Newsletter · 2026-09-13 · Ashutosh Agarwal

Francis Collins Says All of Us Funding Was Cut More Than 70% as CDMOs Reshore - Life-Science Tools Recovery - Week of September 13, 2026

Life-Science Tools Recovery for the week of September 6 to 13, 2026. Podcast synthesis on former NIH Director Francis Collins quantifying a more than 70% cut to the All of Us sequencing program and what it means for research-tools demand, plus CDMO executives at Siegfried and SK pharmteco on US reshoring and rising consumables intensity.

Life-Science Tools Recovery

Week of September 13, 2026: Francis Collins Says All of Us Funding Was Cut More Than 70% as CDMOs Reshore

Weekly Recap, Sunday, September 13, 2026

TL;DR

• The loudest voice this week was the most authoritative one this newsletter has had on research funding: Francis Collins. The man who ran the Human Genome Project, launched the giant "All of Us" study, and led the National Institutes of Health (NIH) for 12 years said the All of Us budget "in the last two years has been cut by more than 70%," and described "unprecedented cuts in support, firing of staff" across NIH, with young scientists getting demoralized and U.S. science leadership "very much in doubt, with China coming up very fast." Academic and government labs are big buyers of our companies' instruments and reagents, so this is a real, quantified demand headwind, and it comes from a source no one can wave away.
• Two contract-manufacturing operators added useful side color: US reshoring is real, and drug-making is getting more data-hungry and more consumables-heavy. Executives at Swiss CDMO Siegfried and Korea's SK pharmteco described buying US plants, squeezing more data out of every experiment, and the mounting appetite for specialty materials (amino acids, solvents, purification media) as drugs get more complex. Encouraging texture for the picks-and-shovels suppliers, though not an order read on any listed name.
• The one number that would settle the sector debate is still outstanding. Whether the bioprocessing order book is turning remains unanswered, and it likely stays that way until Q3 earnings season begins in late October.

What's new

Ranked by what actually moves a portfolio: first the most credible voice imaginable on the research-funding squeeze, then operator color from the contract-manufacturing world.

1. Francis Collins puts a hard number on the funding squeeze, and he built the very projects being cut.

The single most valuable thing on the podcasts this week came from an unlikely place: a reflective, career-spanning interview with Dr. Francis Collins on The Other 80 (September 9). Collins is not a pundit and not a company operator. He is the person who ran the Human Genome Project, launched the "All of Us" precision-health study, and led the NIH for 12 years under three presidents. When he describes the state of American research funding, he is describing machinery he personally built. Treat him as authoritative on the funding environment (the demand base our whole sector sells into) rather than on any single company's orders.

The headline, in his own words:

"The budget for all of us in the last two years has been cut by more than 70%. So the enterprise now is kind of just trying to keep itself afloat instead of building on this with all the next generation things that could be done." (Dr. Francis Collins)

Why that specific project matters to us: "All of Us" is one of the largest genome-sequencing programs on earth. Collins said it has amassed roughly 600,000 complete genome sequences out of a target near a million people, with "tens of thousands of those researchers every day" using the data. That is a direct, high-volume customer for sequencing machines and consumables, exactly the kind of program that keeps Illumina, PacBio, 10x Genomics and their reagent suppliers busy. A budget cut of more than 70% to a flagship buyer is not an abstract policy worry; it is fewer machines bought and fewer reagents consumed.

He was just as blunt about NIH more broadly:

"In the last year and a half, there has been unprecedented cuts in support, firing of staff, deviation of decision-making from scientific priority to political expediency, young scientists becoming increasingly demoralized about whether they have a chance for their own career to flourish and blossom. And our leadership in the world, which was unquestioned, is now very much in doubt, with China coming up very fast." (Dr. Francis Collins)

And, on the customer base itself: "the universities are all now in a pretty bad place." Universities and government labs are among the largest buyers of research instruments and consumables, so "in a pretty bad place" is a plain-English description of a shrinking end market.

Two things to file away. First, this reinforces, from a far more authoritative seat, the same thread flagged last week, when an MD Anderson oncologist relayed the American Society of Clinical Oncology's warning of a proposed ~12% / ~$5 billion NIH cut for fiscal 2027. Collins is not quoting a proposed budget; he is describing cuts that have already landed on the projects he ran. The direction is unchanged; the credibility just went up sharply. Second, as an almost throwaway aside, Collins offered a clean history of the sequencing cost curve that underpins this entire sector: a genome that "initially was $400 million" fell "to the point where you could sequence a genome, maybe for less than a thousand. Now it's down to maybe 200." That $200 figure is the same "couple hundred dollars" the PacBio CEO cited last week for short-read machines, a useful independent confirmation that the cost war is real and roughly where operators say it is.

2. Contract-manufacturing operators: US reshoring is happening, and drug-making is getting more data- and consumables-hungry.

The operator texture this week came from the contract-manufacturing (CDMO) world: the businesses that make drugs on behalf of pharma companies, and therefore buy a lot of the same equipment and consumables our universe sells.

On Molecule to Market (September 11), Stefan Bucardt, Chief Scientific Officer at Swiss CDMO Siegfried, described a US-expansion move: Siegfried closed its acquisition of Noramco "beginning of May," picking up an early-stage R&D site in Athens, Georgia, a large-scale plant in Wilmington, Delaware ("150 cubic meter of asset"), and a starting-materials site in Tasmania. His reasoning went straight to reshoring: "US domestic manufacturing becoming more and more important." That is one more data point in the slow drumbeat of pharma supply chains moving back onshore, a mild tailwind for anyone who sells equipment into US plants.

The more interesting bit was where he sees manufacturing heading. Drug developers, he said, have compressed clinical timelines "by 2 to 3 maybe even 4 years" over the past decade, which leaves far less time to nail down the manufacturing process. His answer is more data, faster:

"You need to be able to... run your experiments more efficiently to get as many data out of an experiment as possible in order to build models more quickly and be able to scale reliably in a much faster timeline." (Stefan Bucardt, CSO, Siegfried)

In plain terms: squeeze more measurements out of every batch, model it, and scale it faster. That is the demand backdrop under the analytical-instruments and process-monitoring suppliers in our universe. He also noted molecules are getting "much larger" and "more complex to synthesize" than 20 years ago, and that Siegfried is "doubling down on flow chemistry" because that was the single biggest technology gap causing it to turn down projects. None of this is tradable directly, but it all points the same way: manufacturing is getting more instrumented and more data-driven.

Separately, on BioSpace's "Denatured" podcast (September 10), Arvind Singh, Global Head of Supply Chain at Korea's SK pharmteco (SK Formteco), walked through what modern drug-making actually consumes. (Note: this episode was sponsored by SK Formteco, so treat it as informed operator color, not independent research.) His point: as drugs shift from simple pills to peptides and cell therapies, they eat far more specialty materials. A peptide, he said, needs "specialized amino acids or amino acid derivatives... resins, solvents, purification" media and coupling reagents, and "can consume large volume of high purity solvents." He gave a vivid yield example: a peptide with "30 steps of synthesis" at 98% success per step still ends up around "55%" final yield, so raw-material quality matters enormously. For cell therapies, "the patient is effectively the supply chain." Why it matters: this is a direct description of rising demand for the specialty consumables, reagents, purification media and cold-chain logistics that names like Avantor, Repligen, Waters and Maravai sell into. It is a "the consumption curve is intact" signal, not an "orders are back" signal, so keep the distinction sharp.

3. China stayed a drug-innovation story, never an instrument-buyer story.

China came up again this week, on The Prof G Pod (September 8, "China Decode... China's Biotech Surge") and Newcomer Pod (September 10, with investor Zavain Dar), and, in an aside, from Collins himself, who warned U.S. leadership is "very much in doubt, with China coming up very fast." But every mention framed China the same way it has been framed all summer: as a rising rival in drug discovery, clinical-trial speed and AI, never as a customer buying Western instruments. This is pundit and investor commentary, clearly labeled as such. The single biggest swing factor for our group, whether Chinese labs are actually purchasing Western machines, remains unanswered.

The debate

Every week we steel-man both sides. This week the bear got the sharper evidence; the bull got only indirect encouragement.

The bull (bioprocessing recovery / sequencing and cell-therapy re-acceleration). The bull's case this week is indirect and structural: the CDMO operators describe manufacturing getting more data-hungry (Siegfried) and more consumables-heavy as modalities grow complex (SK pharmteco), and US reshoring adds a slow tailwind for equipment sold into American plants. That is the same "technology direction is intact" argument made last week: real, but not the same thing as demand turning. The bull's core problem is unchanged and, if anything, sharper: there is still no bioprocessing order read from a company we can actually own.

The bear (funding / China / tariffs / lumpy demand). The bear got the cleanest point of the week, and from the best possible source. Francis Collins, architect of the very sequencing programs our companies supply, said the All of Us budget is down more than 70% and NIH is suffering "unprecedented cuts," with a demoralized, thinning scientific workforce and universities "in a pretty bad place." This is not a proposed budget line; it is damage already done to a flagship customer, described by the person who built it. It stacks directly on top of last week's proposed ~12% / ~$5 billion NIH cut. The brain-drain concern (young scientists leaving research) is a slow, structural negative for the whole research-tools end market.

Our read. The honest scorecard tilts bearish this week, mostly on the strength and credibility of one voice. The bull's evidence this week is entirely indirect (CDMO trends, reshoring), while the bear's is concrete, quantified and authoritative (Collins on a 70%+ cut to a real, sequencing-heavy customer). That said, nothing here changes the underlying setup carried all summer: the long-run technology demand curve looks intact; the near-term order book is dark; and the funding headwind is real and now better-documented than ever. The difference versus last week is simply the mix: last week gave the bull a live operator and a hard price; this week handed the bear its most credible witness. The one question that would settle the whole sector, whether the bioprocessing book-to-bill is turning, is still open. Expect that to stay unresolved until Q3 earnings season begins in late October.

Stocks in play

Every ticker in the coverage universe. This week the signal is almost entirely the funding read-through (which hits every research-tools buyer) plus indirect CDMO color. Where a name has no company-specific signal, the standing debate carries forward.

Companies referenced this week that are not in the coverage universe: Siegfried and SK pharmteco (private/foreign CDMOs) and Noramco (acquired by Siegfried) are not investable here but map the demand direction for the listed consumables, reagents and analytical-tools suppliers above.

Read-throughs

• Academic / research funding: the signal of the week. The loudest bear thread of the summer got its most authoritative voice: Francis Collins, who built All of Us and led NIH, saying that program's budget is down more than 70% and NIH is suffering "unprecedented cuts," firings and a demoralized workforce. This is not a proposed cut; it is damage already landed on a flagship, sequencing-heavy customer, and it stacks on last week's proposed ~12% / ~$5 billion NIH cut. It points the same way for every name that sells into academic and government labs (Illumina, 10x, PacBio, Bruker, Agilent, Bio-Rad, Avantor, Thermo Fisher). The added twist this week: a separate report on This Week in Virology (September 12) that "hundreds of millions of dollars" of NIH infectious-disease research is being shifted to the Department of Defense, money leaving the civilian-research pool that our companies' customers draw on.
• Bioprocessing / advanced-therapy manufacturing (Sartorius, Repligen, Avantor, Maravai, Danaher/Cytiva). The book-to-bill question remains open. But the consumption read-through improved: CDMO operators describe drug-making getting more data-hungry (more measurements per batch) and more materials-heavy (specialty amino acids, solvents, resins, purification media, cold chain) as peptides and cell therapies scale. Structurally good for the consumables and analytics suppliers over time; not an order read. Net: long-run demand direction reinforced; near-term order book still dark.
• Sequencing / NGS competitive dynamics (Illumina vs Element / Ultima / MGI / PacBio / Oxford Nanopore). The new data point is Collins' independent confirmation that a genome now costs "maybe 200," consistent with the short-read economics the PacBio CEO laid out last week. The bigger read-through is on the demand side, not the competitive side: All of Us, one of the largest sequencing programs anywhere, is being starved of funding.
• US reshoring. A quiet but recurring positive: Siegfried explicitly bought US plants because "US domestic manufacturing becoming more and more important." Onshoring of pharma production is a slow, structural tailwind for equipment and consumables sold into American facilities, relevant across the analytical and bioprocessing names.
• China. Discussed as a drug-innovation and AI rival (Prof G, Newcomer, and Collins' aside), never as a buyer of Western instruments. Whether Chinese labs are actually purchasing Western machines remains the biggest open swing factor for the group.
• CGT reagent demand. The SK pharmteco commentary describes the manufacturing logistics of cell therapies (chain-of-identity, cold chain) but gives no demand figure, so cell-and-gene-therapy reagent volumes (relevant to Maravai) stay unconfirmed. Carry forward.

What changed vs last week

Last week (September 6) brought a new PacBio CEO on the record with hard sequencing economics, a quantified NIH-cut proposal, and operator texture on bioprocessing technology. This week the signal came from a single, high-credibility funding voice.

Concretely, what changed:

1. Sequencing color shifted from operator to corroboration. Last week the new PacBio CEO delivered the most substantive sequencing commentary since July. This week the continuity came from Collins independently citing a ~$200 genome, which matches last week's numbers: useful corroboration rather than new news.
2. The funding bear got its most authoritative witness. Last week the funding point came from an oncologist relaying ASCO's warning of a proposed ~$5 billion / 12% NIH cut. This week it came from Francis Collins describing cuts already made, including a 70%+ cut to All of Us, a program he founded. Same direction, materially higher credibility, and now anchored to a specific sequencing-heavy customer.
3. Bioprocessing color shifted from startups to established CDMO operators. Last week's texture came from a two-month-old startup (DNA barcoding) and an AI-bioreactor project. This week it came from real, sizeable CDMOs (Siegfried, SK pharmteco) on reshoring, data-hungry process development and specialty-materials demand. Still not investable directly, but a step up in the seniority of the voices.

Still open, and it's the important part: a bioprocessing book-to-bill from any listed supplier (Sartorius, Repligen, Danaher/Cytiva), a read on Chinese demand for Western instruments, and direct commentary from the four largest names, Thermo Fisher, Danaher, Agilent and Revvity. The scoreboard's funding column got heavier and better-sourced this week, but the one number that would actually settle the sector debate, the bioprocessing order book, likely waits for Q3 earnings season in late October.