A new marketplace listing turns SandboxAQ's Large Quantitative Models into a one-click resource for researchers, and signals how frontier scientific AI is starting to move the way software once did.

by Kasun Illankoon, Editor in Chief at Tech Revolt

[For more news, click here]

For years, the promise of AI-accelerated science has carried a quiet asterisk. The physics-based models capable of predicting how a drug binds to a disease-related protein, or how a catalyst behaves on an industrial surface, have existed for some time. But using them has typically required specialized infrastructure, in-house computational expertise, and often a direct relationship with whoever built the model.

This week, that asterisk got noticeably smaller. SandboxAQ, the AI and quantum technology company that spun out of Alphabet in 2022, announced that its Large Quantitative Models, known internally as LQMs, will become accessible through Google Cloud's Marketplace starting in the third quarter of 2026. The move turns tools once reserved for specialists into something a researcher can reach from inside the conversational AI interface they already use every day, without writing specialized code or standing up new infrastructure.

The Distance Between a Lab and a Marketplace

What separates an LQM from a general-purpose chatbot is the data underneath it. These models are trained on real-world laboratory measurements and grounded in scientific equations, rather than on text scraped from the internet. SandboxAQ has built them specifically for what it calls the global quantitative economy, an umbrella spanning biopharma, energy, advanced materials, and financial services that the company estimates is worth more than $50 trillion. Putting that kind of model inside a marketplace listing is less a technical feat than a distribution decision, but it is the kind of decision that determines whether rigorous science reaches the researchers who need it or stays locked inside a narrow circle of specialists.

Jack D. Hidary, CEO of SandboxAQ, framed the announcement in exactly those terms.

"Bringing our LQMs to Google Cloud’s Marketplace will put the rigor of first-principles science directly into the hands of every researcher, in the tools they already use," he said. "Pairing the reasoning of a frontier model such as Gemini with the quantitative precision of our LQMs is a powerful combination."

Why Google Cloud Wanted In

For Google Cloud, the listing is a way to deepen its footprint in healthcare and drug discovery without building physics-based modeling capability from scratch. Enterprise buyers, particularly in pharmaceutical and biotech research, have been asking cloud providers for tools that go beyond generic machine learning and offer scientifically grounded outputs they can trust.

Brian Goldstein, Vice President of Strategic AI and ISV at Google Cloud, connected the partnership directly to one of the more stubborn problems in medicine.

"Bringing SandboxAQ's Large Quantitative Models to GCP Marketplace is one of the ways we are empowering healthcare researchers to accelerate drug discovery and solve one of the most critical gaps in healthcare today," he said.

The first LQM arriving on the marketplace, expected in the third quarter, is AQCat, built for materials and catalyst discovery. It targets adsorption energy calculation, a measure of how strongly molecules bind to a catalyst surface and typically the first, most expensive step in identifying viable candidates. By letting researchers rapidly screen and prioritize options before committing lab resources to full evaluation, AQCat aims to deliver accuracy comparable to gold-standard methods at a fraction of the time and cost. The stakes are broader than any single lab. Catalysts underpin more than 90 percent of commercially produced chemical products, with direct consequences for green hydrogen, sustainable aviation fuel, fertilizer production, and plastics recycling.

Following behind it is AQPotency, aimed squarely at drug discovery. The model lets researchers computationally identify and rank the most promising binders, the molecules that attach to a specific disease-related target, at a throughput that would be impossible to replicate through traditional lab screening alone. The market context makes clear why that matters. The global drug discovery market was valued at roughly $112 billion in 2025 and is projected to reach about $187 billion by 2034, according to Straits Research, a trajectory that leaves little room for the slow, expensive candidate-screening processes that have defined the industry for decades.

A Pattern, Not a One-Off

The Google Cloud listing is not SandboxAQ's first move to put its models inside tools researchers already rely on. The company recently integrated its LQMs with Anthropic's Claude, and the Google Cloud announcement explicitly builds on that earlier step. Taken together, the two integrations point to a deliberate strategy: rather than betting on a single distribution channel or a single large language model, SandboxAQ is positioning its scientific models to sit alongside whichever frontier AI system a researcher, hospital system, or materials lab happens to be using. It is a channel strategy borrowed from enterprise software, applied to a category of AI that has, until recently, behaved more like a specialized scientific instrument than a product.

Where Berlin and the Gulf Fit In

The timing lands during a period when enterprise AI distribution is being tested on a genuinely global stage. GITEX AI EUROPE 2026, held in Berlin from June 30 to July 1, drew more than 800 companies and roughly 500 investors from over 100 countries to debate exactly the kind of question this announcement answers in practice: how does sovereign-grade, scientifically rigorous AI reach the enterprises and researchers who need it without forcing them to rebuild their own infrastructure.

That question resonates well beyond Europe. Gulf governments, particularly in Saudi Arabia and the UAE, have poured billions into sovereign AI infrastructure with life sciences and advanced materials named as priority sectors, and enterprise buyers across the region have shown a growing appetite for AI tools that arrive pre-validated rather than custom-built. A marketplace listing that lowers the barrier to physics-grounded AI is the kind of development that travels easily from Mountain View to Berlin to Riyadh, because the underlying need, trustworthy computational science without a specialist's overhead, is universal.

What This Means for the Next Researcher

The broader significance of this announcement has less to do with any single product than with a shift in how frontier scientific AI reaches the people who use it. For most of the past decade, running a physics-grounded model meant either building deep in-house computational expertise or partnering directly with the company that built it. A marketplace listing collapses that distance. It means a biotech startup in Boston, a materials lab in Riyadh, or a university research group in Toronto can, in principle, access the same rigorous, lab-grounded modeling capability as a well-resourced pharmaceutical giant, through infrastructure they already pay for and tools they already know how to use. That kind of quiet democratization rarely makes headlines on its own. But it is precisely the sort of structural change, the same one that turned specialized software into commodity infrastructure, that tends to compound over a decade into something researchers look back on as a genuine turning point.

Related Articles:

What 500 Business Leaders Got Wrong About Quantum AI in 2025, and What They Think Now

Banks Are Starting to Finance AI Infrastructure Like They Once Financed Airports

AHOY Establishes North American HQ, Partners With Mila Labs, and Launches Canada's First Physical AI Lab