Originally Answered: DrAnders said. no more hydrogen being produced?
It's essentially correct that no more hydrogen is produced. There might be tiny amounts produced when high-energy collisions cause a nucleus to fission, but the primary trend is for lighter nuclei to be fused into heavier nuclei in the cores of stars.
But there's still plenty of hydrogen. The interstellar medium, from which stars are formed, has a composition mostly determined by the primordial outcome of the Big Bang (75% hydrogen, 25% helium by mass), with some heavier elements that were produced in stars and released by supernova explosions.
Star formation in the interstellar medium tends to occur when the gas is compressed by a shock wave -- either a density wave in a spiral galaxy, or the shock from a large explosion such as a supernova.
I don't think the universe will run out of hydrogen, but other things will happen that decrease star formation over time:
1) Most stars do not become supernovae. Lower-mass stars become white dwarfs, and their mass is normally locked up forever. Over time, mass is transferred from the interstellar medium to stars.
2) As the interstellar medium loses mass, it will become less dense, and therefore star formation will become increasingly rare. Lower star formation will produce fewer supernovae, which means fewer shock waves, which means even less star formation.
Over the long term, the abundance of hydrogen in the universe will decrease. I suspect it won't decrease very much, but I don't know what the "final" abundance is predicted to be.
There's an interesting paper on this subject called "The Galactic Millennium", by Paul Hodge of the University of Washington:
A galactic year is the time it takes the sun to revolve around the center of the galaxy -- roughly 225 million years. Hodge talks about what the universe will be like when it is 1000 galactic years old (i.e., 225 billion years old, as compared to the current age of 14 billion). To make numbers simpler, Hodge talks about the universe 100 billion years from now rather than 225 billion.
Here is part of Hodge's paper:
"Throughout the early parts of the Galactic millennium new stars were born, many of which were formed from gas and dust that was recycled from evolved previous generations. But this process cannot go on forever, as white dwarfs, neutron stars, and black holes are sinks that trap material in forms that cannot be recovered. Eventually the Galaxy will be as devoid of material for forming stars as the elliptical galaxies appear to be now. ... Perhaps, however, as the gas density decreases further, the rate will decrease or maybe even stop abruptly when the gas density drops everywhere below a critical density. ... In A.D. 100 billion star formation probably will have ceased, and there will be no Orion Nebula to light up the Galactic arms (and, anyway, the arms themselves may have faded into the feeble background disk)."
Hodge's paper is fairly short and non-technical, so you might want to take a look at his predictions for the distant future.
I'll add something in response to Richard R's answer. I didn't mention the expansion of the universe here because I don't think it's relevant. The usual statement by cosmologists is that expansion affects the universe as a whole but doesn't affect things on the scale of individual galaxies. To the extent that this is true, the expansion of the universe can be ignored when you're talking about star formation in galaxies. Perhaps the expansion will eventually affect the galaxies (such as in the "big rip" model), but star formation will probably come to an end long before then for the reasons mentioned earlier.
DrAnders, in contradiction with Paul Hodge's article, doesn't think that white dwarfs and neutron stars are a typical "end state" of matter. This is a debatable point. It's true that white dwarfs will become type Ia supernovae if they obtain enough additional matter to exceed the Chandrasekhar limit (about 1.4 solar masses), but this nearly always happens because of accretion of mass from a close binary companion.
Except for the close-binary-companion case, however, white dwarfs will remain white dwarfs forever unless they have a chance collision with another object. Such collisions are incredibly rare, and will not happen unless two objects are headed almost exactly towards each other. Therefore, I suspect that Hodge is correct, and that much matter will be locked up in the three end states he mentions -- white dwarfs, neutron stars, and black holes. On the other hand, you might argue that even if collisions are very rare, they will happen if you consider events over an incredibly long time scale. On yet another hand, however, over such enormous time scales, perhaps the universal expansion will affect galactic dynamics and therefore reduce the chance of collisions. Finally, if the "big rip" model is correct, the end state is entirely different.
In short, we're getting into the realm of considerable speculation and uncertainty. Also, note that Hodge mentions that he expects other astronomers to have some disagreements about his view of the distant future. Nobody has a definitive answer to this question.
For your original question, however, things are simpler. Hydrogen and helium were created in the big bang. Today, we have a bit less hydrogen than there was then, because some of it has fused into heavier elements in stellar cores and supernova explosions. No new hydrogen is being created, but neither have we lost much since the big bang. (This tells you something about the enormous energy in nuclear fusion. Despite the huge amounts of energy released by stars over the last 14 billion years, the loss of hydrogen has been very small. Our oil crisis on earth is playing out over the scale of a couple of centuries, but 14 billion years of stellar energy have barely made a dent in the hydrogen supply.)
DrAnders -- I don't necessarily disagree with you, I just don't know. You're right that there are drag forces that work against white dwarfs in the very long run. But it's also possible that whatever makes the universe expand and even accelerate will eventually affect the internal structure of galaxies. Both of these would take a very long time to have much effect. There's plenty of time in eternity, but the question is which effect is faster. Given our lack of understanding of most of the universe (in the forms of dark matter and dark energy), we're in a poor position to predict the distant future. Hodge's article might be a pretty good description of 100 billion years from now, because that's only an eight-fold extrapolation from the present time. We know that lots of double stars have survived for billions of years, so 100 billion years is not too much of a stretch. At some point (a trillion years? a quadrillion years?) we lack the knowledge to extrapolate. We learned about the acceleration of the expansion only 10 years ago, and I'm sure there will be other surprises in store.
(By the way, it wasn't me who gave your answer a thumbs-down. Whoever did that was unjustified.)
I thought I was done with this answer, but I just read an article in the current Scientific American (March 2008) called "The End of Cosmology", and the authors make predictions about the universe in the far future (and their prediction contradicts my own guess for the future abundance of hydrogen). Here is some information from the article:
a few minutes after the big bang:
2% heavier elements
1 trillion years in the future:
20% heavier elements
100 trillion years in the future:
last star burns out
So there's your long-range forecast for the next 100,000,000,000,000 years -- a time long enough to astonish even astronomers. There's something they don't mention: Whatever matter is locked up in neutron stars or black holes can't be associated with atomic elements, but their article doesn't indicate the fraction of such matter. It does, however, say that eventually (beyond 100 trillion years), the galaxy will collapse into a black hole; so in this, the authors' view of the future is similar to that of DrAnders.
The uncertainty in such predictions can be shown by mentioning the "big rip" hypothesis, which predicts that the entire universe will be torn apart in about 50 billion years. That's a very different scenario for the future universe, and gets back to the overall theme of my answer: We just don't know. I'm very confident that the earth and universe will survive 2012, but am much less certain about 50 billion years out.