JPMorgan’s trading desk issued a note stating that the introduction of Google’s TurboQuant compression technology, which slashes memory requirements by at least sixfold, poses no immediate threat to AI chip consumption. The analysts predict any initial profit-taking in memory stocks will be temporary, with overall demand remaining robust.

TurboQuant can reduce LLM memory usage by a factor of six while accelerating inference speeds up to eightfold, potentially lowering training costs for hyperscalers and large language model operators. This breakthrough has raised questions over future memory hardware needs, prompting industry commentary.

The note invokes the 19th century Jevons Paradox theory, arguing that greater efficiency often drives higher consumption, not less. Following the announcement, memory and storage shares saw a modest pullback as some investors booked gains, though analysts maintain long-term demand will persist.