Newsletter · 2026-08-23 · Ashutosh Agarwal

The Funding Hole Widens as China Poaches US Biotech - Life-Science Tools Recovery - Week of August 23, 2026

The academic-funding leak in life-science tools got a hard number for the week of August 16 to 23, 2026, as the NIH All of Us program said it has lost nearly 80 percent of its base funding since 2023 while China ramps R&D spending to poach US scientists, venture money and IP.

Life-Science Tools Recovery

Week of August 23, 2026: The Funding Hole Widens as China Poaches US Biotech

TL;DR

The academic-funding leak we have been tracking for weeks finally got a hard, program-level number, and it's ugly. The head of the NIH's flagship precision-medicine project, All of Us, said on a podcast that once you strip out one-time top-ups, the program has lost nearly 80% of the funding it used to have since 2023. Real-world consequence: it has stopped enrolling as many people and stopped collecting as many DNA samples. That is not an abstract budget line, it is a direct cut to the demand for sequencing, sample-prep kits and lab consumables that names like Illumina, 10x Genomics, PacBio and Bio-Rad sell into research programs. This is the most concrete piece of the "fewer funded labs = fewer customers" story we've had all summer.

China showed up on the podcasts for the first time in weeks, but not as an instrument buyer. As a poacher. A California life-sciences advocate laid out, in plain terms, how China is spending $500 billion a year on R&D and dangling 20 years of stable funding to pull American scientists, venture money and intellectual property across the Pacific, while U.S. NIH funding "has not come back" and California plans to cap its research tax credits after 2030. The near-term read is a slow bleed of the domestic research base that buys Western lab equipment; the longer-term read is that a bigger and bigger slice of instrument demand ends up inside China, where Western tools names face tariffs and homegrown rivals.

What's new

The dominant theme this week was money: who is funding science, who isn't, and where the science (and the equipment orders that follow it) is going to end up. Below are the five items with real signal, ranked by how much a portfolio manager should care, each labeled by who is speaking and how much weight they deserve.

1. The clearest, most quantified demand hit of the summer, from the NIH program that literally buys sequencing. The clearest demand signal of the summer came on Mendelspod Podcast, "All of Us Comes of Age. And So Does Its Funding: Josh Denny on the Next Phase of Precision Medicine" (August 18, 2026). The guest, Josh Denny, who runs the NIH's All of Us Research Program (a massive government effort to build a research database of a million-plus Americans' DNA and health data), walked through what the funding squeeze has actually done to his program:

"when we go just to base appropriations, we will have lost nearly 80% of the funding that we used to have... as the funding has [successively] decreased since the 2023 year... we've had to make a number of cuts. It's mostly, most directly influenced the number of participants we've enrolled and the amount of data we can generate."

He put a physical number on it: the program has 880,000 people enrolled but DNA and other biospecimens from only about 600,000, and the gap, he said, "is largely... a money gap." There are willing volunteers; there just isn't cash to collect and process their samples.

Why it matters, and why it's different from every other funding headline we've had. Most of the NIH story so far has been about appropriations, top-line dollars in a bill. This is the first time we've heard the funding squeeze translated into fewer samples physically collected and sequenced. That is exactly the demand our sequencing and sample-prep names live on. All of Us is a real, recurring buyer of DNA sequencing (historically Illumina's core turf), genotyping arrays, and the collection tubes, reagents and consumables that Bio-Rad, 10x Genomics, PacBio and the broader research-tools group sell. When Denny says enrollment and biospecimen collection are being cut for lack of money, he is describing a purchase order that isn't being written. Label it precisely: Denny is an NIH program leader, not a tools-company executive, and he named no vendor and gave no instrument-spend figure. He even flagged some hope: the program is "very bipartisan," and one-time 21st Century Cures Act money plus ongoing talks in Congress could cushion the base-appropriations cliff. But this is the most direct, quantified line we've had between the Washington funding fight and an actual reduction in demand for the gear this sector makes.

2. China appears as a competitor for talent and capital, not as a customer. On Health UnaBASHEd: BIO 2026 China, Clinical Trials, and California's Moment of Reckoning (August 21, 2026), broadcast on Healthcare NOW Radio from the BIO industry convention in San Diego, Sam Torbati, an advocate for California's life-sciences industry, laid out a stark competitiveness picture. His central worry:

"as the federal government vacillates on NIH funding, SBIR, STTR. These are signals to our biggest competitor, China, to approach our academic institutions... and offer them sometimes 20 years of funding, stable funding. Our scientists and researchers can continue to do their research here in California, but they will be funded and financed by China. And obviously that intellectual property now belongs to China."

He stacked up the imbalance bluntly: China has "committed $500 billion a year in R&D," while California plans, "post 2030... [to be] divesting in biotech by capping R&D tax credits." He said venture-capital firms, even storied Bay Area ones, are "moving more money and investing in Chinese based biotech startups than here in California," because China "can get through the clinical trials process in under five years" versus "at least eight years" in the U.S., against a backdrop where a single American drug takes "12 to 15 years, $2.6 billion." And the recurring refrain: U.S. "NIH funding... which has not come back" is the thing that "de-risks" academic research, and without it, capital pulls back.

Why it matters for our names. This is a two-speed read-through, and both speeds are important. Near term, it reinforces item 1: a shrinking, capital-starved U.S. academic and early-biotech base means fewer domestic labs writing purchase orders for instruments and consumables. Longer term, it describes demand physically relocating to China, where Western tools companies (Thermo Fisher, Danaher, Agilent, Illumina, Waters, Mettler-Toledo, Revvity) do sell, but into a market shaped by tariffs, "buy-local" pressure, and increasingly capable domestic rivals. Label it carefully: Torbati is an industry advocate making a policy argument at a convention, not a tools operator, and he said nothing about instrument orders. Treat it as sentiment and structural backdrop, not a demand data point. But it is the most substantive China discussion we've had in weeks, and its direction is unhelpful for the U.S. research-instrument demand base.

3. Last week's "politicize the grants" worry got a sequel, with a surprising twist toward "maybe this part is fine." Seven days ago, a science-policy journalist warned about a White House push to let political appointees, not scientists, steer NIH funding. This week, on This Week in Virology, "Ebola Virus Brains and Contagious Catfish Cancers" (August 23, 2026), three working scientists (Rich Ebright, Vincent Racaniello and Alan Dove) dug into a specific, related proposal: the NIH wants to stop giving grant applications precise numerical scores and instead sort them into three buckets, "good, fair, and poor."

The fear is obvious, and they named it: if a huge pile of grants all land in the "good" bucket, they can't all be funded, and, in Racaniello's framing, "the political appointee could come in" and pick the ones that fit an ideology. They called it, as many do, "cover for further politicizing science."

Here's the twist, and it's why this is a nuance rather than a fresh negative. Ebright, who has actually sat on grant-review panels, argued the mechanics of the change aren't crazy: the system "is much more like NSF" (the National Science Foundation, which already works this way); grants are "already binned"; and chasing an exact "pay line" (the score cutoff above which grants get funded) has long produced meaningless score-compression, where a "92" and a "91" are indistinguishable. His verdict, assuming honest administrators: "not necessarily a bad idea." The real problem, all three agreed, is trust, not method, as one put it: "the tragic part about all this is that three of us are sitting here saying we don't trust the people at NIH."

Why it's only a nuance. This doesn't change any covered company's numbers this quarter. But it refines last week's story: the process change to grant scoring is defensible on the merits and mirrors existing practice; the risk is entirely about who's holding the pen. For the academic-exposed instrument names (Bruker, Agilent, 10x, Bio-Rad, PacBio's research side), the takeaway is that this particular reform is less of a clean negative than it looked last week, and the funding level (item 1) matters far more than the scoring mechanism. Label: academic scientists / science journalists, informed pundits, not operators, and not a demand read.

4. A vivid, operator-grade picture of manufacturing capacity running behind demand, in radiopharma. On Off Script: A Pharma Manufacturing Podcast, "Can Radiopharma Manufacturing Keep Pace with Demand?" (August 18, 2026), the CEO of Perspective Therapeutics, a company developing radioactive cancer drugs, described a corner of pharma manufacturing where demand is badly outrunning supply. His blunt summary: "manufacturing right now is suffering from extraordinary shortage of capacity."

The details are the kind of thing this sector loves, because they show how physical and hard-to-scale advanced-therapy manufacturing is. To shield workers from radiation you need "leaded glass," and it has "a very limited number of suppliers," the thick kind takes "six months to cool down" after Corning makes it, and "they only make it once a year." Standing up a facility "takes about four years." He sized the pipeline pressure behind all this: "about 16 phase 3 programs... around 40 phase 2 programs... over 50 phase 1 programs and over 140 preclinical programs... All of them using isotopes." And the capacity math is brutal: only about "30%" of U.S. lead-shielded manufacturing suites are certified for the highest sterile-manufacturing standard, and only about 30% of those are actually fit for late-stage or commercial work. On the contract-manufacturing (CDMO) firms that fill this gap, he said there is "clearly... a massive demand driver behind the scenes," with CDMOs "absolutely backlogged with pipeline," amid a "war on talent" that is "incredibly fierce."

Why it matters, and why to weight it carefully. This is a fresh, credible operator saying, in effect, that in his slice of the world, demand for specialized manufacturing capacity, testing and supply chains is far ahead of supply. That is the structural "picks and shovels stay scarce and valuable" story the bioprocessing bulls tell. But two big caveats. First, radiopharma is a niche; its bottlenecks (hot cells, isotope supply, leaded glass, cyclotron time) are not the mainstream bioreactors, filters and chromatography resins that Sartorius, Repligen, Avantor and Danaher's Cytiva sell. The read-through is thematic ("advanced-therapy manufacturing is capacity-constrained"), not a direct order signal for our names. Second, he is a drug developer describing his suppliers' demand, not a tools operator quoting a book-to-bill. Genuinely encouraging color on the durability of manufacturing-capex demand, not a covered-name data point.

5. A bioprocessing-podcast operator confirms the funding chill, and reminds us the analytics demand is real. On Smart Biotech Scientist, "279: Why Nanovesicles Outperform Exosomes: Scalable Drug Delivery Beyond Injectable Vaccines with Christopher Locher - Part 1" (August 18, 2026), a show explicitly for process-development and manufacturing leaders, founder Christopher Locher described how the funding environment is squeezing his vaccine company. His line on Washington: "now with the funding rescinded and the NIH kind of rechanging its focus out of vaccines, it's been really challenging to try to raise money for our products." More broadly, he said "the goalpost has been moved," investors who once backed preclinical ideas "now... want to see a phase two clinical study, proof of concept," with "a lot more competition and... less money available."

Two useful tools read-throughs sit inside his story. On consumables, recalling the COVID period: "We couldn't get filters for concentrating and purifying our products. That was a big bottleneck... A lot of products were made in China and a lot of shipping from China was delayed," a reminder (historical, but instructive) that bioprocessing filtration is a genuine choke point when supply tightens. And on analytics, describing how he characterizes his product for regulators: "a lot of it is with mass spec," plus ELISA testing for quantifying and stability, a plain-English example of why every advanced-therapy program is stuffed with measurement instruments (the turf of Waters, Bruker, Agilent, Danaher's SCIEX). He also flagged that "a lot of people have left the FDA," pushing him toward European collaborations.

Why it's a modest, mixed signal. The funding-chill and higher-investment-bar comments reinforce the bear's near-term case (less biotech money = fewer new programs buying gear). The mass-spec/ELISA and filtration comments reinforce the bull's structural case (any program that does advance is instrument- and consumables-hungry). Label: adjacent operator, a drug developer on a bioprocessing show, credible on his own supply chain, but not a covered name and not an order-book read.

The debate

We steel-man both sides every week. This week the concrete facts pile up on the bear side.

The bull (bioprocessing recovery / CGT and NGS re-acceleration). The structural "picks and shovels" story keeps getting reinforced by credible operators. A radiopharma CEO describes demand so far ahead of manufacturing capacity that leaded glass is on a once-a-year backorder and CDMOs are "absolutely backlogged." A bioprocessing-podcast founder reminds everyone that any program reaching the clinic burns mass spec, ELISA and filtration capacity. And on the cell-and-gene-therapy front, the commercial demand is visibly real: a markets podcast walked through Vertex and CRISPR Therapeutics' sickle-cell gene therapy, CASGEV, doing $76 million in the second quarter, "up 78% from the prior quarter, up 151% year over year." Growing CGT sales eventually pull through more reagents, more testing, more manufacturing. The bull doesn't need every academic lab flush; it needs biopharma and advanced-therapy demand to keep compounding, and the operators who spoke this week all said it is.

The bear (China / academic funding / tariffs / lumpy capex). And yet the hard, verifiable news this week was almost entirely negative for demand. The head of the NIH's flagship genomics program said it has lost nearly 80% of its base funding and is physically collecting fewer DNA samples, a direct cut to sequencing and consumables demand. A California advocate described the domestic research base being hollowed out as China spends $500 billion a year and lures away scientists, venture money and IP with 20-year funding offers, while U.S. NIH money "has not come back." A regulatory attorney said smaller companies are so cash-strapped it's "all we can do to keep our clinical trial funded" (DarshanTalks, August 20). The developments with a number attached this week all point the near-term academic and early-biotech demand curve down.

Our read. This week tilts bearish on the facts and neutral-to-bullish only on the long-run structure. The concrete, fresh, checkable developments (an 80% funding cut at a real sequencing customer, and China openly poaching the U.S. research base) both point the near-term academic and early-biotech demand curve down, and they land squarely on the names with the most university/government exposure (Bruker, Agilent, 10x, Bio-Rad, PacBio's research side, and the research-programs side of Illumina). Against that, the structural manufacturing-capacity and advanced-therapy stories stayed intact and even got vivid new color, though from niches (radiopharma) and adjacent operators rather than the order books of the names we own. The honest posture: the academic/genomics-funding leg of demand got measurably worse and now has a number attached, while the industrial/advanced-therapy leg still looks structurally fine. Keep marking the model to the two things we actually learned this week: a quantified funding cut hitting sequencing/consumables demand, and China accelerating as a competitor for the U.S. research base.

Stocks in play

Every ticker discussed by name, or with a direct, specific read-through, this week, with an honest label on who was speaking. Every read below is second-hand, from an NIH official, an industry advocate, adjacent operators, or pundits.

Read-throughs

Academic / genomics funding, the week's hard, verifiable update, and it's a demand negative with a number on it. Last week the story was "the NIH headline budget held, but it funds fewer labs." This week it got specific and physical: the NIH's All of Us precision-medicine program has lost nearly 80% of its base funding since 2023 and is now collecting DNA and biospecimens from only ~600,000 of 880,000 willing enrollees "because of a money gap." That is a direct reduction in sequencing runs, sample-prep kits and lab consumables at a real, recurring buyer, the most concrete link yet between the funding fight and demand for what Illumina, 10x, PacBio and Bio-Rad sell. Offsetting nuance: the program is genuinely bipartisan, and one-time Cures Act money plus live congressional talks could soften the cliff.

China, back after weeks, but as a rival, not a customer. The read is a slow hollowing-out of the U.S. research base (China at $500B/yr R&D, luring VC, talent and IP with multi-decade funding; U.S. NIH money "not come back"; California set to cap R&D tax credits post-2030). Two implications for the group: near term, fewer U.S. labs and early-biotechs buying instruments; longer term, demand shifting into China, where Western tools names (TMO, DHR, A, ILMN, WAT, MTD, RVTY) sell but face tariffs and rising local competition.

Bioprocessing / advanced-therapy manufacturing (Sartorius, Repligen, Avantor, Maravai, Danaher/Cytiva). There was no fresh bioprocessing book-to-bill read this week; the slippage Danaher warned about weeks ago (toward 2027) has no update. But the structural "manufacturing stays scarce and equipment-hungry" case got vivid new color: a radiopharma CEO describing capacity so tight that leaded glass is a once-a-year backorder and CDMOs are "backlogged," and a bioprocessing-podcast founder pointing to mass spec, ELISA and filtration as choke points. Both are adjacent operators in niches, not covered-name order signals, so the long-run demand curve looks intact while the near-term book-to-bill remains unconfirmed.

Cell & gene therapy demand (Maravai, Repligen, Avantor, Danaher). A modest positive at the drug-developer level: Vertex/CRISPR's sickle-cell gene therapy CASGEV grew to $76M in Q2, +78% sequentially. Commercial CGT ramping is, over time, supportive of reagent, testing and manufacturing demand, but this is a pundit read on the therapy's revenue, not a supplier's order book, so treat it as directional backdrop.

Reshoring / CDMO capacity. A regulatory attorney reported larger pharma/device companies "setting up shop or building more in the United States" under "America first," questioning reliance on Chinese and Indian API makers (with the BioSecure Act adding friction), while smaller companies can't afford to reshore and are just trying to keep trials funded (DarshanTalks, August 20). Separately, a GLP-1 podcast (retail commentary) amplified a Bloomberg framing that China's WuXi AppTech has a peptide-manufacturing "chokehold," with Lilly and Novo pouring billions into U.S. capacity. Net: a slow, structural tailwind for domestic bioprocessing/fill-finish demand, offset by cash-strapped small biotech. All adjacent/pundit sources, backdrop, not a covered-name read.

FDA capacity, a quieter risk worth noting. Two separate operators/pundits flagged that "a lot of people have left the FDA" (DOGE-driven cuts), with reviewers pressured to move faster and talent reluctant to join. Not a direct tools read, but a slower or more erratic FDA affects the pace at which new therapies advance, and therefore the timing of the equipment and reagent orders that follow them.

What changed vs last week

Last week (August 16) ended a hair bearish-on-facts, bullish-on-hope: the concrete news was the NIH funneling money into ~3,000 fewer grants, while the encouraging item was a second-hand read that Thermo Fisher's early-stage equipment orders were rebounding on life-sciences M&A. A cell-therapy CMO reaffirmed the CGT-manufacturing bottleneck (naming Lonza), and a macro strategist saw equipment/R&D spending growth cooling slightly. This week, three things shifted:

1. The funding negative went from a grant-count estimate to a physical, program-level demand cut. Last week: "~3,000 fewer grants." This week: the NIH's own flagship genomics program says it has lost nearly 80% of base funding and is collecting fewer DNA samples for lack of money. That is a straight line from Washington to fewer sequencing runs and consumables orders, a more concrete, more clearly demand-relevant negative than last week's grant math, and it lands on the sequencing names specifically.
2. China re-entered the conversation, but as a competitor, not the demand catalyst bulls want. This week an industry advocate made it the centerpiece: $500B/yr R&D, poaching scientists and venture capital, IP migrating overseas. It's the first substantive China discussion in weeks, and its direction is negative for the U.S. research base.
3. The NIH "politicization" worry got a more balanced sequel. Last week a journalist warned about political appointees steering funding. This week, working scientists examined the specific grant-scoring change and concluded the mechanism is defensible and NSF-like: the risk is trust in the people, not the method. A small but genuine softening of one strand of last week's bear case (the funding level still matters far more).

What did not change: Last week's open question, whether Danaher's soft bioprocessing guide was company-specific or sector-wide, got no fresh read, so it stays open. The one thread that carried forward and strengthened is the academic/genomics-funding squeeze, now the clearest, most quantified demand story of the period.