Plutux
Anthropic Built a Wet Lab to Run Its Own Drug Programs. The Losers Are Not Charles River or WuXi — They Are the AI Biotechs Without a Pharma Partner. insight cover
Private CompanyCRL · WUXAY · RXRX•14 min read

Anthropic Built a Wet Lab to Run Its Own Drug Programs. The Losers Are Not Charles River or WuXi — They Are the AI Biotechs Without a Pharma Partner.

Anthropic has quietly stood up a physical biology laboratory in the San Francisco Bay Area to run drug science internally, Reuters reported on Sept. 18, 2026, making the Claude maker the first frontier-AI developer to operate simultaneously as a tool vendor and a first-party drug sponsor. The pivot from biosecurity gatekeeper — Anthropic's Responsible Scaling Policy v3.0 still gates CBRN capabilities at ASL-3 — to wet-lab drug developer does not displace the order books of Charles River Laboratories or WuXi AppTec in 2026: CRL just raised full-year guidance on the strongest biotech book-to-bill in nearly four years, and WuXi AppTec grew H1 2026 revenue 38.9% year over year and lifted guidance by 14%. The actual read-through is structural: standalone AI biotechs like Recursion — Q2 2026 revenue down 60% year over year as the Roche collaboration rolls off — face a two-front squeeze as compute-rich AI labs (Anthropic, Alphabet's Isomorphic) run their own pipelines and pharma giants (Eli Lilly, Novo Nordisk) internalize Claude as the new standard tool.

Published Sep 18, 2026Updated Sep 18, 2026

Charles River Q2 2026 reported revenue

$1.00B

Down 2.7% reported; organic growth +0.1%, vs. -2.7% a year earlier

Charles River DSA organic growth

+0.2%

Q2 2026; driven by higher study volume in regulated safety assessment

Charles River 2026 organic guidance

0.0% to +1.0%

Raised from -1.5% to -0.5% on Aug. 5, 2026; EPS guidance up $0.25 at midpoint

WuXi AppTec H1 2026 revenue

RMB 28.9B

Up 38.9% year over year; backlog up 23.6% YoY to RMB 59.77B

The Sept. 18, 2026 Reuters scoop by Jeffrey Dastin and Michael Erman confirms what Anthropic's job board had been signaling for months: the AI lab behind Claude has built an in-house wet lab in the San Francisco Bay Area and is doing \"real lab work\" — running physical biology experiments, not just in-silico predictions. Head of life sciences Eric Kauderer-Abrams told Reuters the operation mirrors a typical biotech setup: a mix of in-house execution and external partners, with Claude directing robotic units to automate bench science and human scientists providing safety oversight.

Anthropic's stated scope is narrow on purpose — early pre-clinical research on rare and \"undruggable\" targets, no clinical trials yet, focus on diseases \"industry would not address\" — which caps any near-term volume impact on traditional CROs while still validating the strategic question every contract researcher is now asking.

The Pivot From Biosecurity Gatekeeper to First-Party Drug Developer

Anthropic spent2024 and most of 2025 defining the most-cited voluntary AI safety framework in the industry. The Responsible Scaling Policy v3.0, released Feb. 24, 2026, still gates chemical, biological, radiological, and nuclear capabilities at AI Safety Level 3 — protections the company formally activated in May 2025 and updated as recently as August 2026. Less than five months later, the same organization is hiring molecular biologists, RNA scientists, and protein-and-nucleic-acid characterization experts to run bench experiments on its own balance sheet.

  • May 7, 2026 (SynBioBeta): Kauderer-Abrams publicly confirms Anthropic has \"opened wet labs to run its own basic research,\" targeting a 10× compression of life-sciences R&D timelines.
  • June 30, 2026: Anthropic launches Claude Science (the workbench for scientists) and announces its own pre-clinical drug programs focused on neglected diseases — the first frontier-AI lab to do so publicly.
  • April 2026: Anthropic closes the roughly $400 million all-stock acquisition of Coefficient Bio, an eight-month-old New York startup, bringing in biotech operational talent (target selection, modality choice, portfolio planning).
  • 2026 hires: John Jumper, the Nobel-winning architect of AlphaFold, joins the life-sciences team alongside procurement and biochemical characterization staff.
  • Sept. 16, 2026: Novo Nordisk announces a Claude-based drug development partnership — Anthropic now signs Big Pharma customers while building its own pipeline in parallel.
Anthropic is not the only frontier-AI player running internal pipelines. Alphabet's Isomorphic Labs already has multi-target deals with Eli Lilly and Novartis worth nearly $3 billion combined (Jan. 2024) and is preparing AI-designed oncology candidates for clinical trials. Anthropic's entry formalizes a two-track market — AI labs as both vendor and sponsor — rather than a one-off experiment.

Why Charles River and WuXi Are Not the Story (The Data)

The obvious read — that an AI lab running internal biology work will eat into contract research organization order books — does not survive a quarter of 2026 financials. Biotech funding has now grown for four consecutive quarters, reaching $34.3 billion in Q2 2026, and that money is what flows into preclinical CRO pipelines. Both of the largest Western and Chinese CRDMO/CRO players are raising guidance, not bracing for disruption.

Charles River Q2 2026 reported revenue

$1.00B

Down 2.7% reported; organic growth +0.1%, vs. -2.7% a year earlier

Charles River DSA organic growth

+0.2%

Q2 2026; driven by higher study volume in regulated safety assessment

Charles River 2026 organic guidance

0.0% to +1.0%

Raised from -1.5% to -0.5% on Aug. 5, 2026; EPS guidance up $0.25 at midpoint

WuXi AppTec H1 2026 revenue

RMB 28.9B

Up 38.9% year over year; backlog up 23.6% YoY to RMB 59.77B

Charles River Laboratories — Reported Quarterly Revenue

Revenue stabilized near $1.0B through 2025–2026; Q2 2026 saw the highest net book-to-bill in nearly four years.

Unit: USD millions

Q2 2024

1,009.8

Q3 2024

1,009.8

Q4 2024

1,002.5

Q1 2025

984.2

Q2 2025

1,032.1

Q3 2025

1,004.9

Q4 2025

994.2

Q1 2026

995.8

Q2 2026

1,004.1

WuXi AppTec's trajectory is even starker. The Chinese CRDMO posted RMB 12.44 billion in Q1 2026 (up 28.8% year over year) and RMB 16.46 billion in Q2, lifting full-year 2026 guidance by 14% on Aug. 2, 2026. The BIOSECURE Act — signed into law Dec. 18, 2025 as part of the FY2026 NDAA — has so far been more of a tailwind than a headwind: a U.S. District Court granted WuXi AppTec a preliminary injunction on Aug. 7, 2026 barring DoD enforcement of its 1260H listing, while federal procurement restrictions cannot bite new contracts until OMB completes the Biotechnology Companies of Concern list (no later than Dec. 2026) and the FAR is updated (potentially mid-to-late 2028).

WuXi AppTec — Quarterly Revenue (CNY)

Despite BIOSECURE Act headwinds, WuXi AppTec revenue compounded through 2025 and accelerated in H1 2026.

Unit: CNY millions

Q3 2025

11,982

Q4 2025

12,599.4

Q1 2026

12,435.8

Q2 2026

16,461.7

The textbook \"AI displaces CROs\" thesis misses the supply-chain reality: AI labs running early pre-clinical programs in rare and neglected diseases do not bid for the same safety-assessment, IND-enabling tox, or large-scale manufacturing slots that Charles River Laboratories and WuXi AppTec sell. They are doing work CROs were never asked to do — which means order-book cannibalization is a2028–2030 story, not a 2026 one.

The Real Losers: AI Biotechs Without a Pharma Partner

If Anthropic-as-sponsor doesn't hurt CRL or WuXi directly, it does change the competitive geometry for the standalone AI-native biotech platforms that built entire business models around selling AI-discovered molecules back to pharma. The numbers at Recursion Pharmaceuticals — the most-watched public AI biotech since the NVIDIA partnership in 2023 — show the strain.

Recursion Q2 2026 total revenue

$7.67M

Down 60% year over year from $19.22M; Roche collaboration expirations

Recursion Q1 2026 total revenue

$6.47M

Down 56.1% YoY; collaboration revenue collapse

Recursion TTM operating cash flow

Negative

$361M TTM free cash flow burn; R&D spend of $89.6M in Q2 2026 alone

Recursion Q2 2026 net loss

$131.0M

Diluted EPS -$0.25; vs. -$0.41 a year earlier, but on a much smaller revenue base

Recursion's August 5, 2026 quarterly release offered a small bright spot — Genentech opted in to its first neuroscience target from the Roche collaboration, and a fifth progress-based milestone triggered a $4 million payment in February 2026 — but the structural picture is deteriorating. With pharma now able to license Anthropic's Claude Science and similar platforms directly, the value of \"discovery-as-a-service\" intermediaries is compressing. Compare this with Charles River Laboratories, whose Q2 2026 demand improvement was broad-based across \"global biopharmaceutical and small and mid-sized biotechnology clients\" — including the AI-native biotechs that still need regulated toxicology work the AI labs cannot do themselves.

AI Drug Discovery Competitive Landscape — Sept. 2026
PlayerStatusPipeline / PartnershipCapacity Signal
Anthropic (private)Tool vendor + first-party sponsorWet lab open Sept. 2026; neglected-disease programs$30B+ ARR (April 2026); $65B ARR by July 2026
Alphabet / Isomorphic LabsTool vendor + first-party sponsorMulti-target deals with Eli Lilly, Novartis, J&J$3B pharma partnerships; $2.1B Series B (Apr. 2025)
Xaira Therapeutics (private)AI-native sponsorFoundational models; protein drug design$1B launch funding (Apr. 2024)
Recursion Pharmaceuticals (RXRX)Platform-as-a-serviceGenentech option Aug. 2026; pipeline thinningQ2 2026 revenue $7.67M, -60% YoY
Charles River Labs (CRL)Traditional CROPre-clinical safety, RMS, manufacturingQ2 2026 organic growth back to +0.1%
WuXi AppTec (WUXAY)CRDMOSmall-molecule CRDMO; backlog RMB 59.77BH1 2026 revenue +38.9% YoY

The Supply Chain Reads Through

Three layers matter for the rest of the value chain. Upstream, the AI compute infrastructure play belongs to NVIDIA, whose announced $1 billion drug discovery lab partnership with Eli Lilly — plus its residual stake in Recursion — positions it as the picks-and-shovels beneficiary regardless of which AI lab wins. Midstream, the platform layer is bifurcating: capital-rich frontier AI labs (Anthropic, Alphabet/Isomorphic) absorb the platform-plus-sponsor economics; standalone AI biotechs (Recursion, and private peers like Insitro) face a narrowing lane. Downstream, the Big Pharma winners are those who lock in platform access early — Eli Lilly holds deals with Isomorphic (Jan. 2024, $45M upfront per multi-target) and NVIDIA; Novo Nordisk just signed the Anthropic partnership (Sept. 16, 2026).

  • NVIDIA — Upstream AI compute; $1B pharma lab with Eli Lilly; $50M equity stake in Recursion from the 2023 collaboration
  • WuXi AppTec — Downstream CRDMO; backlog of RMB 59.77B and BIOSECURE injunction in hand after Aug. 7, 2026
  • Charles River Laboratories — Downstream CRO; DSA segment recovery confirms biotech-funding-to-CRO pipeline transmission still works
  • Recursion Pharmaceuticals — Direct competitor to first-party AI labs; revenue model shows what happens when a partnership expires
  • Eli Lilly — Pharma; holds multi-target Isomorphic deal (Jan. 2024) and NVIDIA AI lab partnership
  • Novo Nordisk — Pharma; first publicly disclosed Claude-based drug development partnership (Sept. 16, 2026)

Short-Term Catalysts vs. Long-Term Structural Shift

Days to quarters: the Reuters scoop does not change2026 CRO fundamentals. Charles River Laboratories is exiting its post-pandemic trough with the strongest net book-to-bill in nearly four years; WuXi AppTec is compounding38% H1 growth and raising guidance by 14%. Biotech funding at $34.3B in Q2 2026 (the fourth straight quarter of growth) is what fills preclinical pipelines, and AI labs running neglected-disease programs do not draw on that same dollar pool. Watch the next two earnings calls (CRL Q3 in early November, WuXi Q3 around the same window) for confirmation.

One to three years: the structural shift is real, but it cuts in a direction most CRO bulls do not expect. The threat to the Charles River Laboratories order book is not Anthropic taking work in-house — it is Anthropic empowering every small-cap biotech to run more candidates into IND-enabling studies faster, which is good for volume but bad for CRO pricing power. The threat to WuXi AppTec is geopolitical, not technological: BIOSECURE Act enforcement timing (FAR update potentially mid-to-late 2028) matters more than Anthropic's lab. The threat to Recursion Pharmaceuticals is existential — its Q2 2026 revenue collapse is the early signal of what happens to a platform when compute-rich AI labs build the same capability in-house and pharma signs platform deals directly.

Investors who want exposure to the AI drug discovery trade are better served owning the picks-and-shovels layer — NVIDIA for compute, Eli Lilly and Novo Nordisk for the Big Pharma adoption signal — than the standalone AI biotechs whoseplatform value is structurally being absorbed by the frontier labs themselves.

Investable Read-Through

CCharles River LaboratoriesCRL--
--Vol --
-
Mixed
  • Q2 2026 demand environment \"broad based across global biopharmaceutical and small and mid-sized biotechnology clients\" — DSA segment organic growth back to +0.2%, the strongest book-to-bill in nearly four years.
  • Full-year 2026 organic guidance raised by 150bps to 0.0%–1.0% on Aug. 5, 2026; non-GAAP EPS guidance lifted $0.25 at midpoint — Anthropic's wet lab is not displacing order book in2026.
  • Structural long-term risk: AI labs accelerating candidate flow into preclinical tox could mean more volume but compressed pricing — the multi-year CRO margin story turns into a throughput story.
WWuXi AppTecWUXAY--
--Vol --
-
Bullish
  • H1 2026 revenue RMB 28.9B, up 38.9% YoY; backlog up 23.6% YoY to RMB 59.77B — backlog visibility through 2027 even with BIOSECURE overhang.
  • Aug. 7, 2026 U.S. District Court injunction blocks DoD enforcement of the 1260H listing; full-year 2026 guidance raised 14% on Aug. 2, 2026.
  • Near-term: AI labs are not yet customers, but their tool offerings (Claude Science, Isomorphic platforms) accelerate global biotech R&D pipelines — WuXi is the CRDMO beneficiary of any AI-driven candidate volume.
RRecursion PharmaceuticalsRXRX--
--Vol --
-
Bearish
  • Q2 2026 total revenue $7.67M, down 60% YoY as the Roche collaboration rolls off — direct evidence that pharma is no longer paying standalone AI biotechs full platform margins.
  • TTM free-cash-flow burn of $361M against $1.86B market cap — multiple quarters at this cash burn rate stress the balance sheet by mid-2027.
  • Aug. 5, 2026 Genentech opt-in on the first neuroscience target shows the partnership model still works deal-by-deal, but does not reverse the structural pressure from Anthropic and Isomorphic absorbing the platform layer.
LEli LillyLLY--
--Vol --
-
Bullish
  • Holds a multi-target deal with Alphabet's Isomorphic Labs ($45M upfront, Jan. 2024) plus the NVIDIA AI drug discovery lab partnership — two of the three largest AI drug platforms locked in.
  • Anthropic's Claude Science launch (June 30, 2026) and tool-vendor-plus-sponsor model gives Eli Lilly another platform option without depending on any single AI vendor.
  • Days-to-quarters: pipeline acceleration from AI tools should compress clinical timelines; one-to-three years: structural advantage over smaller pharma peers who cannot afford multi-platform AI deals.
NNovo NordiskNVO--
--Vol --
-
Bullish
  • Sept. 16, 2026: Anthropic partnership to use Claude across the discovery and development pipeline — first major pharma-name disclosure of an Anthropic drug deal.
  • TTM revenue $329.4B with 35.3% net margin provides the cash flow to out-spend any AI biotech competitor on platform access.
  • Long-term: AI-driven candidate acceleration in cardiometabolic space aligns with Novo Nordisk's core diabetes/obesity franchise and the Valo Health cardiometabolic collaboration.
NNVIDIANVDA--
--Vol --
-
Bullish
  • Picks-and-shovels play: $1B pharma drug discovery lab partnership with Eli Lilly plus the residual $50M equity stake in Recursion (2023 deal) — exposure regardless of which AI lab or biotech wins.
  • Short-term: AI drug discovery market estimated at $3–4B in 2026, growing at 23–28% CAGR toward $17.56B by 2031 — compute layer captures a disproportionate share.
  • Long-term: as frontier AI labs (Anthropic, Isomorphic) absorb platform economics, NVIDIA's infrastructure position becomes the durable layer of the value chain.

Plutux is not an investment adviser. Market data and AI-generated analysis are for information and education only, not investment advice. Disclaimer

© Plutux Technology Limited 2026