It's not
just an analogy, and it contains the essence of the story, but it's also not the whole story.
In physics we talk about the "degrees of freedom" of a system -- this is just the count of all of the independent ways that it can move. For each degree of freedom of a system you can calculate the average energy in that degree of freedom. By the equipartition theorem, at thermal equilibrium, all the degrees of freedom will have the same average energy, which will be T (if you measure temperature in units of energy).
So if you think about dropping a bouncy ball in a tube and it bounces until it slowly comes to rest, it has these degrees of freedom -- the internal degrees of freedom of the atoms of the ball, the internal degrees of freedom of the atoms of the floor/tube -- and then two really obvious degrees of freedom, the center-of-mass position of the ball, which gains an energy scale due to the gravitational force, and the center-of-mass momentum of the ball, which trades energy with this position degree-of-freedom.
Statistical mechanics says that as this system progresses, the location of the energy will slowly become more uncertain until it is on-average-evenly distributed across all of the degrees of freedom. That's why it bounces lower and lower: there is so much energy in the two "main" degrees of freedom -- maybe half a joule? -- whereas in the vibrations there is something closer to 10^-21 J of energy at room temperature.
But the flip side of dissipation is always fluctuation -- this is in fact the subject of a major theorem! So the fact that this can randomly lose energy to these other degrees of freedom means that those degrees of freedom are also randomly kicking the ball. As you can imagine with ~20 orders of magnitude difference between the two, they don't kick this ball by all that much. But you have a lot of experience with a lot more tiny balls that are bouncing off the ground all the time. Take a deep breath. There they are.
If everything were to come to its minimum energy configuration, why are these air molecules so stubbornly not falling to the floor? Well, they are trying to! But they are so light that they are being kicked back upwards by these random thermal kicks, so high that they can in principle go the many kilometers to the uppermost atmosphere.
(Of course if they could go all that way in a single kick then air would have to be so non-interactive that we could not use it to talk to each other... the mean free path in air is actually about 68 nm, so in practice every air atom is getting its random thermal kicks from other nearby air atoms. But the ultimate origin of these random thermal kicks is the random kicks of the floor on the few hundred nanometers of air sitting above it, and that energy comes from the Sun and is mostly conserved as these atoms collide with each other -- but a tiny bit is often converted to little photons of infrared light that sometimes escape the atmosphere.)
With that said as others have noticed, the free-particle energy relation in special relativity is E = γ m c². Famously, at rest, this factor γ = 1/√(1 − (v/c)²) is 1 and the energy of a particle at rest is E = m c². But as v gets closer and closer to c, v → c, this energy grows without boundary, E → ∞. So there is no finite temperature where a kinetic degree of freedom would exceed the speed of light. Indeed you can solve for v, as 1/γ² = 1 − (v/c)². So the velocity corresponding to any given total energy is v = c √(1 − (mc²/E)²). For a rest particle with E = mc² this is v = 0 as you would expect; or when the kinetic energy first gets to mc² we would have E = 2mc² and thus v = c √(3/4) = 0.866 c.