SandboxAQ – Weekly Recap

SandboxAQ is a quantum- and AI-focused software company operating at the intersection of high-performance computing, security, and advanced materials research. This weekly summary reviews a series of updates that underscore the company’s push into physics-based AI, quantum chemistry, AI security, and energy-related applications.

During the week, SandboxAQ highlighted research conducted with NVIDIA showcasing quantum chemistry simulations running on NVIDIA’s latest Blackwell GPUs. The work, featured via an ArXiv paper, indicates that high-precision quantum chemistry workloads can achieve 90–95% GPU utilization on hardware originally designed for large language models. This suggests that Blackwell GPUs could be repurposed for demanding scientific simulations, potentially lowering infrastructure costs and broadening access to SandboxAQ’s quantum chemistry and materials modeling tools in industries such as pharmaceuticals, chemicals, and advanced materials. While no commercial terms were disclosed, the collaboration raises SandboxAQ’s profile within the GPU-based computing ecosystem.

The company also amplified its focus on AI security by promoting an AI Security Posture Management (AI-SPM) checklist aimed at addressing growing gaps in enterprise AI governance. Citing that a large majority of organizations run AI in production but most have not performed comprehensive AI security assessments, SandboxAQ is positioning its tools to help enterprises locate “shadow AI,” improve visibility across code and pipelines, and align AI deployments with compliance and risk-management requirements. This reinforces the company’s push into AI security and governance, an area likely to become increasingly important as regulators and enterprises tighten controls around AI usage.

Strategically, CEO Jack Hidary’s participation in a “Beyond LLMs” fireside session underscored SandboxAQ’s emphasis on Large Quantitative Models (LQMs), or physics-based AI systems geared toward deterministic, high-precision problems. The company framed the next phase of AI as shifting from content generation to simulation and modeling, with applications in drug discovery, energy, materials science, and national security. This focus aligns SandboxAQ with mission-critical and regulated sectors where differentiated capabilities and long-term contracts are possible, though development timelines and execution risks may be significant.

Finally, SandboxAQ gained additional industry visibility through recognition of an article on AI’s role in battery design as one of Battery Power Online’s top stories of 2025. The coverage highlights the company’s engagement in AI-driven battery and materials innovation, which is drawing attention from automotive and energy-storage markets. While largely reputational at this stage, such thought leadership can support future partnerships and pilot projects.

Overall, the week’s developments reinforce SandboxAQ’s positioning in deep-tech AI and quantum-related software, spanning scientific simulation, AI security, and energy-focused innovation, with the potential to strengthen its long-term commercial prospects if these initiatives translate into scalable customer adoption.