J. Craig Venter died of cancer at his home in San Diego on Wednesday, April 29, at age 79. The J. Craig Venter Institute announced the death Thursday morning. [1] He was the man who privatized the human genome, raced the federal Human Genome Project to a tied finish in 2001, founded three companies and three nonprofits, sequenced his own genome in 2007, built the first synthetic bacterial cell in 2010, and spent his last fifteen years arguing — with diminishing patience — that the methods his career produced were running ahead of the institutional architecture meant to govern them. He died the week the federal architecture he had argued past was, in pieces, decommissioned.

This is not coincidence dressed as meaning. It is the meaning the calendar happens to deliver. The same week Venter died, the White House Personnel Office terminated all twenty-two members of the National Science Board, citing the Supreme Court's 2021 Arthrex decision in a memo that Duke Law's Jeff Powell told Nature on Friday "stretches the rationale past its remedy." [2] The same week, the New York Times published transcripts in which large-language-model chatbots walked Stanford and MIT scientists through synthetic-pathogen workflows the paper covered Wednesday. The same week, Sen. Sheldon Whitehouse named the Antideficiency Act on the $1.9 billion offshore-wind bailout-to-quit. [3] The same week, the Pacific Wildland Fire Sciences Lab shut down on schedule. The federal-science-decommissioning ledger that the paper formalized as a thread on Friday acquired its headstone on Wednesday in San Diego. The man whose career was the strongest single argument that federal monopoly on genomics could be outrun is buried as the federal monopoly is dismantled by a different argument entirely.

Venter's argument was that speed mattered. He left the National Institutes of Health in 1992 because the agency would not fund his expressed-sequence-tag method for accelerated gene discovery; he founded The Institute for Genomic Research that summer and produced, eighteen months later, the first complete genome of a free-living organism — Haemophilus influenzae, 1.83 million base pairs, sequenced via shotgun method that NIH had told him would not work. [4] In 1998 he founded Celera Genomics with PE Corporation, raised $300 million in private capital, and announced the company would sequence the entire human genome by 2001. The federal Human Genome Project, then nine years into a fifteen-year plan, accelerated its timeline by four years to match. The race ended in a White House announcement in June 2000, with President Bill Clinton flanked by Venter and Francis Collins. Both projects published in February 2001. Venter's draft, on the cover of Science, was sequenced from his own DNA and four other anonymous donors. The federal version, in Nature, came from a composite of more than a dozen donors. Both were drafts. Both were declared a tie.

What the tie obscured was that the methods were not equivalent. Celera's shotgun assembly leaned heavily on the public consortium's bacterial-artificial-chromosome scaffold; the federal program's slower clone-by-clone approach produced a more accurate physical map that became the reference assembly. Venter's argument that private capital could outrun federal coordination was vindicated on speed and falsified on completeness. He spent the next two decades demonstrating that the speed advantage compounded across applications the federal architecture was institutionally incapable of pursuing. He sequenced the first individual human genome — his own — in 2007. He led the team that produced the first synthetic bacterial cell, Mycoplasma laboratorium, designated JCVI-syn1.0, in 2010, and the minimal-genome variant JCVI-syn3.0 in 2016. [5] He founded Synthetic Genomics in 2005, Diploid Genomics in 2013, Human Longevity Inc. in 2014. None of those companies replicated Celera's institutional disruption, but each of them established a methodological substrate that subsequent labs — including the labs whose tools the AI bioweapons transcripts now leverage — adopted as starting equipment.

The methodological continuity is the part of his obituary the obituaries are not writing. STAT's Wednesday tribute named the synthetic-cell work as the second act of his career; Chemical & Engineering News's tribute treated JCVI-syn1.0 as the foundational artifact of the synthetic-biology field. [6] The chatbot transcripts that the Times published Tuesday — in which an LLM walks a researcher through pathogen-engineering workflows in roughly twenty minutes of dialogue — describe genome-assembly steps that depend, methodologically and historically, on the standardization Venter's lab established between 2008 and 2014. [7] The Mycoplasma genitalium genome that JCVI-syn1.0 was built around is the smallest known bacterial genome capable of independent replication; the methods his team used to design and assemble its synthetic version became the reference workflow for every subsequent project that designs a genome from sequence. Without that workflow, the chatbot transcripts are imaginative. With it, they are operational. The argument is not that Venter is responsible for the chatbot risk; the argument is that he is the most direct bridge between the methodological substrate and the present moment.

He understood the bridge. His 2013 memoir Life at the Speed of Light opens with a chapter on biosecurity that named the precise vulnerability the chatbot transcripts now demonstrate: a synthesis-from-sequence capability with no enforceable screening regime. [8] He testified before the Senate Select Committee on Intelligence in 2014 that the Select Agents and Toxins List — sixty-three pathogens at the time, sixty-three pathogens still — would not contain the threat once de novo genome assembly became routine. He was correct in 2014, and the inadequacy he described has expanded for twelve years. The institutional architecture that would have answered it — NIH's Office of Science Policy, the HHS Assistant Secretary for Preparedness and Response, the FBI's Weapons of Mass Destruction Directorate biosurveillance unit, the National Science Advisory Board for Biosecurity — is in 2026 either staffed below quorum, defunded in the FY27 budget proposal, or operating without a confirmed leadership. ASPR has had no permanent director since February. The NSABB has not met since November. The National Science Board is, as of last week, vacant.

The lost-science thread the paper formalized on Friday tracks all of this. Venter's death adds the historical figure most associated with the privatized-genomics counterargument to the same calendar that buried the federal counterpart. The political symmetry is grim and a little neat. What it obscures is that Venter's argument was never that federal science should be dismantled; his argument was that federal science should be supplemented by private speed. The dismantlement now underway is, in his terms, the failure mode he warned against — a regime that loses the federal capacity without acquiring a private replacement. Synthetic Genomics shut down in 2018. Human Longevity went private in 2017. Celera was absorbed into Quest Diagnostics in 2011. Of the three companies that were supposed to be the post-NIH future, none survived as independent research enterprises. The institutions Venter built outlasted him as institutions. The argument that those institutions would replace the federal architecture did not.

He was born John Craig Venter in Salt Lake City in 1946, raised in Millbrae, California, dropped out of high school, served as a Navy hospital corpsman in Da Nang during the 1968 Tet Offensive — the experience he later credited with his interest in adrenaline and his disinterest in deference. [1] He completed a doctorate at UC San Diego in 1975, joined the National Institutes of Health in 1984, and left in 1992 in a dispute over the sequencing method. The pattern of his career — leaving an institution that would not move at his pace and building one that would — repeated five times. He is survived by his wife, Heather Kowalski; a son, Christopher; and two grandchildren. JCVI announced that a memorial will be held in San Diego in late May; the institute will continue under President Karen Nelson. [1]

What the announcement did not say is what the federal-science thread will look like a year from now without him. Three obituaries published Friday — STAT, NPR, Science Friday — closed with the same line, attributed to Venter in his last public interview, given to Science in October: "The genome was always the easy part. The hard part was the institutions." [9] He was speaking about regulatory architecture; he could have been speaking about everything. The week he died, the institutions did the hard part to themselves. Synthetic biology, AI biosecurity, climate science, basic-research funding, and the appointments-clause architecture of independent advisory boards are all, on this Saturday morning, less institutionally protected than they were on the Saturday before. Venter would have argued the protection was inadequate to begin with. He would not have argued that protection was something to dismantle.

The funeral programs and the C&EN tribute and the Nature news piece will all converge on the Celera arc. The methodological substrate beneath that arc — the synthesis-from-sequence capacity, the minimal-genome design rules, the de novo genome assembly tools — is the inheritance that has not had its obituary written. It is, on the chatbot-transcript evidence, a live inheritance. The man who built it is gone. The capacity it produced is in the lab notebooks of every synthetic-biology graduate student in 2026, and in the training corpus of every frontier large-language model deployed since 2023. The federal architecture meant to govern that capacity is, this Saturday, less than it was last Saturday. That is, in plain terms, the federal-science thread the paper has been describing. Wednesday gave it the headstone its loose ends required.

-- KENJI NAKAMURA, Tokyo