EDIT: Regarding trustworthiness of NL figures, see EDIT after the figure. [end EDIT]
The Netherlands seem rather advanced in actual IPv6 usage of the pool, with IPv6 traffic share currently at about 40% for an instance I recently added there, vs. just 26% IPv6 adoption rate according to Google (though there is some skew in favor of IPv6 in this case as I enabled IPv6 full throttle well before enabling IPv4 as well, and then ramped up the netspeed setting slowly only for IPv4 to probe the conditions in this zone I didn’t have prior first-hand experience with). Maybe @marco.davids had a hand in promoting IPv6 NTP in the Netherlands despite the pool-side imbalance? 
The NL zone seems to be extremely well served in general, with a bitrate/packet rate of typically less than 400 kbit/s / 500 pps for both IPv6 and IPv4 at full throttle being the lowest by some margin among the country zones I have/had servers in so far. The best so far had been Germany, with typically less than about 1.3Mbit/s / 1900 pps with IPv6 and IPv4 both at full throttle, and after being in the pool for some time.
EDIT: I need to revisit my numbers for the Netherlands. They are too good to be true. Triggered by some other oddities, I now realize that the provider seems to employ some sort of DDoS protection, and likely for IPv4 only. It started with three monitors on the production system not getting through anymore, which I initially put down to some routing/connectivity issues. But now with the server added to the beta system with its much higher number of monitors, even more of them fail. It looks suspiciously as if a portion of the IPv4 NTP traffic might fall prey to the DDoS system, noticeably impacting the absolute number for IPv4 NTP, thus potentially and significantly skewing the IPv6/IPv4 ratio in favor of IPv6. Need to dig deeper, e.g., contact the provider. Also, for a “reality check”, I’d be interested to hear from other people with servers in the NL zone as to how much traffic they see, for both IPv6 and IPv4 in absolute numbers, as well as what their derived IPv6 share is. [end EDIT[
Neither is anything in comparison to, e.g., some Asian zones. E.g., SG with sustained bursts easily exceeding 10Mbit/s at a 12Mbit setting for IPv4 (IPv6 essentially negligible in comparison, even at full throttle netspeed setting). I faintly remember South Korea being similar, or even higher.
Which still is nothing compared to what has been reported for China, where a 512kbit netspeed setting, the lowest officially* available, can apparently easily result in double- or even triple-digit Mbit/s actual traffic. Which would overwhelm any basic server not only in the Asia/Pacific region, where not only bandwidth, but especially traffic volume is generally more expensive than, e.g., in North America or Europe. It is interesting to see those differences in traffic costs based on cloud service provider offers. There is an older explanation by Cloudflare as to the why, but a web search will also yield more recent examples and explanations.
E.g., I had to take an instance out of the MY zone recently as both its bandwidth limit as well as monthly traffic quota were easily exceeded even at a 1kbit netspeed setting for both IPv6 and IPv4. Malaysia as such seems to have a relatively high IPv6 adoption rate, but there are only 3 or 4 IPv6 servers in the zone. Which results in them being overloaded much of the busier time of the day. (I’ve seen short sustained peaks of above 3Mbit/s, which artificial bandwidth limits aside is ok for typical cloud infrastructure, but the corresponding packet rate may be challenging, e.g., for typical consumer Internet routers.) I.e., the IPv6 servers are constantly phasing out of and into the pool so that there often is only a single server in the zone, or the pool even falls back to the Asia continent zone because there is no active server in the country zone. And obviously, when a server is the only one active in a zone, the netspeed setting is essentially irrelevant…
I was surprised, though, to see that a 12 Mbit IPv4 netspeed setting in India yields an average of slightly above 2Mbit/s only (and almost unnoticeable traffic at a 1kbit netspeed setting), given that some older threads in this forum referred to the IN zone as being underserverd as well (but maybe I misunderstood/misremember). Unfortunately (and strangely), none of my instances in India comes/came with IPv6. Would have been interesting to see what the IPv6 traffic share is in India, given the generally high IPv6 adoption rate there.
Typically less than 200kbit/s / 200 pps for IPv6 full throttle in Japan, with an IPv6 adoption rate of about 55% (according to Google), but I don’t have IPv4 there, so can’t say what the IPv6/IPv4 relation is, nor what the NTP traffic situation is overall. I got 800GB monthly traffic quota (in+out) for pretty much the lowest-end VPS conceivable there, but at just a few Euro more per year, I could easily get (and have) upper single- or double-digit TB traffic volume, or even unmetered, in Europe or North America.
* I.e., from among the values offered by the server management page, without resorting to low-level access to pool internal mechanisms.
