"Shall we play a game?"
In the wake of the North Korean nuclear test, there are some worrisome aspects that will come into play.
First, testing is not a friendly act in the post-Cold War era. It's one thing to walk into a bar with a concealed weapon. It's another thing to haul it out, fire a shot into the ceiling to verify that it's loaded, and yell, "WHO WANTS A PIECE OF THIS?" Li'l Kim has just done that, and now everyone is discreetly reaching for their hardware. (UPDATE: Apparently, Li'l Kim's pistol may have had a blank chambered. H/T to my partner in crime for finding this.)
Second, there is the question of "who's next?" Japan is a shake-and-bake nuclear power--one estimate I've read puts Japan having a credible nuclear deterrent within six months of the Japanese government giving the go-ahead. They have a breeder reactor that manufactures large amounts of plutonium (Reactor-bred "Ploot" is, by definition, bomb-grade), and they have a first-rate technical infrastructure; the rest is merely a matter of money, man-hours, and political will.
South Korea and Taiwan are not in that league--they do not have breeder reactors--but they have good technicians, and they can always whip out hard currency for what William Burroughs called "The Yard Sale at the End of History." So they could probably have nukes inside of two years.
Now, the second-order effects:
South Korea and Japan don't like each other overmuch. If one goes nuclear, the other will. If Japan goes nuclear, Taiwan will (assuming that they are not already practicing what I call "CCWMD"), simply because nobody in eastern Asia trusts the Japanese (the 1931-1945 period may seem to be ancient history to an upstart nation like the United States, but to the cultures of the Pacific Rim, it's just last week). The People's Republic of China will be extremely unhappy if Japan, South Korea, and Taiwan go nuclear.
Some sort-of good news: multilateral (as opposed to bilateral) nuclear stability has never been tested in an international crisis, only in RAND political-military games. Some really bad news: in those RAND games, multilateral nuclear stability seems to be a figment of everyone's imagination; it "fails deadly" about 95% of the time as players pursue their victory conditions vis-a-vis Country A without thinking about how Countries B, C, D, and E view their actions. It seems to be a natural human blind spot--because the same players, with their experiences from earlier games to inform and admonish them, did the same damn thing in subsequent games.
Moving from the purely geopolitical to the tactical end of nuclear warfighting ("It's all fun and games until somebody gets their eyes melted out!"):
One thing that Japan, the Koreas, and Taiwan would need is early warning (and, in that theater, it would be not-so-early warning due to the short distances times of flight involved) in order to maintain some minimal level of crisis/prompt launch stability (the old "use 'em or lose 'em" issue from the late Cold War). Early warning systems tend to be incredibly expensive; the best solution may be for these nations to buy access to our own warning net. (Pay-per-view launch detection: the new growth market for the United States Strategic Command.) We spent billions per year during the Cold War just to buy an extra 10-15 minutes of tactical warning time against the USSR. The same amount of money (after indexing for inflation) would give 30 seconds to 2 minutes additional in northeast Asia--which would just about enough time to execute one predetermined nuclear strike option on initial alert.
Because of the short timelines involved, a lot of "early warning" would be strategic (i.e., intelligence of possible intent to attack) as opposed to tactical (detection of missile launches). Strategic warning is ambiguous at best. With relatively small forces on all sides, generation stability (the sensitivity of force survivability to whether the force is on day-to-day alert or generated alert) assumes critical importance.
On the one hand, generated forces may be more survivable (additional mobile missiles on the move or in hide sites away from the main garrisons, additional command and control assets deployed, et cetera). On the other hand, generating forces to increase their survivability may convince the other side that you intend to launch a first strike.
One potential response to this concern is to design a missile force whose day-to-day posture is more or less indistinguishable from its generated posture. This, however, will tend to make even minor crises into nuclear ones, as there would be no way to know if the other side's nuclear forces are cocked and locked or not.
So, if America does nothing:
At some point in the not-too-distant future, Japan, South Korea, Taiwan, North Korea, and China will be in a multidirectional Mexican stand-off, most likely with relatively accurate nuclear-armed ballistic missiles. All sides would be able to deliver a signifcant strike from a day-to-day alert posture and would have next to no tactical warning available. The only thing missing would be the spark.
How does one say "Able Archer" in Japanese, Hangul, or Mandarin?
Inaction is failure; failure is not an option. We must lead.