How AI moving into the lab gets read as compute demand

Built around news that a major model company has set up a physical lab of its own, this traces a 3-year progression that has moved from helping summarize papers, through research support, down to running experiments. The central thesis is not about drug discovery itself but about a chain of links: once profits are proven in a new industry, demand for compute and memory picks up again. However, this logic stands on a four-step assumption running from proof to rising usage and on to capacity expansion and back-end process volumes; the proof for the first step has not yet been presented, and the rest is filled in with expectations. The usage-surge data cited as quantitative evidence is also out of date, and the point that an index's rise coincides with the company's entry is closer to a coincidence of timing than to causation. It also sets out, on its own account, the opposite reading that a company heading for a stock listing may be inflating its growth story, and that is worth keeping in mind while reading.

This is a note: my own write-up after reading several sources. It has no source link; events checked against a source are kept as records.