I know this was an old post. Nonetheless...
Not everything with large quantities of hydrogen is explosive. E.g., many hydrocarbons, such as are in a piece of paper, are reasonably stable, unless you light them.
Rather than assuming the device would be a hydrogen bomb, nuclear or chemical, it is appropriate to do a scientific literature search:
They aren't talking pure hydrogen - they are talking about hydrogen atoms on silicon surfaces.
Hydrogen lithography (HL), the removal of hydrogen atoms (depassivation), on hydrogen-passivated silicon surfaces is becoming an important technique in next-generation device designs
However, I think we are still talking about single-layer-surface coding, which limits storage capacity. I imagine many kinds of volumetric coding might conceivably create higher storage densities.
Hope that clarifies the issue, though I admit the article is too technical for me to bother trying to read in full.
I wonder if, realistically speaking, any such technology, even with fairly good error-correction coding, would have a fairly limited lifetime, because it might be gradually degraded by various forms of radiation.
Nonetheless, the projected availability of such technologies introduces interesting privacy concerns - in that virtually everything you do might be recorded and stored. The articlie mentions quantum computing arrays
Together, HL and HR unlock an array of new possibilities including the creation of hundreds of precisely placed identical qubits for quantum computation
which might mean that all current codes will be broken - so assume that any encrypted stuff recorded now might be decrypted later.