Part 01

Industry, academia, and the ones who vanish

Doctoral destinations one year out, and ten years out. Half go to industry, a quarter to academia, and the academic share does not fall at graduation. It falls over the decade afterwards.

Doctoral graduates 2000–2023 · measured one year after the degree year

Where a computing doctorate lands is usually discussed as a two-way split between industry and academia. It is at least a four-way split, and the fourth category, people the US labor market simply cannot see, is larger than government and research institutes combined.

Destination throughout is the position held one year after the degree year, so that doctoral internships and pre-doctoral employment do not get counted as placements.

The academic share does not collapse. It erodes, slowly, and later.

One year after the doctorate, the academic share is roughly flat across two decades of cohorts. The change shows up further out. Of the 2000–2009 cohorts, 25% were still in academia ten years on; of the 2015–2019 cohorts, 19%. The exit is not happening at graduation. It is happening in the years after, and it is happening faster for each new generation of graduates.

The pay gap is not subtle

At the same career stage the industry–academia gap is roughly a factor of two and a half, and it falls straight out of the employer table below: the large technology firms cluster near or above $190,000, the universities near $70,000. Pay here is the career panel’s estimate, so read the ratios rather than the dollar amounts.

Who actually hires them

The concentration is remarkable: a handful of firms absorb a large share of an entire national research pipeline.