Gate NVIDIA/Spark announcement on engineering supportability — eng-only until then, then hand to GTM

Situation

With NVIDIA engagement happening at very high levels over the next week, Peter decided to say nothing publicly and not pre-announce anything until CIQ has something it can support — something it would actually ship and point a customer at the support org for. Until then the Spark/Rocky-on-Spark work stays purely engineering (get it on Nathan, Wolford, Westley, and Peter own Sparks; validate it works), explicitly not a Bjorn/GTM item. Once it is supportable, GTM is unleashed.

Reasoning

This is Peter recurring production-readiness gate: engineering owns is-it-real-and-supportable; GTM owns tell-the-world, and the second cannot begin before the first is true. The discipline is sharper because NVIDIA is a top-level relationship — a premature claim there costs more than to a normal customer, so the bar is something we would ship and support, not it-boots-on-my-Spark. He separates internal momentum (validate broadly on the team Sparks) from external signaling (GTM) to keep energy alive internally without leaking outward prematurely.

Additional Context

Peter confirmed. Same gate as the 5/27 Computex production-ready condition, now applied to NVIDIA/Spark. Work just became possible (alive on the Spark via Max LiveUSB) but not yet reproducible/supportable.

Observed Evidence

Max DM 6/11: "I dont want to pre-announce anything at all... something we can support before we tell nvidia." "this is purely eng right now and not something for Bjorn." "so its about getting it to nathan. and me. and wolford. and westley." Bjorn DM 6/9: "its alive... not at all reproducible yet."

Matching Patterns

Confidence Breakdown

Reasoning Depth Analysis