Raise model, summer 2027

Anchor high because the profile justifies it and because the investors are already asking. The floor holds even if nothing new is counted and growth merely stays at 47%, which makes walking away credible. The stretch needs two consecutive quarters of acceleration and a separately priced AI line, both of which are achievable on this timeline but neither of which is done.

Evidence now supports the top of this band rather than the middle. The band itself is not extended, because above 20x requires the AI revenue line. Originally revised upward for one structural reason rather than any metric: with Windows attaching at 50% of new business and a strong MSP motion, the Apple-only comparison stops applying and the comparable set becomes cross-platform IT operations. That removes the cheapest comp from the analysis. A down round requires pricing below 7.5x on $114M, or 6.3x on $135M. Nothing in the comparable set prices below 12.9x except Jamf at 3x, and Jamf grows at 12% to 16%. At 47% growth with 120% retention, that comparison does not apply.

The difference between about 15x and about 19x. Real money, and no longer existential, which is the important part. The round does not depend on the new platform, so it can be scoped for whether it works rather than for how it presents.

Understated ARR is good news. ARR reporting that omits a signed platform partnership and an entire channel is a finance-systems problem, and a quality of earnings review finds it immediately. The danger is not the number, it is what an investor concludes about rigour everywhere else once they find it, because that conclusion gets applied to retention, attach, margin and the pipeline too. It converts upside into a discount and adds weeks to a process.

This is the most urgent item on the page. One reconciled ARR definition, every channel inside it, restated history so the growth rate is calculated on a consistent basis, signed off by finance, well before the first meeting. Done properly it raises both the base and the growth rate. Done late it becomes the reason the round takes an extra quarter.

The only unmet gate above is separately identifiable AI revenue, and the existing code product is the one thing that can produce it before summer 2027. It is live, it has a published price, and even $1M to $2M of ARR attributed to it converts the platform claim from a diagram into a revenue line. Folding it into a larger unreleased platform removes the only asset capable of closing that gap. Build the graph, keep the meter running on the product that already has one.

Apple management with Windows attached is a product. The same thing sold direct, through MSPs, and embedded inside another platform's offering is infrastructure, and infrastructure is what carries a 20x multiple. This is a better and more defensible version of the platform claim than anything in the proposed R&D operating system, and it is already true rather than planned.

NinjaOne is endpoint and remote monitoring at its core. Endpoint management across Apple and Windows, plus EDR, plus vulnerability management, plus compliance automation, plus identity is a wider consolidation claim than theirs, sold through the same channels. That is the argument for pricing near a comparable that is 4.4x larger: the stack collapses more of the buyer's spend. It only holds once identity carries real customers, which is why turning it on before the raise matters more than it appears to.

Deel negotiated at arm's length with full visibility into alternatives and settled above the direct price. That is market evidence that direct pricing has headroom. A pricing uplift on renewal lifts ARR, retention and gross margin with no product work at all, and it is the only lever on this page that compounds into the base before a summer 2027 raise. On $114M, a modest uplift is worth more to the valuation than a year of building an unreleased platform, because it raises the number the multiple multiplies.

It already exists, it already has a published price, and a modest revenue line from it proves the graph generalizes commercially without requiring a second go-to-market motion built from nothing. That is a far better use of it than absorbing it into an unreleased platform aimed at a buyer the company cannot reach.

Nothing resembling a six-stage operating system produces revenue inside nine months, and it cannot produce it inside four. That is not a judgement on the thesis, which reads as broadly correct about where software development is going, and the context-graph architecture is genuinely the same shape as the context model this company already shipped for IT. It is a statement about sequencing.

The raise stands on the core business: multi-channel distribution, 85% margin, 120% retention, and acceleration. The platform is the answer to what gets built with the proceeds, which is a question every growth investor asks and most companies answer badly. Presented that way it strengthens the round instead of having to carry it, and the vision survives intact rather than being cut down to fit a deck.

What to do in the meantime: build the graph narrowly with a handful of design partners, and keep the existing code product metered so the engineering-surface claim has a revenue line against it however small. That is enough to make the next-phase story credible without diverting the capacity the four base levers need.