Fusion
If hydrogen fusion (including deuterium and tritium) was easy, we wouldn't be here. Relatively small masses of hydrogen would have "burned" to iron very early in the formation of our galaxy; not enough time for our sun, our planet, and complex life to form.
"Cold Fusion" aka "Low Energy Nuclear Reactions" assumes that deuterium fusion is easy - that is, it can happen with electron-chemistry energy levels (electron-volts) and solid-matter atomic spacings (nanometers)...
... rather than the 300 KeV, 300 femtometer deuteron spacings encountered (for tens of nanoseconds) in 100 million Kelvin temperature fission-triggered fusion weapons, or the vastly slower (gigayear rates) proton-proton reactions that occur in the dense 15 million Kelvin temperature, 15 times-denser-than-lead plasma cores of gigameter-diameter stars like our Sun.
"Temperature" is a crude measure of particle energy and velocity. Thermal velocity in this case, particle accelerators and lasers can also cause high energy particles. For a hot gas, the thermal energy of the particle is 1.5kT (0.5kT in 3 dimensions).
k is the Boltzmann constant, 8.62E-5 eV/Kelvin. eV (electron-volts) is the appropriate energy unit for atoms, molecules, and chemistry ... but way too small for atomic nucleus interactions.
So, the energy per particle in a thermonuclear blast is 1.5*8.62E-5*1E8 eV = 13 KeV, and a head-on collision of two opposite-direction particles is 26 KeV.
For comparison, the temperature of a welding torch flame is around 3500 Kelvins, about 0.5 eV, and boiling water is 370 Kelvins, about 0.05 eV.
Large magnetic containment fusion power reactor proposals assume a mixture of naturally occuring deuterium and artificially-created tritium. Tritium has a 12 year half-life; it does not occur naturally. Just as well; in terawatt fusion-fuel quantities, tritiated water is a horrendous radiological health hazard. Tritium is currently a waste product of fission nuclear reactors, and while "burning it up" in a fusion power reactor might be a worthwhile improvement in fission safety, separating and purifying reactor-sourced tritium will be hazardous and less than 100% efficient. Tritium decays into helium 3 and a beta particle (electron).
Fusion reactions (if and when they happen) leak neutrons, which can be captured in a thick lithium blanket to produce tritium. However, neutron capture at best has a "gain" of 1.3, so another source of tritium may be needed.
Outside of a nucleus, the half-life of neutrons is around 15 minutes.
How much energy is needed to fuse two deuterons? ... sometimes ...
Two 10 KeV nucleons approaching exactly head-on will fuse, but if they are slightly mis-aimed, they will scatter. The aiming circle for fusion is measured as an area of 5 "barns", or an area of 500 square femtometers, or 5e-28 square meters. The scattering cross section is a thousand times larger.
By comparison, the cubic lattice spacing of a palladium crystal is 0.386 nanometers (3.86 Ångstroms), so the area of a lattice cell is 1.5e-19 square meters, 300 million times larger. The ratio of a 260 square meter tennis court one grain of sand.
Consider "undulations" in that palladium "tennis court", positive spikes of electric field near the palladium atomic nuclei, which would repel very slow "thermal" deuterium or tritium nuclei - those will tend towards the valleys between the palladium atomic spikes. But those are gradually curved VERY WIDE VALLEYS compared to the size of a palladium or deuterium or tritium nucleus. All of those nuclei will REPEL each other, not attract. If the nuclei are like needles sticking up, the chance that TWO deuterium needles will perch point down on the upwards pointy needles of palladium crystal atoms is VERY VERY VERY VERY VERY small. Like "probably never in the lifetime of the universe" small.
Such tiny probabilities are numerical, not "convenient hand-waving-false-analogy", which LENR advocates trot out because they cannot do the atomic modelling, and WANT TO BELIEVE. They put a lot of poorly measured energy into their experiments, measure statistically variant energy escaping their experiments, and call the variations "Low Energy Nuclear Reactions" ... instead of inevitable measurement noise.
Stanley Pons and Martin Fleischmann were world-class chemists in 1989, when they announced their "N Fusion" results. Few scientists were able to replicate anything like the Pons/Fleischmann results, and the ones who did usually found experimental errors, such as faulty sensors. If world class scientists fail to get it right, and succumb to life-long confirmation bias rather than education and critical self-education and improvement, what hope is there for "LENR Joe" in his garage?
"Atomic/chemical" scale is vastly larger, and vastly lower energy, than nuclear scale. As of this writing, the only way to create hydrogen/deuterium/tritium fusion at large scale is with a fission bomb. Even the huge ITER Tokamak consumes more input energy (magnetic field, support systems, and plasma heating) than nuclear fusion energy output.
Alternatives like Generation IV nuclear reactors, with continuous fuel recycling and long-life-isotope deactivation can be safer and smaller than a fusion reactor, perhaps someday deactivating stored fission "waste".
Wikipedia article, Nuclear Fusion
Graph of reaction rates. Note that at 10 keV temperatures, the reactivity of D-T is
A better alternative may be artificial alternate-genome "fuel plankton" grown in currently dead areas of the oceans. If we can create plankton genomes that cannot be exploited by viral phages in seawater, we can vastly increase primary productivity in the world's oceans, creating far more sea life and plankton-derived carbon fuel, while drawing down excess atmospheric CO₂. Reprocess CO₂ emissions, close the carbon loop, and fuel our existing fleet of carbon-fuel vehicles with "green fuel".