Tokamaks and the reason they are still not efficient

Well, a large number of countries, after the break even ( actually 60% of output over input energy) in energy of the prototype tokamak in JET joined into creating ITER, a prototype Tokamak design designed to have output energy in megawats.

If interested you should go to the FAQ of the link given for ITER .

There exist alternate projects:

Of the "magnetic confinement concepts" for fusion (mainly tokamaks and stellarators) the main advantage of ITER and its tokamak technology is that for the time being, the tokamak concept is by far the most advanced toward producing fusion energy. It is consequently pragmatism that dictated the choice of the tokamak concept for ITER. Stellarators are inherently more complex than tokamaks (for example, optimized designs were not possible before the advent of supercomputers) but they may have advantages in reliability of operation. The W7-X Stellarator, presently under construction in Greifswald, Germany, will allow good benchmarking against the performance of comparable tokamaks. These results will be incorporated in decisions about how DEMO, the next-generation fusion device after ITER, will look.

The "inertial fusion concepts" are something quite different. These technologies have mainly been developed to simulate nuclear explosions and were not originally planned to produce fusion energy. The inertial fusion concept has not demonstrated so far that it offers a better or shorter path than magnetic confinement to energy production. In Europe, the Euratom Framework Programs do not fund research on inertial fusion, but the program maintains a "watching brief" on developments.

Efficiency in tokamaks rises with dimensions, and that is why ITER is much larger than JET.

The motivation for pursuing fusion is clear, but there are currently several main physics and engineering challenges:

• Confinement time: An operational reactor requires a long energy confinement time, $\tau_E$. An empirical scaling law for confinement time has been found to depend on the size of the tokamak as $\tau_E \propto R^{2.04} a^{1.04}$, where $R$ is the major radius and $a$ is the minor radius of the tokamak. Therefore bigger is better for confinement time. Can we create sufficiently long confinement times for energy production?
• Edge localized modes (ELMs): These are quasi-periodic disruptions due to the self-organization of the plasma. They release large heat loads onto the containment vessel but also diminish the build-up of impurities in the plasma. Can we control ELM production?
• Plasma facing component survival: There is a risk that the reactor will melt or erode with prolonged runtime. Can we design material that survives prolonged neutron bombardment and exposure to de-confined plasma?
• Tritium breeding: Tritium is one of the ingredients for fusion fuel of first-generation reactors. With a half-life of 12 years it does not survive very long in nature and must be freshly produced in the lab. Can we breed Tritium from fusion neutrons and Lithium in situ? Second-generation reactors will burn only Deuterium, a much more abundant isotope of Hydrogen than Tritium.
• Feedback Control: The Wright Brothers were the first to fly a fixed wing aircraft in part because they introduced a novel way to control their aircraft. Can we design clever control tools and algorithms that suppress and control plasma instabilities, enabling stable “flight” of the reactor?
• Cost: Can we build such a fusion reactor that competes economically with existing renewable and non-renewable power plants (taking the external cost of non-renewable energy into account, if necessary)?

A lot of progress has been made, but these current problems need to be overcome before tokamaks will be used as generators.

To get a lot of energy from a fusion reactor you need lots of D-T fusion events per second, and this means a combination of reasonably high density and very high temperature. This is extraordinarily difficult to achieve. In particular as you try and increase the plasma density it gets increasingly difficult to maintain the plasma in a stable state.

There have been studies of using other geometries for the magnetic fields, and these may be easier to use. See this article for details or Google for many similar articles. However it's early days and as far as I know no-one has actually used this type of geometry in a fusion reactor.