of programming, a check, and a network. Nobody at the accelerator writes the code, owns the roadmap, or picks up the pager. Useful, but it is not building.
// them
Venture Capital
- Horizon
- 10-year fund, 10x or zero
- Exit posture
- IPO or bust. Failed exits are written off; middling ones are pushed to sell.
Growth at all costs to justify the next mark. Cycles force exits regardless of whether the business is ready. Founders churn through boards, not customers.
// us
Cobalt Glacier
- Horizon
- We conceive it, build it, staff it, and keep running it.
- Exit posture
- No spin-out and no fund clock. Sale, recap, or public listing when it clearly beats continued ownership, decided per brand, with the operator, against underwritten alternatives.
We build companies the way a studio does and then keep them the way an owner does. Shared platform underneath, independent brands on top, and operator economics that roll through whatever outcome a brand eventually earns.
// Leadership
Operators, not allocators.
// Operating Partners
We're hand-picking a small group of senior operators to build a brand with us.
If you know a market cold and want to co-found the product that fixes it, the studio brings the design, engineering, infrastructure, and distribution. You bring the domain. This is a build seat, not a quarterly advisory check-in.
// Where we fit
Studio, fund, accelerator, or roll-up?
The distinctions that actually change how a brand gets built and who ends up running it.
// For investors & long-duration allocators
If studio economics are your mandate, let's talk.
Cobalt Glacier isn't raising outside capital today. When we do, a small priority list of family offices, endowments, and RIAs hears first. Request the data room for per-brand economics, studio cost per launch, and the framework we use to decide what gets built, under NDA. Or schedule a 30-minute intro with a partner.
- · Written thesis brief and portfolio one-pager on day one
- · Read-only data room by end of week two
- · Named partner as your single point of contact