Ramp data shows Jev gained a point of adoption in a month
Card spend across Ramp's customer base says a lot about who is trying Jev, and nothing about who kept it.
Ara Kharazian posted Ramp's top SaaS vendors list for October 2026 and says Jev captured an entire percentage point of market adoption within a month, growth he describes as unheard of. The figure comes out of Ramp's view of what its customers are paying for, so it is a spend signal, not a usage or quality one.
Jev captured an entire percentage point of market adoption within a month (unheard of growth)
On the downside, new and cheaper models are pushing AI spend down
Kharazian frames the number as one half of a split in AI spending. New and cheaper models are pushing spend down, he says, with Jev's jump and continued gains by routers on that side of the ledger. Pulling the other way is adoption of expensive image and video generation models, and he names Higgsfield and Runway as examples of vendors on the upside.
TypeSafe AI quoted the post with the words "fastest growing" and nothing else. That is the vendor amplifying a third party chart, not publishing anything new, and the chart is not theirs.
What a point of adoption actually measures
Ramp's data is corporate card and bill pay activity across its customer base. A percentage point of market adoption means more of those businesses put a Jev charge through in October than in September. It does not say how much they spent, whether the workloads stayed in production, or whether any of them got the accuracy they wanted. A team that signed up, ran a weekend evaluation and churned counts the same as a team that moved a pipeline over.
That matters because the public record on Jev this month has been mixed in exactly the places a spend chart cannot see. A red team reported a 43.5% attack success rate via prompt injection, and Theo's own $1,000 benchmark found Jev Router almost 5x slower on his workload. Trial volume and retention are different series.
For people building on it, treat this as evidence that Jev is being evaluated widely right now, which is useful if you are deciding whether your team will be alone on it. Treat it as nothing at all about whether the evaluations are going well. Kharazian published a single month of change on one dataset, and one month is not a trend.