SandboxAQ Files for IPO Following $4.5B Valuation Milestone

Recently, SandboxAQ, a quantum AI startup, announced that it has raised $450 million in its Series E funding round. This round attracted investments from industry giants such as Google, NVIDIA, and BNP Paribas, bringing the total funding to $950 million and valuing the company at $5.75 billion (approximately RMB 41.93 billion).

SandboxAQ, a company spun off from Alphabet, Google's parent company in 2022, develops artificial intelligence models using quantum computing technology and has successfully developed Large Quantitative Models (LQMs). Its quantum models serve various fields such as life sciences, financial services, and navigation, and can be used to process large numerical datasets, perform complex calculations, and conduct statistical analysis. It is continuously pushing technological boundaries and holds significant potential application value.

Amidst the booming wave of AI development, what unique advantages does SandboxAQ possess that continuously attract investors' attention? And in the field of life sciences, what innovative approaches has it adopted to achieve technological breakthroughs and inject new momentum into the healthcare industry? Let’s decode these one by one.

SandboxAQ is a quantum AI enterprise and is regarded as a pioneer in the field of quantum technology. Since its establishment, SandboxAQ has consistently focused on the research and development of quantum AI solutions, committed to promoting the application of quantum technology in the commercial sector.

Quantum computing, with its properties such as quantum superposition and quantum entanglement, possesses exponential acceleration capabilities when handling complex computational tasks. The integration of quantum computing with AI (i.e., "Quantum AI") can significantly enhance the learning ability and computational efficiency of artificial intelligence.

SandboxAQ was co-founded by former Google CEO Eric Schmidt and former Alphabet employee Jack Hidary. Chairman Eric Schmidt served as both CEO and chairman of Google and held key positions at Google's parent company, Alphabet, wielding significant influence in technology, business, and policy.

CEO Jack Hidary, with a background in neuroscience from Columbia University, has deep expertise in functional brain imaging and artificial neural networks. He is an authority in the field of quantum computing, and his book, *Quantum Computing: An Applied Approach*, has become a key textbook for both undergraduate and doctoral courses. Additionally, he is a serial entrepreneur who successfully led EarthWeb/Dice to an IPO and co-founded VistaResearch, achieving remarkable success in corporate operations and innovation.

In the field of technology and data intelligence, a "Sandbox" is an isolated and restricted runtime environment specifically designed for secure testing and technological innovation. The sandbox thus becomes a cradle for technological innovation, where new software, algorithms, and systems can be repeatedly validated and iterated within this environment.

The name SandboxAQ is derived from this concept. Guided by the "sandbox" spirit, the company focuses on the development of quantum computing models, aiming to build a stable and reliable experimental platform for technological innovation.

In the medical field, SandboxAQ has demonstrated strong innovation capabilities. By leveraging the deep integration of quantum and AI, its quantum model LQMs is continuously accelerating the new drug development process and addressing challenges in drug research. Meanwhile, the magnetocardiography (MCG) device developed based on LQMs is unlocking innovative applications in the field of heart disease diagnosis, improving diagnostic accuracy and precision.

In the traditional drug discovery process, virtual screening is often required to handle billions of compounds. This process relies on large pre-enumerated datasets and requires the use of various methods to evaluate and filter compounds, with only dozens to hundreds of compounds eventually entering the experimental testing and validation stage. Not only is the computational cost high, but it is also impossible to determine whether better ligands exist beyond the dataset, presenting significant limitations.

To tackle the challenges in drug development, SandboxAQ has established the biopharmaceutical molecular simulation department AQBioSim, which relies on large quantitative models (LQMs) to provide solutions that enhance efficiency and quality throughout the entire lifecycle of drug discovery and development.

SandboxAQ Launches Generative AI Application IDOLpro, Effectively Solving This Challenge. As a cloud-based simulation software package from SandboxAQ, IDOLpro is an LQM that combines public data with physics-based simulations. Leveraging AWS infrastructure, it integrates diffusion models with multi-objective optimization to guide and design drug molecule generation.

From a technical perspective, LQMs play a crucial role in the drug molecule screening and generation process. By training on vast amounts of chemical data and molecular structure libraries, they can simulate interactions and dynamic behaviors between molecules, quickly identify target molecules that align with the desired mechanism of action (MoA), and optimize 3D drug molecules within minutes using AI algorithms to create customized molecules, further enhancing the accuracy and efficiency of drug development.

In practical tests, evaluated using benchmark datasets such as crosstocked and Binding MOAD that assess the performance of protein-ligand docking methods, the models generated by IDOLpro achieved 3.4 times the binding affinity of industry-leading methods, with the compounds produced for the first time even outperforming experimentally validated molecules.

In addition, in the screening exploration of drug research and development, SandboxAQ has developed an innovative AQFEP (Advanced Quantum Free Energy Perturbation) technology based on LQMs, which also brings new breakthroughs to drug screening.

Traditional drug screening methods are not fully developed. Although the Relative Free Energy Perturbation (RFEP) method is widely used for binding affinity prediction, its high computational cost and reliance on structurally similar compounds with known activity limit the screening scope and flexibility. On the other hand, the Absolute Free Energy Perturbation (AFEP) method, in theory, can identify hit rates more accurately but suffers from slow processing speeds, making it difficult to meet the practical needs of rapid docking and scoring in virtual screening.

AQFEP technology effectively overcomes the aforementioned drawbacks by combining active learning with a rigorous physics-based scoring function, fully leveraging the advantages of RFEP and AFEP while maintaining both efficiency and accuracy. This technology does not require reference molecules, and within an integrated workflow, AQFEP can "unlock" molecules at the forefront, enabling highly efficient screening of large chemical libraries and significantly improving the hit rate of virtual screening. In screening efforts for neurodegenerative diseases and cancer, AQFEP demonstrates exceptional performance in hit identification and lead optimization, offering extremely high application value.

Currently, SandboxAQ has established partnerships with two top academic research institutions and expanded its collaborations with large biopharmaceutical companies, including AstraZeneca, Sanofi, and the University of California, San Francisco, to identify new biomarkers and optimize the clinical development of drugs in the pipeline.

In recent years, cardiovascular diseases have continued to threaten global human health. According to statistics from the World Health Organization, about 18 million people die from cardiovascular diseases each year, accounting for more than 30% of the total global deaths.

In the field of cardiovascular disease diagnosis, the traditional electrocardiogram (EKG) has been used for over 150 years as an important tool for monitoring the heart's electrical activity, playing a critical role in diagnosing cardiac rhythm abnormalities. However, its detection of electrical pulse signals is susceptible to interference from body tissues, affecting diagnostic accuracy. In the United States alone, 8 million emergency patients visit clinics annually due to chest pain, yet less than 5% are successfully diagnosed through standard tools like EKGs.

In response to this challenge, SandboxAQ has launched CardiAQ, a groundbreaking magnetocardiography (MCG) device. This device detects cardiac magnetic signals using high-performance sensors and employs large quantitative models (LQM) to effectively eliminate electromagnetic interference, overcoming the limitations of traditional EKG. It can capture details of cardiac electromagnetic activity more effectively, helping to detect abnormal patterns and potential heart disease conditions, providing doctors with more comprehensive diagnostic support.

In practical applications, CardiAQ is convenient and efficient, requiring no dedicated space, cooling, or shielding. Measurements can be completed in just a few minutes, making it highly suitable for instant bedside testing. This non-contact innovative diagnostic method, powered by the deep integration of high-performance sensors and advanced AI technology, transforms the diagnostic approach for heart diseases. It not only provides patients with more timely and accurate medical services but also holds the potential to enhance the prevention and treatment of heart disease globally, saving lives.

Currently, two well-known research hospitals are using the CardiAQ device for clinical studies to test its functionality. At the same time, SandboxAQ has entered into a technical collaboration with Mayo Clinic to jointly explore innovative applications of artificial intelligence-driven magnetocardiography (MCG) in the field of heart disease diagnosis, aiming to improve the level of cardiac diagnosis.

In recent years, the AI pharmaceuticals market has shown vigorous development, with its scale continuously expanding and strong growth momentum. According to data from Research And Markets, the global AI pharmaceuticals market size reached $1.04 billion in 2022, and it is expected to reach nearly $3 billion by 2026, with an average annual compound growth rate as high as 30%. By 2032, the global AI drug research and development market size is projected to exceed $20 billion, indicating a very broad future prospect.

An increasing number of companies are entering the AI pharmaceuticals field, including startups like XtalPi and Insilico Medicine that focus on AI-driven drug development, as well as traditional pharmaceutical companies such as Hengrui Medicine and CSPC Pharmaceutical Group, which are actively embracing AI technology through strategic partnerships and equity investments to accelerate the innovation of drug development.

In this competitive landscape, SandboxAQ has demonstrated unique advantages and development potential in the field of drug research and development through the deep application of LQMs quantitative large models. With forward-thinking strategic thinking, cutting-edge technical means, and an avant-garde founding team, SandboxAQ has continuously gained the favor of many investors, carving out its own unique path of development.

In a broader sense, the rise of SandboxAQ aligns perfectly with the current global technological fervor for quantum AI research—countries like China and the U.S. are actively investing, and tech giants such as Baidu, Google, and Microsoft are all entering the field. The unique properties of quantum AI, such as superposition and entanglement, can indeed address problems that traditional AI struggles to solve. For instance, quantum AI can simultaneously process multiple possibilities in optimization problems, rapidly solving complex tasks that would traditionally take years or even decades to compute. Potential application scenarios include logistics, transportation, chip manufacturing, energy distribution, and drug development.

The potential of quantum AI is a technical entry point, but its significance goes beyond that. It may alter our understanding of intelligence, creativity, and the future. In the more distant future, could quantum AI usher in an entirely new era of AI-driven drug discovery? We will wait and see.