Highlights
For decades, animal testing has been the undisputed, mandatory gatekeeper between the lab bench and human trials. But that status quo is finally cracking. The FDA Modernization Act 2.0 of 2022 prompted the FDA to make changes on September 21, 2026. Alongside traditional animal testing, organ-on-chip technology and innovative human cell models can also be used for drug safety evaluation. Building such capabilities early will help companies adapt faster as regulatory expectations continue to evolve.
It could enable companies to identify safety risks sooner and advance promising therapies into clinical development with greater confidence. On the surface, this may seem like a change in terminology. The agency essentially replaced terms like "animal tests" with the much broader "nonclinical studies/trials." But in the regulatory world, language defines intent. A simple terminology change is often a change in policy. Such changes influence how drug sponsors design their pipelines, where investors direct capital and which technologies gain mainstream credibility.
A key question is how regulatory agencies are balancing innovation with patient safety. In this context, it is important to note that the FDA has not abandoned animal testing.
Animal experiments are still important and, in some cases, indispensable. However, it does assume a change in mindset: animal tests are no longer viewed as the only method for obtaining proof of drug safety. The question is whether the chosen method is effective in obtaining the necessary proof. Animal models have played a significant role in medical breakthroughs, yet they cannot perfectly replicate human biology. The regulatory change, therefore, does not replace animal studies with a new default. Instead, it allows sponsors to select a combination of animal, human-derived in vitro and computational evidence that best answers the research question.
The scope of biotechnology and computational science has increased to such an extent that what is possible now might not have existed even a decade ago.Scientists can now cultivate tissues from human cells, recreate aspects of human biology on microchips, create computer models of biological processes, and apply artificial intelligence technology to predict drug actions and toxicity. These methods are labelled New Approach Methodologies, or NAMs, and sometimes offer more relevant information about humans than traditional animal models. Over time, data from these models could be used to create virtual patients or digital bio-twins that simulate how different types of patients may respond to a therapy. These technologies are still emerging and will need careful validation. However, they could eventually improve dose selection, patient stratification and safety assessment.
For pharmaceutical companies, the opportunity goes beyond replacing one type of study with another. It is to use more human-relevant evidence at the right stages of development, helping teams make decisions earlier, reduce avoidable rework and focus investment on candidates with stronger prospects. When used early in discovery, these methods can help scientists screen compounds, compare candidates and identify human-specific toxicity before investing significant time and money. In later phases, organ-on-chip systems and human cell models can provide additional information on organ toxicity, drug metabolism and biological response. Computational models can also support exposure prediction, dose selection and clinical monitoring. Together, these methods can reduce unnecessary experiments and stop unsuitable candidates from moving forward.
Alongside the rule change, the FDA launched a database that shows how NAMs have been used in regulatory reviews. This gives drug applicants practical examples and signals the agency’s willingness to consider these methods. However, sponsors must still demonstrate that their chosen approach is reliable and fit for purpose.
The influence will go beyond the drug-developers. Companies developing digital biology platforms, AI tools for predicting toxicity, advanced in vitro devices, and micro-physiological systems (Organ-on-a-chip) are likely to see growing interest from both investors and pharmaceutical companies. Contract research organisations and life science service companies will need to gain practical experience in model development and validation, data science, computational toxicity, and changing regulatory processes.
This evidence-first approach also recognises that different therapies require different methods. A monoclonal antibody, gene therapy, and a conventional small molecule may each require a different combination of animal studies, human-derived assays, and computational models.
In simple terms, the regulatory question is shifting from "Did you test this in animals?" to "Does the evidence reliably predict human safety?”. Animal models are not going away, but their position as a default approach is beginning to change. Going forward, they are one tool in a much larger toolbox. The transition will take time. New models must be standardised and validated, and both regulators and sponsors need experience interpreting the results. Even so, the direction is clear: animal studies will remain important, but they will increasingly form part of a broader, more human-relevant evidence strategy.