Mitsui & Co. Subsidiary Improves the accuracy of AI models from 65% to 74% through collaboration on confidential data. It is expected to significantly reduce the time and cost associated with Drug Discovery

Fortanix® Inc., a leader in data-first cybersecurity and pioneer of Confidential Computing, today announced that Xeureka Inc., a subsidiary of Mitsui & Co. established to engage in early drug discovery research using a state-of-the-art technology, has successfully leveraged Fortanix and NVIDIA Confidential Computing technologies to significantly improve the accuracy of AI models, and is expected to reduce the time and costs associated with drug discovery.

The Pharmaceutical Industry Faces Significant Hurdles

Established in 2021 by Mitsui & Co., one of Japan’s largest trading companies, Xeureka provides computational and AI drug discovery support services for pharmaceutical and biotech companies. The organization aims to speed up the discovery of new drugs that can save lives in the face of significant challenges that hamper the pharmaceutical industry: research and development costs are high, often averaging $1 billion to $2 billion per new drug, and development times can average between 10 and 20 years. Additionally, success rates are infamously low, often below 1% on average.

Xeureka aims to use advanced technology across all aspects of drug development to improve both the efficiency and effectiveness of new drug research. AI technologies are core to its mission, requiring extensive data sets to power the AI models. However, the data used in pharmaceutical research often includes highly confidential, sensitive, and personally identifiable data, while stakeholders must comply with various global privacy regulations. To successfully leverage this data in full, Xeureka needed a way to pool data from multiple stakeholders while maintaining its confidentiality during use in AI models to power research.

Fortanix and NVIDIA Confidential Computing Powers Accelerated Drug Discovery

Xeureka partnered with Fortanix and NVIDIA to demonstrate the ability to integrate disparate datasets from multiple sources while maintaining the security and confidentiality of the original sources. Specifically, Fortanix provided the control plane to enable security policy management and orchestrated the use of highly confidential data in a secure environment. Fortanix’s Confidential Computing capabilities include:

• Data sharing and access control for fast and collaborative data processing to ensure privacy for multiple parties without revealing sensitive information.
• Modern key management to ensure unified encryption, regardless of the location of the data across hybrid multi-cloud environments.
• Workflow management to define, manage and track the data collaboration process flow and ensure privacy and security guardrails.

In addition, NVIDIA’s secure enclaves with Confidential Computing-enabled NVIDIA H100 Tensor Core GPUs provided the execution environment to run the data, algorithms and AI models, guaranteeing that highly sensitive data assets remained secure and private while in use.

A Significant Leap for AI Model Accuracy

By enabling confidential data collaboration, Xeureka improved AI model accuracy from 65% to 74% in certain parameters, which will ultimately reduce the time and costs associated with drug discovery. This is also proof of how AI can enhance the effectiveness and efficiency of new drug development while improving the lives of millions of consumers across the globe.

"This is a perfect example of the power of Confidential Computing -- rapidly processing extremely large volumes of real patient data required by AI models with the highest level of privacy and security possible," said Anand Kashyap, CEO and co-founder of Fortanix. "We are proud to partner with Xeureka in its quest to improve drug discovery, and we look forward to more advancements in the years to come."

"NVIDIA H100 Tensor Core GPUs configured in Confidential Computing mode have proprietary hardware-based cryptographic engines, firewalls, and remote verification to ensure the integrity of the TEE and validate that their confidential workloads are protected while in use on the GPU – in the most performant way possible today,” said Daniel Rohrer, VP of Software Product Security, Architecture and Research at NVIDIA. “NVIDIA technology uniquely provides companies with the ability to speed up drug discovery and development with a high level of security without compromising performance."

"AI is foundational to the initiatives we've launched to tangibly transform pharmaceutical research and development, but models require extremely large amounts of data that often contain sensitive or personal information," said Hiroki Makiguchi, CTO at Xeureka. "Confidential Computing from Fortanix and NVIDIA essentially alleviates those privacy and security concerns while also improving model accuracy, which will prove to be a win-win situation for the entire industry.”

About Fortanix

Fortanix is a global leader in data security. Our unified platform, powered by Confidential Computing, future proofs your data security and makes it simple for organizations to discover, assess, and remediate cybersecurity risks. Our post-quantum-ready cryptographic solutions help organizations thwart cloud and AI data exposure threats and enable secure innovation. Trusted by leading brands and government agencies, Fortanix empowers enterprises to secure their most sensitive data at rest, in motion, and in-use and remain compliant with regulations worldwide. For more information, visit www.fortanix.com.

View source version on businesswire.com: https://www.businesswire.com/news/home/20240822160236/en/