Richard Jaenisch surveys pharma's biggest recent AI deals

Manufacturing Intelligence: Sizing Up Pharma's Biggest AI Deals, Part One

Three regulatory actions are converging to move biologics, gene therapy, and the ATMP platform forward quickly

Regulatory Actions Impacting Biologics

June saw regulators look back at their achievements, take stock of drug shortages, and advance treatments for rare diseases.

Regulatory Actions for June 2026

The use of artificial intelligence is facing harder questions of governance and data quality

PharmTech Weekly Roundup—July 17, 2026

How evolving regulatory expectations are reshaping pharma development and manufacturing.

Drug Digest April 2026

PharmTech Weekly Roundup—August 7, 2026

 featured articles about how policymakers are rewriting the rules for how drugs get made, including an approval for a messenger RNA (mRNA) vaccine and the White House’s blueprint pushing deregulation. Manufacturers are wrestling with whether their own systems, partnerships, and data can keep pace with the evolving regulatory system.

, the first mRNA flu vaccine, after initially refusing to review the application earlier in the year, saying more data were needed. Will this clearance test out the platform's manufacturing case against a backdrop of federal mRNA funding cuts?

In late July 2026, the White House Office of Science and Technology Policy released the report, 

, which holds up pharma's recent deregulation as a model for other sectors while pushing reshoring and new funding pathways beyond the standard NIH grant.

 for more on recent regulatory actions and upcoming agency meetings.

On August 4, Pathos AI agreed to pay $125 million upfront to a subsidiary of Alphamab Oncology for rights to an experimental treatment for triple-negative breast cancer. The 

 may see the AI-driven “Foundry” platform reshaping how oncology assets get selected, as cross-border deals increasingly hinge on manufacturing coordination across jurisdictions.

Dr. Hui-Yin (Harry) Li of Wilmington PharmaTech told 

 that sponsor-contract manufacturing organization relationship breakdowns rarely stem from weak quality agreements but from unclear communication and escalation pathways, making tiered governance the real differentiator.

 AI tools have arrived, but organizational readiness, employee sentiment, and data governance are now the real bottleneck.

Across this week's stories, it appears that policy and platforms are advancing quickly, but governance, data, and manufacturing coordination will be essential for success. For the full stories, visit 

Click the video to above to watch this week’s highlights.

Videos

This week’s content included Moderna's mRNA flu approval, CMO governance tips, OSTP’s deregulation push, a $2 billion oncology license, and July's regulatory wrap-up.

Partha Anbil is at the intersection of the Life Sciences industry and Management Consulting. He has over 30+ years of experience in Life Sciences. He is also a Life Sciences industry advisor at MIT, his alma mater. He held senior leadership roles at WNS, IBM, Booz & Company, Symphony, IQVIA, KPMG Consulting, and PWC. Mr. Anbil has consulted with and counseled Health and Life Sciences clients on structuring solutions to address strategic, operational, and organizational challenges. He is a diplomat/fellow at MIT CSAIL. He is a healthcare expert member of the World Economic Forum (WEF). He was a member of the IBM Industry Academy, a very selective group of professionals inducted into the academy by invitation only, the highest honor at IBM.

Deepak Manjarekar is a veteran Data & AI leader with over three decades of experience driving business transformation through data. As Head of Coforge's Global Data HBU, he leads the company's Data & AI growth strategy, building on leadership roles at Mphasis and KPIT, where he successfully scaled high-growth data businesses.

Life sciences organizations are caught in a familiar bind: the pressure to generate faster, richer insights keeps growing, while the regulatory and privacy obligations that govern how data is handled grow right alongside it. Traditional data warehouses—the structured, relational databases that have anchored enterprise reporting for decades—were built for a different era. They excel at curated, structured reporting, but they were not designed to absorb the sheer variety of data that modern drug development and manufacturing demand: genomics files, medical images, high-velocity sensor streams from production lines, real-world evidence from claims databases, and the free-text narratives that populate pharmacovigilance case reports.

A cloud data Lakehouse offers a practical way through this tension. The term refers to an architecture that combines the low-cost, flexible storage of a data lake with the governance, reliability, and query performance traditionally associated with a data warehouse. Rather than forcing organizations to choose one or the other, a Lakehouse sits in between—storing raw data in open, vendor-neutral file formats while layering on the controls and structure that regulated industries require. Crucially, it does not demand a wholesale replacement of existing systems. The most successful migrations in life sciences have been deliberate and phased, building trust incrementally across technology, organizational culture, and the regulatory environment.

The core technology choices in this transformation hinge on adopting open storage formats, guaranteeing transactional robustness, instituting disciplined metadata and lineage controls, and ensuring seamless interoperability with existing reporting environments and AI platforms. Equally critical are the organizational decisions: securing committed executive sponsorship, establishing shared cross-functional stewardship, and navigating the transition period when legacy and modern systems must operate in parallel. All of this must remain resilient under the industry's regulatory pressures: GxP expectations, HIPAA and GDPR privacy obligations, and 21 CFR Part 11 requirements for electronic records.

The data landscape in life sciences has changed faster than most legacy architectures can accommodate. A single phase III clinical trial today might draw on electronic data capture (EDC) records, patient-reported outcomes, central laboratory feeds, wearable sensor streams, imaging data, and real-world data sourced from external partners, all simultaneously.

 A biologics manufacturing network must continuously synthesize batch records, process analytical technology (PAT) signals, environmental monitoring readings, and laboratory information management system (LIMS) results to maintain quality and meet release timelines. A traditional warehouse can handle selected slices of this information, but only after labor-intensive extract-transform-load (ETL) processes—the pipelines that pull data from source systems, reshape it, and load it into the warehouse—have run their course. Those pipelines introduce delay, and in a world in which clinical teams need to act on site performance data the same day it is generated, delay is a strategic liability.

The Lakehouse paradigm addresses this directly (Table I). Raw data land once they are in scalable cloud object storage, preserving the original form. From there, curated layers apply standardization, quality controls, and access policies in a structured progression. By leveraging open table formats and transaction logs, organizations gain consistency, “time travel” (the ability to query data exactly as it appeared at a specific point in the past, which is directly relevant to audit and inspection readiness), and reproducible analytics. Compute power can be scaled up or down independently of the workload: a compliance dashboard has very different performance requirements from a genome-wide association study. The strategic prize is not cheaper infrastructure, though that matters. It is the creation of governed data products that can power faster trials, sharper safety surveillance, and more defensible regulatory submissions from a single, unified foundation.



Table I: A Life Sciences Reference Architecture

Table I. A Life Sciences Reference Architecture⁴,⁵

Abbreviations: AI/ML, artificial intelligence/machine learning; CDISC, clinical data interchange standards consortium; CRM, customer relationship management; CTMS, clinical trial management system; EDC, electronic data capture; EHR, electronic health record; RWE, real-world evidence; ERP, enterprise resource planning; eTMF, electronic trial master file; HL7, health level seven; FHIR, fast healthcare interoperability resources; LIMS, lab information management system; MES, manufacturing execution system; QMS, quality management system; SDTM/ADaM, study data tabulation model/analysis data model. 


Adapted from: Databricks. What Is the Medallion Lakehouse Architecture. Available at docs.databricks.com/aws/en/lakehouse/medallion. Accessed Aug 6, 2026.

What This Looks Like in Practice: A Lesson from Roche

Roche encountered a challenge that will feel familiar to many. The organization had built a clinical-trials data hub that housed more than 360 studies and over 1 billion documents.

 By any standard, it was a massive and valuable data estate. Yet, the platform was beginning to buckle under its own scale. Legacy silos prevented EDC data (i.e., structured case-report-form records) from non-CRF sources, such as laboratory outputs, imaging files, and pharmacokinetic datasets. Manual handoffs slowed the pace of new capabilities. Compliance and reporting demanded heavy manual effort. And as data volumes exceeded 50 terabytes, the underlying architecture simply could not keep up.

Roche's response was methodical rather than dramatic. The team migrated the platform to a cloud-native architecture on AWS, integrating tools to support real-time ingestion and harmonization of both structured and unstructured clinical data. Crucially, they did not attempt a single cutover. Instead, they ran a phased implementation: first validating the new environment at small scale for security and access controls, then deploying a full production-like environment for performance tuning, and finally executing a validated, GxP-compliant go-live with about 2 days of migration downtime. The outcomes were concrete: release cycles shortened from weeks to days, search and ingestion performance improved substantially, and the platform achieved end-to-end compliance traceability across all trials. Total cost of ownership fell through cloud auto-scaling and infrastructure optimization.

The Roche example is instructive because the approach was disciplined. The team validated before they cut over. They ran parallel environments long enough to build confidence. They treated compliance as a design requirement, not a post-migration checklist item. That discipline is the thread that runs through every successful Lakehouse migration in regulated industries.

Migration Strategy: Phase the Work Around Risk and Value

A pragmatic migration begins by sorting existing warehouse workloads into three distinct groups (Table II). The first comprises stable, highly regulated reporting assets (e.g., batch disposition reports, submission-ready datasets, validated quality dashboards) that demand stringent validation before any change. These should migrate only after data lineage and reconciliation have been definitively proven in the new environment. The second group includes high-value analytical workloads that are currently bottlenecked by warehouse costs or latency: clinical enrollment forecasting, manufacturing process monitoring, safety signal detection. These are strong candidates for early migration because the business case is clear and the tolerance for some transitional complexity is higher. The third group encompasses exploratory workloads (omics analysis, natural language processing over medical literature, real-world evidence studies) in which the Lakehouse delivers immediate net-new value, because legacy warehouses were never designed to handle them at all.


Table II: Illustrative Prioritization of Life Sciences Workloads

Table II. Illustrative Prioritization of Life Sciences Workloads

Adapted from: Kah M, Mui J, Ames M, et al. Migrating a Research Data Warehouse to a Public Cloud: Challenges and Opportunities. J Am Med Inform Assoc. 2022;29(4):592-600.


The phased sequence that emerges from this segmentation—assessment, pilot, parallel run, validated cutover, optimization—is not merely a project management preference. It is a risk management strategy. Running the old and new systems in parallel for a defined period allows quality and regulatory teams to compare outputs, identify discrepancies, and sign off on equivalence before the legacy system is retired. Every parallel run should have a strict exit plan tied to measurable criteria: reconciled key metrics, accepted performance benchmarks, security sign-offs, and hard retirement dates for the redundant pipelines. Indefinite duplication drains budgets and creates confusion about which system holds the authoritative version of a given dataset.

In regulated environments, technology choices carry weight beyond performance benchmarks. They must also support auditability, long-term data accessibility, and the ability to demonstrate to an inspector exactly where a number came from and what transformations it passed through.

Open storage formats are the foundation. Parquet is a column-oriented file format that stores data in a highly compressed, efficient structure optimized for analytical queries; it has become the de facto standard for lakehouse storage. Unlike proprietary formats tied to a single vendor's platform, Parquet files can be read by a wide range of tools, which matters for long-term data accessibility and regulatory submissions.

On top of Parquet, open table formats such as Delta Lake, Apache Iceberg, and Apache Hudi add a crucial layer of reliability.

 Think of these as the equivalent of a database transaction log for a data lake: they ensure that data operations are atomic and consistent (properties collectively described as ACID [Atomicity, Consistency, Isolation, Durability] compliance), support schema evolution as data structures change over time, and enable the time-travel queries mentioned earlier. For a quality team that needs to reconstruct exactly what a dataset looked like on the day a batch was released, this capability is a compliance requirement.

The medallion architecture—a layered approach that organizes data into bronze (raw, unmodified), silver (standardized and quality-checked), and gold (curated, business-ready) tiers—provides the structural framework for applying these technologies in a governed way. Each layer should have explicit business ownership, defined quality thresholds, and documented permitted use cases. The bronze layer is not simply a staging area; it is the immutable record of what was received from source systems, and it may itself be subject to retention and access controls. The gold layer is not simply a reporting database; it is a governed data product with a defined steward, approved business definitions, and audit trails.

Comprehensive metadata and data catalogs are non-negotiable in this architecture. Whether a clinical data manager preparing a submission package or a manufacturing quality engineer investigating a deviation, every user must be able to determine where a dataset originated, what transformations it underwent, and whether it is approved for the decision at hand. Without this, the lakehouse becomes a sophisticated data swamp rather than a trusted analytical foundation.

Performance engineering deserves deliberate attention as well. Clinical dashboards require predictable, rapid response times, whereas complex omics pipelines demand massive, burstable compute. Partitioning—dividing large datasets into smaller, manageable chunks based on meaningful attributes, such as study identifier, country, or manufacturing batch—should reflect actual query patterns rather than storage convenience alone. Data observability tools that continuously monitor for data freshness, volume anomalies, schema drift, and referential integrity are essential for catching problems before they affect downstream decisions.

For life sciences, governance is not an afterthought appended to the end of a project. It is the core of the migration strategy itself, and it must be designed in from the beginning.

Every data domain requires a clearly accountable owner and steward—a designated individual or team responsible for ensuring the quality, accuracy, and proper use of the data within that domain. Governance policies must draw firm boundaries between exploratory research environments, non-GxP analytical workflows, and tightly controlled GxP-relevant reporting. GxP, or Good Practice, refers to the collection of quality standards and regulations that govern pharmaceutical manufacturing, laboratory operations, and clinical trials. Data supporting batch release, regulatory submissions, or clinical decision-making falls under GxP expectations and therefore demands a significantly higher level of validation than data used for internal analytics or business intelligence.

Where electronic records and signatures are involved, procedures must mandate rigorous access controls, audit trails, and system validation consistent with 21 CFR Part 11, the FDA regulation that establishes standards for electronic records and signatures in regulated industries.

 HIPAA-regulated data demands minimum-necessary access, de-identification, and strong vendor controls.

 GDPR obligations must be woven into the design of global studies, particularly where data subjects are European residents.

Validation should be risk-based rather than uniform. Reports used for regulatory submissions or batch disposition require exhaustive test scripts, traceability matrices, and formal change control. Exploratory analytics pipelines do not. A validated lakehouse does not mean freezing innovation; it means maintaining a clear separation between locked-down production data products and agile experimental sandboxes, with a defined pathway for promising analytics to mature into governed assets. This separation is sometimes called a “promotion pathway” and is one of the most important governance constructs an organization can establish, because it allows the scientific community to move quickly without inadvertently compromising the integrity of regulated data products.

Technology is the easier half of a lakehouse migration. The harder half is organizational.

The most successful programs dismantle traditional silos and form cross-functional squads that blend data engineers and architects with clinical subject-matter experts, regulatory operations professionals, and privacy officers. This is not simply good project management practice. It reflects a genuine shift in how data work gets done: from a model in which IT builds pipelines for business users to consume, to one in which data products are co-owned by the teams that depend on them. That shift requires new roles, including data product owners, catalog stewards, and responsible AI practitioners, as well as new accountability structures to support them.

Executive sponsorship is essential, and not merely as a formality. Migrations of this scope inevitably surface legacy metric inconsistencies, ownership disputes, and governance gaps that have accumulated over years. Resolving them requires authority that sits above any individual function. Without visible executive commitment, these issues tend to stall in committee rather than get resolved.

Training must extend well beyond instruction in new software tools. It must encompass the new responsibilities that come with a lakehouse operating model: understanding what it means to own a data product, using a data catalog to assess fitness for purpose, and applying responsible AI practices when building on governed data assets.

A lakehouse migration in life sciences is not, at its core, a technology project. It is a business and compliance transformation enabled by cloud architecture. The organizations that do it well are the ones that resist the temptation to treat it as an IT initiative and instead engage clinical, quality, regulatory, and commercial stakeholders as co-architects of the outcome.

The most effective programs preserve what legacy warehouses do well (trusted reporting, quality controls, strict governance) while expanding the enterprise's ability to work with the full range of scientific, clinical, and operational data generated by modern drug development and manufacturing. They start with high-value use cases that demonstrate the architecture's potential without exposing the organization to unnecessary risk (Table III). They prove the approach through targeted parallel runs before cutting over. They invest heavily in metadata stewardship, because data that cannot be understood and trusted will not be used. And they scale through reusable data products, so that each new use case builds on the governed foundation rather than starting from scratch.


Table III: Illustrative Life Sciences Use Cases

Table III. Illustrative Life Sciences Use Cases⁸

 Adapted from: Kah M, Mui J, Ames M, et al. Migrating a Research Data Warehouse to a Public Cloud: Challenges and Opportunities. J Am Med Inform Assoc. 2022;29(4):592-600.


The result, when the work is done with discipline, is a data foundation capable of powering faster trials, safer products, and responsible AI, and one that holds up under regulatory scrutiny when it matters most.

Disclaimer: The views expressed in this article are those of the authors and not of the organizations they represent.

1. Armbrust M, Ghodsi A, Xin R, et al. Lakehouse: a new generation of open platforms that unify data warehousing and advanced analytics. Proc CIDR. 2021.

2. Nambiar A, Mundra D. An overview of data warehouse and data lake in modern enterprise data management. Big Data Cogn Compute. 2022;6(4):132.

3. Harby AA, Zulkernine F. Data Lakehouse: a survey and experimental study. Inf Syst. 2025;127:102460.

4. Databricks. Lakehouse architecture and medallion architecture technical documentation. Accessed July 29, 2026. 

5. Inmon B, Levins M. Evolution to the data lakehouse. Databricks Blog. Published May 19, 2021. 

6. Datavid. From bottlenecks to breakthroughs: Modernizing clinical data at scale—Roche case study. Accessed July 2026. 

7. Harby AA, Zulkernine F. From data warehouse to lakehouse: a comparative review. 2022 IEEE Int Conf Big Data. 2022:2889-2898.

8. Kahn MG, Mui JY, Ames MJ, et al. Migrating a research data warehouse to a public cloud: challenges and opportunities. J Am Med Inform Assoc. 2022;29(4):592-600.

9. US Food and Drug Administration. Data integrity and compliance with drug CGMP: questions and answers guidance for industry. Published 2024.

10. US Food and Drug Administration. Electronic systems, electronic records, and electronic signatures in clinical investigations: questions and answers guidance for industry. Published 2023.

11. US Food and Drug Administration. Decentralized clinical trials for drugs, biological products, and devices: guidance for industry, investigators, and other stakeholders. Published 2023.

12. US Department of Health and Human Services. Summary of the HIPAA privacy rule. HHS Office for Civil Rights.

13. European Medicines Agency. Guideline on computerized systems and electronic data in clinical trials. Published 2023.

14. International Council for Harmonization. ICH E6(R2) good clinical practice. Published 2016.

15. International Council for Harmonization. ICH E6(R3) good clinical practice draft guideline. Published 2023.

16. Clinical Data Interchange Standards Consortium. CDISC standards for clinical research data, including SDTM and ADaM.

17. Linux Foundation. Delta Lake, Apache Iceberg, and Apache Hudi open table format project documentation.

18. Amazon Web Services, Microsoft Azure, Google Cloud. Documentation on cloud data lakes, analytics governance, encryption, and healthcare/life sciences compliance controls.

Articles

Life sciences data demands now span EDC, imaging, omics, wearables, PAT/MES streams, RWE, and pharmacovigilance narratives, straining legacy warehouses and ETL latency. Cloud lakehouse architectures unify low-cost object storage with warehouse-grade governance, ACID transactions, and performance, enabling auditable “time travel,” scalable compute, and governed data products. Successful adoption hinges on phased, risk-based migration with parallel runs, rigorous metadata/lineage, and compliance-by-design for GxP, 21 CFR Part 11, HIPAA, and GDPR. Roche’s AWS modernization demonstrated faster releases, improved performance, traceability, and lower TCO.

A phased playbook for moving life sciences data warehouses to a compliant, AI-ready cloud Lakehouse, balancing GxP rigor with faster insights.

From Warehouse to Lakehouse: A Life Sciences Playbook for Regulated, AI-Ready Data Modernization

melanoma diagnosis. the doctor examines the patient's mole | Image Credit: ©Evgeniy Kalinovskiy -stock.adobe.com

The FDA announced on August 6, 2026 that it has approved Replimune’s Tudriqev (vusolimogene oderparepvec-wtpg) for the treatment of unresectable advanced cutaneous melanoma in adults who experience disease progression with a programmed death receptor-1 (PD-1)-blocking antibody-based regimen.

 Tudriqev, which is approved for use in combination with nivolumab, is a genetically modified oncolytic viral therapy. The treatment was approved through the agency’s accelerated approval pathway and was granted Breakthrough Therapy and Priority Review designations.

“This is a transformative moment for Replimune, marking years of pioneering research to bring TUDRIQEV to patients desperately in need of new treatment options for advanced melanoma,” said Sushil Patel, PhD, CEO of Replimune, in a press release.

 “We are deeply grateful to the melanoma community, including patients and investigators who participated in the clinical trial, for their tremendous support of this approval. We also would like to thank the FDA for recognizing the urgency to get this therapy to patients. Replimune is now focused on critical activities to deliver TUDRIQEV to as many patients as possible.”

According to the FDA, “Anti-PD-1 refractory melanoma is advanced skin cancer that no longer responds to a widely used class of immunotherapy drugs. Despite treatment, the cancer may continue to grow because of developed mechanisms to evade the immune system’s ability to detect and destroy it.”

The therapy is based on a modified herpes simplex virus type 1 (HSV-1) engineered to target and destroy cancer cells. The virus replicates itself inside a tumor and breaks the cancer cells apart, while also stimulating the body’s immune system to attack the cancer. It is injected directly into tumors once every 2 weeks for 8 consecutive doses. Dosage is determined by the size of the tumor. Used in combination with nivolumab, which is an anti-PD-1 immunotherapy, the FDA believes Tudriqev may create an anti-tumor immune response in patients. Nivolumab is given intravenously to patients beginning at the third week of treatment.

“For patients with advanced melanoma that has stopped responding to PD-1 blocking therapy, the prognosis is often devastating, and options have been far too limited. Clinicians managing these patients know this urgency firsthand,” said Karim Mikhail, B. Pharm., M.S., acting director of the FDA’s Center for Biologics Evaluation and Research, in a press release.

 “Today’s important milestone gives oncologists a meaningful new tool—and more patients a fighting chance.”

“Advanced melanoma patients have few options after anti-PD-1 therapy and face high morbidity and poor survival outcomes,” said Michael K. Wong, MD, PhD, primary investigator of the IGNYTE study and Former Physician in Chief and Professor of Oncology at Roswell Park Comprehensive Cancer Center in Buffalo, NY, in a press release.

 “With the approval of Tudriqev, we have a potent oncolytic immunotherapy that can be used in a broad population, including BRAF-naïve or pretreated, and adjuvant relapsed patients. Importantly, the therapy in combination with nivolumab drives immune activation with a favorable risk-benefit profile and can be injected into superficial and visceral lesions.”

The drug’s approval was based on data from an open-label, multiregional, single arm clinical trial that showed 1 in 4 patients responded to treatment. This positive response lasted a medium of nearly 14 months. The trial consisted of 140 adults with Stage IIIB, IIIC, or IV unresectable advanced melanoma whose disease had progressed for at least 8 consecutive weeks of being treated with an anti-PD-1-based therapy.

The agency also considered input from clinical experts and patient advocates in its approval decision. A meeting to discuss the application was held on July 30, 2026, and included patients, patient advocates, clinicians, and independent experts.

Adverse reactions reported with use of Tudriqev were mild and included fatigue, fever (pyrexia), infections, chills, musculoskeletal pain, nausea, diarrhea, injection site reaction, headache, cough, influenza-like illness, rash, vomiting, itching (pruritus), arthralgia, constipation, decreased appetite, dizziness, dyspnea, hemorrhage, edema, and abdominal pain. There is a risk of spreading a herpes infection to others or a herpes infection developing in patients.

“The approval of RP1 in combination with nivolumab is a meaningful step forward for patients with advanced melanoma who have failed to benefit from immunotherapy,” said Sam Guild, President of AIM at Melanoma, in a press release.

 “Every year, more than 8500 people die of melanoma in the US, and new treatment options are urgently needed.”

Some of the challenges associated with the manufacture of this type of product is the live-virus good manufacturing practice burden. Because this is a genetically engineered, replication-competent HSV-1, production may require dedicated biocontainment infrastructure, such as BSL-2 viral vector suites, segregated air handling, and strict facility controls to prevent cross-contamination with other biologics on site. The treatment also comes with cold-chain and site-of-care logistics due to the fact that it is injected intratumorally in a clinical setting, which requires cold-chain distribution and just-in-time delivery coordination with infusion/injection centers.

FDA approves new engineered viral immunotherapy for patients with treatment-resistant advanced melanoma. Press release. FDA. August 6, 2026. Accessed August 7, 2026. 

https://www.fda.gov/news-events/press-announcements/fda-approves-new-engineered-viral-immunotherapy-patients-treatment-resistant-advanced-melanoma

Replimune announces FDA accelerated approval of Tudriqev in combination with nivolumab for unresectable advanced cutaneous melanoma after progression on an anti-PD-1-based regimen. Press release. Relimune. August 6, 2026. Accessed August 7, 2026. https://ir.replimune.com/news-releases/news-release-details/replimune-announces-fda-accelerated-approval-tudriqevtm

Robilotti E, Zeitouni NC, Orloff M. Biosafety and biohazard considerations of HSV-1-based oncolytic viral immunotherapy. 

https://pmc.ncbi.nlm.nih.gov/articles/PMC10546393/

Beyer S, Lipinski K. Specialized know-how and capacity to support demand for oncolytic viruses. 

. March 19, 2020. https://www.pharmasalmanac.com/articles/specialized-know-how-and-capacity-to-support-demand-for-oncolytic-viruses

FDA granted accelerated approval to Replimune’s intratumoral oncolytic HSV-1 therapy Tudriqev (vusolimogene oderparepvec-wtpg) plus nivolumab for adults with unresectable advanced cutaneous melanoma progressing after PD-1 blockade. Breakthrough Therapy and Priority Review supported the decision. In the single-arm, open-label IGNYTE study (n=140; stage IIIB–IV), objective response was ~25% with median response duration ~14 months. Adverse events were largely mild flu-like and local effects, with herpes transmission/infection risk. Manufacturing demands include live-virus GMP biocontainment and cold-chain, just-in-time clinic logistics.

The genetically modified oncolytic viral therapy, used in combination with nivolumab, was approved by the agency after data showed 1 in 4 patients responded to treatment.

The FDA Gives Accelerated Approval to Melanoma Treatment

Senior Director
Education, Outreach, and Digital Experience
Open Biopharma Research and Training Institute

In the second half of a two-part deep dive on major pharma-AI partnerships with 

’s Chris Cole, Richard Jaenisch, senior director of education, outreach, and digital experience at Open Biopharma, argues that generative AI's promise in drug discovery is real but narrower than the hype suggests. Jaenisch draws a sharp distinction between AI tools that mine existing pharmaceutical datasets to surface overlooked structures — where scientists stay in control and can validate the science — and generative approaches he likens to "chimpanzees writing Shakespeare," where sheer volume substitutes for scientific rigor. The former is producing real pipeline candidates, including one drug now in Phase III; the latter, he says, hasn't yet demonstrated meaningful success. (

Watch Part One on why Roche's raw GPU infrastructure buildout and Sanofi's regional talent hub represent fundamentally different bets on where AI adoption actually stalls in pharmaceutical manufacturing.

In this episode, we discuss AI drug discovery hype, agent security risks, and why pharma's AI workforce bets remain unproven.

Manufacturing Intelligence: Sizing Up Pharma's Biggest AI Deals, Part Two 

2:17 – Where the line is today between what generative AI can do in discovery and formulation and what people imagine it does

9:38 – What’s feeding this disconnect, be it exaggerated capabilities or pure hype

11:16 – Similarities between the ethical conversations around AI and those around embryonic stem cell use

20:17 – Whether or not the Pope voicing opinions on AI has any sway, operationally, for pharma companies

22:09 – Where AI agents actually save time today in pharma where the security risk is just not worth it

31:23 – If there’s a pharma company or situation currently “hamfisting” enough money to make the security risk worth it

37:05 – With workforce repositioning as the metric that determines whether any of these investments pay off, what a pharma company doing this correctly looks like in two years

Jaenisch also draws an unexpected parallel between the current AI ethics debate and the embryonic stem cell controversy, noting that Pope Leo's recent commentary on AI has forced pharma leaders, Catholic or not, into conversations about morality and technology they weren't previously having. He connects this cultural moment to concerns about "AI psychosis" and over-reliance on chatbots as substitutes for human judgment and community.

On agentic AI specifically, Jaenisch is blunt about the security exposure created by ungoverned deployment. Most agents in active use today are simple scraping or reporting tools, he says, but coding agents that are increasingly used to build unauthorized software bolt-ons represent a growing risk, since the same capabilities that help non-experts build tools also help bad actors find exploits in them. He points to real-world incidents of agents breaking sandbox environments and warns that context-window limitations can cause agents to quietly drift from their original instructions over time.

Pressed on whether any pharma company has invested enough to meaningfully move the needle on AI-driven workforce transformation, Jaenisch is skeptical of headline infrastructure spending and argues the real signal will be sustained, team-based investment in people, not individual empowerment. "If everyone's stirring the bucket, the bucket keeps moving," he says, describing how team-oriented AI adoption preserves institutional knowledge even as individuals leave. He predicts the industry will have clearer answers on ROI and workforce strategy within two years, once companies begin reporting real financial results rather than partnership announcements.

The Implications of Orzeyful (oveporexton) approval by the FDA

The FDA approved Orzeyful (oveporexton) tablets on August 5, 2026 for the treatment of narcolepsy type 1 in adults, marking the first therapy designed to treat the disorder as a unified condition rather than a collection of separate symptoms.

 It is also the first approved medicine that works by directly restoring signaling from orexin, the brain chemical whose loss causes the disease.

Narcolepsy type 1 affects an estimated 1 in 2000 people in the United States. It results from the destruction of brain cells that produce orexin, a messenger responsible for regulating wakefulness, sleep, and muscle tone.

 Its absence produces a cluster of disabling symptoms: excessive daytime sleepiness, cataplexy, sleep paralysis, hallucinations at the edge of sleep, and fragmented nighttime rest. Existing treatments have historically addressed these symptoms individually, using stimulants to manage wakefulness or other agents to blunt cataplexy, without correcting the underlying biological deficit.

Why Does the Mechanism Matter More than Symptom Management?

 Rather than masking symptoms, it activates the same brain receptor that orexin itself would normally stimulate, restoring a signal the body can no longer produce on its own. In 2 randomized, double-blind, placebo-controlled 12-week trials enrolling 273 adults, patients taking the 2 mg dose showed improved ability to stay awake during the day relative to placebo, along with reduced daytime sleepiness, fewer cataplexy episodes, and measurable improvement across the disorder's broader symptom profile. The most frequently reported side effects were insomnia, increased urinary frequency and urgency, and excess saliva production, with a low rate of treatment discontinuation due to adverse effects.

Tiffany Farchione, MD, director of the Division of Psychiatry in the FDA's Center for Drug Evaluation and Research, framed the approval as a shift in treatment philosophy, in a press release,

For too long, people with narcolepsy type 1 have had to manage a complex, lifelong neuropsychiatric condition with treatments that only address pieces of it. This new drug is the first medicine that impacts the underlying biology of the disease, treating narcolepsy type 1 as a whole,” she stated.

What Should Development and Manufacturing Teams Take from This Approval?

This approval is a signal of where central nervous system drug development could be headed, toward therapies engineered against a disorder's specific molecular deficit rather than its downstream symptoms.

 Orexin receptor agonism has been an area of longstanding scientific interest, and a first approved product in this class offers a validated mechanism that other developers working in broader neuropsychiatric indications should study closely, both for its clinical trial design and its regulatory pathway. The drug's route through Breakthrough Therapy Designation and Priority Review also reflects how the agency is pushing to expedite review for therapies addressing serious, underserved conditions where no mechanistically targeted option previously existed.

There are practical downstream considerations as well.

 Orzeyful has been recommended for scheduling under the Controlled Substances Act, meaning it cannot be lawfully marketed until the Drug Enforcement Administration finalizes that scheduling decision. The product also carries a labeled interaction with strong CYP3A inhibitors, underscoring the ongoing importance of drug-interaction data in supporting safe use once a novel mechanism reaches the market. Its safety and effectiveness have not been established in patients under 18 years of age, leaving pediatric development as an open question for the orexin receptor agonist class going forward.

How Are Patients and Advocates Responding to the Approval?

Patient advocacy groups have framed the approval as a turning point for a community long limited to symptom-focused treatment.

 Julie Flygare, president and CEO, Project Sleep and a person living with narcolepsy type 1, said in a press release,

 "For those of us living with narcolepsy type 1, symptoms reshape everyday life and extend far beyond what most people understand about the condition. Orzeyful's approval is a historic moment that expands our treatment choices and gives me great hope for the future and for our community." 

FDA approves first drug to treat the full range of narcolepsy type 1 symptoms. US Food and Drug Administration. News Release. August 5, 2026. 

https://www.fda.gov/news-events/press-announcements/fda-approves-first-drug-treat-full-range-narcolepsy-type-1-symptoms

U.S. FDA approves Takeda’s ORZEYFUL™ (oveporexton), the first and only medicine to treat the underlying cause of narcolepsy type 1. Business Wire. August 5, 2026. 

https://www.businesswire.com/news/home/20260805760759/en/U.S.-FDA-Approves-Takedas-ORZEYFUL-oveporexton-the-First-and-Only-Medicine-to-Treat-the-Underlying-Cause-of-Narcolepsy-Type-1

FDA approval of oveporexton (Orzeyful) for adult narcolepsy type 1 introduces the first therapy positioned to treat the disorder as a unified pathophysiologic entity by directly restoring orexin signaling via receptor agonism. In two 12-week randomized, double-blind, placebo-controlled trials (n=273), the 2 mg dose improved wakefulness and reduced daytime sleepiness and cataplexy, with broader symptom benefit. Common adverse events included insomnia, urinary frequency/urgency, and hypersalivation, with few discontinuations. Regulatory acceleration and pending DEA scheduling shape launch timing.

The FDA approved the first orexin receptor agonist for narcolepsy type 1, which validates a new mechanism-driven approach shaping CNS drug development and regulatory pathways.

Why the Orzeyful Approval Matters Beyond One Drug Class

The FDA on Aug. 5, 2026, approved Moderna's mFlusiva (mRNA-1010) for adults 50 and older, granting approval for two age groups (adults 50 to 65 and adults 65 and up) on the condition that Moderna run an additional clinical trial in the older group.

 The company said the vaccine will be available for this fall's flu season.

Moderna's mRNA shot was nearly 27% more effective against seasonal flu than a standard flu shot, according to clinical trial data from a study enrolling more than 40,000 adults aged 50 and older.

 Reactogenicity ran higher than with standard-dose comparators, including higher injection-site pain and systemic symptoms, though events were predominantly mild-to-moderate and resolved within one to two days.

 For adults aged 65 and older, immunogenicity data comparing mFlusiva against Fluzone High-Dose showed superior antibody responses at day 29 and at six months, the data underpinning that group's accelerated approval.

The review was not straightforward. The FDA initially refused this year to review Moderna's application before it reversed course a week later, with agency officials saying they wanted more data because Moderna had compared its vaccine with a standard flu shot in adults aged 65 and older whereas the CDC recommends a high-dose vaccine for that age group.

 An amended, age-stratified application followed, and the FDA's Vaccines and Related Biological Products Advisory Committee voted 9-0 that benefits outweighed risks ahead of the final decision.

That regulatory maneuvering echoes what our 

 series has tracked this year in tariff- and funding-driven shifts affecting mRNA programs, and it lands against a specific policy backdrop: HHS canceled roughly $500 million across 22 BARDA-funded mRNA vaccine projects in 2025, a decision that also forced a stop-work order on 

. The approval is likely to face pushback from allies of the health secretary, underscoring that mFlusiva's clearance came despite, not because of, the current federal posture toward the platform.

What Does the Approval Mean for Manufacturing?

mFlusiva extends Moderna's mRNA-lipid nanoparticle platform beyond COVID-19 into a second major respiratory indication, following the 2

That trajectory has been built on real infrastructure: Moderna's 

, all now underpinning a third product line rather than a single pandemic response.

The core manufacturing argument covered in our look at 

engineering challenges in commercial-scale mRNA production

—shared production infrastructure and rapid reformulation to match circulating strains—gets a real-world test as Moderna scales mFlusiva for fall 2026 distribution.

1. Lovelace, B. "FDA approves 1st mRNA flu shot, from Moderna." NBC News, Aug. 5, 2026. 

2. "FDA approves first mRNA-based flu vaccine." CNN, Aug. 5, 2026. 

3. "FDA Advisory Committee Supports Moderna mRNA Flu Vaccine Review." AJMC, 2026. 

4. "FDA Advisory Panel Votes 9-0 in Favor of Moderna's mRNA Flu Vaccine." BioPharm International, 2026. 

5. "A new type of flu shot is coming after FDA approves mRNA-based vaccine." Washington Post, Aug. 5, 2026. 

FDA approval of Moderna’s mFlusiva (mRNA-1010) introduces the first mRNA-based influenza vaccine for adults ≥50 years, with age-stratified authorization and a postmarketing trial requirement in those ≥65. In a >40,000-participant study, relative vaccine effectiveness was ~27% higher than a standard comparator, with increased but mostly mild-to-moderate, short-lived reactogenicity. For ≥65, accelerated approval relied on superior immunogenicity versus high-dose influenza vaccine at day 29 and six months. The decision arrives amid mixed federal signals toward mRNA programs.

FDA approves Moderna's mFlusiva, the first mRNA flu vaccine, testing the platform's manufacturing case amid federal funding cuts.

What Does Moderna's mRNA Flu Vaccine Approval Signal?

Dr. Hui-Yin (Harry) Li is president and CEO of Wilmington PharmaTech.

The pharmaceutical timeline from discovery to market is a long, complex process. Sponsor companies often use contract organizations, such as contract manufacturing organizations (CMOs), to perform a bulk of the manufacturing processes. Because the sponsor companies, in the eyes of regulators, have the ultimate responsibility to ensure the safety, quality, and efficacy of drug products, maintaining a healthy relationship with CMOs that includes oversight is imperative.

But how is this relationship created and maintained in a way that benefits both sponsor and contractor? 

 spoke with Dr. Hui-Yin (Harry) Li, president and CEO of Wilmington PharmaTech, to understand how these relationships are built and nourished.

PharmTech:How do pharmaceutical companies establish clear quality agreements with CMOs, and what governance structures have you seen effectively align expectations on both sides of the relationship from day one?

A strong sponsor–CMO relationship starts with a well-defined quality agreement that clearly assigned responsibilities for good manufacturing practice (GMP) compliance, quality systems, change control, deviations, investigations, corrective actions and preventive actions (CAPA), regulatory communications, audits, and product release. The agreement should be developed collaboratively and aligned with the commercial contract, technology transfer plan, and regulatory strategy.

Quality agreements typically receive substantial attention during contract negotiations, while governance structures often receive less. In my experience, more challenges arise from unclear communication pathways, decision-making authority, or escalation processes than from the agreement wording.

A quality agreement sets the foundation. Effective governance is what keeps the relationship aligned throughout the program. Strong partnerships establish a tiered structure from the outset with executive sponsors, operational project teams, and quality representatives. Regular joint meetings, defined escalation pathways, and clear decision-making responsibilities keep both parties aligned. Early agreement on communication expectations, risk tolerance, timelines, and success criteria prevents misunderstandings later on.

What communication cadences and data-sharing practices are most effective in building transparency with a contract manufacturer, particularly when deviations or supply disruptions arise?

Transparency is built through consistent communication long before a problem occurs. For active development or manufacturing programs, weekly operational meetings and monthly management reviews are often effective. These meetings should cover project status, manufacturing schedules, quality metrics, change controls, supply chain risks, and upcoming milestones.

When deviations, investigations, or supply disruptions occur, timely communication matters even more. In my experience, sponsors rarely get upset because a deviation occurred; they get upset when they learn about it late. Most sponsors understand that manufacturing is complex and issues can happen. What they expect is prompt notification, transparency, and a credible resolution plan. Sharing data, root-cause analyses, risk assessments, and CAPA plans maintain trust and facilitates collaborative problem solving. Sponsors are far more understanding when informed promptly and involved appropriately, rather than learning about issues after significant delays.

How can sponsors companies maintain meaningful technical oversight without creating an adversarial dynamic or micromanaging the CMO's operations? What governance mechanisms work well?

The most productive sponsor–CMO relationships focus oversight on outcomes, compliance, and risk management rather than day-to-day operations. Sponsors should recognize that CMOs have established quality systems, manufacturing expertise, and operational processes that support multiple clients and regulatory requirements.

A common challenge arises when sponsors manage individual manufacturing activities rather than focusing on process performance, quality outcomes, and key risks. The CMO was selected for its expertise, which should be leveraged while maintaining appropriate visibility into quality, compliance, and program risks. The most effective relationships allow the CMO to execute within its established quality systems and operating procedures while providing sponsors with the oversight they need.

Sponsor companies retain regulatory accountability for product safety, quality, and efficacy, making structured, transparent CMO partnerships essential. Effective engagement begins with collaboratively built quality agreements covering GMP responsibilities, change control, deviations, CAPA, audits, and release, aligned to commercial and tech-transfer plans. Governance—clear decision rights, escalation paths, and tiered forums from executive to operational teams—prevents most friction. Proactive cadence, rapid deviation notification, and sharing of root-cause and CAPA data sustain trust. Long-term alignment and continuous improvement emerge from forecasting transparency, joint performance reviews, and post-campaign lessons learned.

PharmTech spoke with Dr. Hui-Yin (Harry) Li, president and CEO of Wilmington PharmaTech, about steps that should be taken to build a productive relationship between sponsor and contract organizations.

Featured Content

Manufacturing Intelligence: Sizing Up Pharma's Biggest AI Deals, Part One

Regulatory Actions Impacting Biologics

Regulatory Actions for June 2026

PharmTech Weekly Roundup—July 17, 2026

Drug Digest April 2026