Anthropic is extending its AI commercialization efforts into the life sciences sector. On Tuesday, the company launched a specialized AI product for scientists and pharmaceutical firms called Claude Science, highlighting protein 3D structure rendering and drug discovery as key application areas. This move aims to open up new revenue streams ahead of its planned initial public offering.
This marks Anthropic's first product specifically designed for scientists and will be available to individual paying users and enterprise subscribers globally. Eric Kauderer-Abrams, head of life sciences at Anthropic, stated in an interview:
"We believe that by working in science, and particularly in life sciences and healthcare, we can have a scaled positive impact on humanity."
Commercially, Anthropic has already signed collaboration agreements with several pharmaceutical companies and acquired the biotech startup Coefficient Bio this past April. Novo Nordisk is using Claude for drug discovery, clinical documentation, and regulatory submissions, while AstraZeneca is leveraging Claude to scale its research and development efforts.
Anthropic also released the latest version of its Sonnet model, Sonnet 5, on Tuesday. The company claims its agent performance is now close to that of its top-tier Opus model, but at a lower cost. This positioning is designed to fill the gap between high-end flagship models and lightweight alternatives, offering enterprise clients a balance of performance and affordability.
While expanding its customer base and upgrading its core technology, Anthropic is also re-evaluating the profitability of its existing key partnerships.
According to a report, Anthropic recently renegotiated parts of its agreement with Amazon.com (AMZN), shifting Claude's pricing model from compute-hours to a token-based billing system. This change in billing logic is expected to increase the cost for Amazon's use of Claude in its shopping, coding, and office AI products.
Drug Discovery as a Primary Focus
Regarding specific applications, Eric Kauderer-Abrams explained that Claude Science can accelerate early-stage drug discovery processes, including tasks like molecular design. He revealed the company's next step is to focus on the clinical stage, while also exploring ways to improve physical lab experiments and studying the potential for integrating robotics.
Kauderer-Abrams positioned the launch of Claude Science as an effort to bridge the gap between the technological frontier and practical application.
"There's quite a gap between what is possible today and what most people can access and actually use. The main goal of this product is to close that gap as much as possible, to let scientists in all fields of science get the most out of the maximum capabilities of the technology that exists."
The importance of AI in the pharmaceutical field continues to grow. Eli Lilly has previously invested in Nvidia chips and, earlier this year, took a stake in Insilico Medicine, a company focused on AI-driven drug discovery.
New Sonnet Version Targets Cost-Effective Applications
Anthropic's simultaneous release of Sonnet 5 is aimed at providing a high-value option. The company states it delivers agent performance nearing its top Opus model but at a significantly reduced cost.
This strategy is intended to occupy the space between premium, high-performance models and more basic, lightweight ones, giving business clients a choice that balances capability with expenditure. Sonnet 5 is available immediately across all plans: it is the default for free and Pro users, and is also accessible to Max, Team, and Enterprise subscribers.
To demonstrate Sonnet 5's practical capabilities, Anthropic compared it using two benchmarks: BrowseComp, which tests agent search abilities, and OSWorld-Verified, which assesses computer operation skills.
Across different levels of compute investment, Sonnet 5 showed consistent improvements over its predecessor, Sonnet 4.6. Opus 4.8 remains the most accurate option in these tests, but Sonnet 5 offers a very competitive level of performance at a much lower price point.
Concurrently, Anthropic's expansion with products like the non-technical agent tool Cowork and the coding assistant Claude Code has raised concerns about job displacement in sectors like software engineering, consulting, and law. The company's commercial growth is now under dual scrutiny from both market forces and regulatory bodies.
Revising Pricing to Boost Margins from Key Partners
As it pursues new customers and technological upgrades, Anthropic is also taking a closer look at the profitability of its central alliances. Reports indicate the company recently renegotiated portions of its deal with Amazon, changing Claude's pricing structure from a compute-hour basis to a token-based system.
This shift in billing methodology is anticipated to raise the costs incurred by Amazon for utilizing Claude across its shopping, programming, and office productivity AI offerings.
Considering Anthropic's recent financial performance, where its inference infrastructure gross margin surged to over 70%, this pricing adjustment is not only a direct move to enhance profitability and optimize its cost structure but also signals to the market the company's strong bargaining power.
In the context of its push toward an IPO, this effort to extract greater profits from a core ally demonstrates that Anthropic has fully transitioned from an early phase of scale expansion to an intense focus on earnings quality and commercial monetization.
Biosecurity Risks Prompt Access Controls
As AI capabilities continue to expand, Anthropic faces escalating security scrutiny alongside its accelerated commercialization efforts.
CEO Dario Amodei noted on Tuesday that societal awareness of the risks posed by AI in biology remains insufficient compared to the cybersecurity domain. He and other AI developers have warned that bad actors could potentially use such models to create novel pathogens or biological weapons.
Amodei stated that only vetted individuals or institutions should be permitted to use powerful biological AI models with potentially dangerous applications.
"Inside pharmaceutical companies, people have been handling dangerous biological materials, they have their own protocols. We can take inspiration from those protocols—perhaps we shouldn't open it up to anyone before verifying who they are."
It is important to note that Claude Science currently operates on existing models and does not possess advanced biological capabilities. Anthropic plans to establish a trusted access program for biological applications, modeled on practices in the cybersecurity field.
IPO Prospects Influence Market Expectations
Anthropic is currently in a critical pre-IPO phase. Reports suggest the company could complete its public listing as soon as this year, with a potential valuation exceeding $1 trillion. The company closed a funding round last month with a pre-money valuation of $900 billion.
On the competitive front, OpenAI has also prioritized scientific applications as a strategic focus. In April, it launched GPT-Rosalind, a cutting-edge reasoning model built specifically for biological research, drug discovery, and translational medicine, concentrating on converting medical research into clinical treatments.
The launch of Claude Science represents a substantive step for Anthropic in transforming AI capabilities into sustainable enterprise revenue. However, maintaining a balance between accelerating commercialization and managing biosecurity risks remains an unavoidable core challenge as it approaches the public markets.
Comments