Pfizer, Anthropic and longevity scientists back AI in health

Pfizer, Anthropic and a group of longevity researchers reported separate developments this week that involve artificial intelligence in drug design, model building and aging research. The announcements included a potential AI-designed drug candidate, a new version of a large language model and claims that AI could change how age-related disease is studied.

Pfizer’s chief executive said the company is reviewing a molecule its teams generated with AI. Pfizer has paid up to $350 million to an external small-molecule design firm since 2020, entered a collaboration in January to refine open-source biomolecular models on its internal data, and its venture group has backed longevity-focused efforts. The company reported first-quarter results that referenced AI accelerating work across the organization.

Any candidate designed with AI must still complete preclinical testing and controlled clinical trials before it can be approved for patient use. Companies can produce candidate molecules quickly, but development timelines remain governed by laboratory testing, regulatory review and trial outcomes.

At an invite-only financial services event in New York, Anthropic introduced Claude Opus 4.7 and presented a set of agent products aimed at banks, including a financial-crime tool developed with a payment-services provider. Anthropic’s chief executive reported that the company’s internal security system has identified tens of thousands of previously unknown software vulnerabilities and warned of a six- to 12-month window for governments and large enterprises to patch systems before competing models narrow a claimed lead. The company also reported rapid first-quarter revenue growth on an annualized basis and a pre-IPO valuation that exceeded $1 trillion in April.

On a podcast, biomedical gerontologist Aubrey de Grey and immunology professor Derya Unutmaz described how AI tools are being applied to disease research and aging. Unutmaz predicted many diseases could be addressed within 10 to 15 years and cautioned clinicians that use of AI in medicine may soon be expected as standard practice. De Grey estimated about a 50 percent chance of reaching ‘‘longevity escape velocity’’ — a point where gains in life expectancy outpace aging — by the late 2030s if current research paths succeed.

Near-term tests for the field include whether Pfizer’s AI-designed molecule advances into formal trials, whether Anthropic’s assertions about lead times hold up to independent benchmarking, and whether experimental longevity results observed in animals translate into human outcomes. Each result would affect how quickly AI-driven therapies move from the lab into clinical practice.

Investments and product launches show firms are integrating large models into drug discovery pipelines, packaging agents for regulated industries and applying AI to biological data sets that were previously hard to analyze. Regulators, access to large-scale compute and the limitations of biological data continue to shape development timelines and commercial rollouts.