SandboxAQ
SandboxAQ helps enterprises solve physics and chemistry problems using quantitative AI models.
SandboxAQ develops Large Quantitative Models (LQMs)—physics-aware AI systems that solve complex problems in drug discovery, cybersecurity, and GPS-denied navigation. The company spun out from Alphabet in 2022 with six years of internal R&D and serves Global 1000 enterprises through cloud-delivered software subscriptions. Unlike general-purpose LLMs, SandboxAQ's models incorporate quantitative reasoning for measurable real-world outcomes in regulated, calculation-intensive domains.
Problem solved
Enterprises lack AI systems that incorporate physics-based reasoning and quantitative logic for complex problems in drug discovery, cryptographic vulnerability, and GPS-denied environments.
Target customer
Global 1000 enterprises in pharma, biotech, telecom, healthcare, and defense sectors requiring advanced computational modeling for drug discovery, cryptography management, and resilient positioning.
Founders
J
Jack Hidary
CEO
Founded EarthWeb (1995, IPO 1998), co-founded Vista Research (2001, acquired by McGraw-Hill 2005), led quantum technology initiatives at Alphabet (2016-2022). Studied philosophy and neuroscience at Columbia University.
E
Eric Schmidt
Chairman
Former CEO of Google, long-term investor and board chairman of SandboxAQ since spinout from Alphabet.
Funding history
Seed
Undisclosed
March 2022
Led by Breyer Capital, Eric Schmidt
· T. Rowe Price, Guggenheim Investments, Marc Benioff, David Siegel, Section 32, Parkway Venture Capital
Series A
$500M
February 2023
Led by Unknown
· Unknown
Series D
$300M+
December 2024
Led by Fred Alger Management
· T. Rowe Price, Mumtalakat, Parkway Venture Capital, Breyer Capital, Rizvi Traverse, Yann LeCun, IQT
Series E
$450M+
April 2025
Led by Google, NVIDIA
· Ray Dalio, Horizon Kinetics, BNP Paribas, Breyer Capital, T. Rowe Price, Rizvi Traverse
Total raised:
$950M+
Industries
Pricing
Subscription-based with usage pricing. Customers access Large Quantitative Models through cloud platforms (Google Cloud) with monthly/annual fees plus compute-based charges. Enterprise pricing not publicly disclosed.
Notable customers
AstraZeneca, Sanofi, Vodafone Business, SoftBank Mobile, Mt. Sinai Health System, Wix
Integrations
Google Cloud, NVIDIA compute infrastructure, enterprise cloud platforms
Tech stack
jQuery (JavaScript libraries)
Open Graph
HTTP/3
Linkedin Insight Tag (Analytics)
Hotjar (Analytics)
Google Analytics (Analytics)
reCAPTCHA (Security)
HSTS (Security)
Google Workspace (Email)
jsDelivr (CDN)
Cloudflare (CDN)
Marketo Forms (Marketo Forms)
Reddit Ads (Advertising)
Linkedin Ads (Advertising)
Google Tag Manager (Tag managers)
Webflow (Page builders)
Greenhouse (Recruitment & staffing)
Website
Competitors
OpenAI / GPT-4
General-purpose language models lacking physics-aware reasoning; SandboxAQ specializes in quantitative, equation-based domains like chemistry and physics.
Schrodinger
Focuses narrowly on computational chemistry; SandboxAQ operates across drug discovery, cybersecurity, and navigation with physics-based models.
Quantum computing startups (IonQ, Rigetti)
Pure quantum hardware plays; SandboxAQ integrates AI with quantum-inspired algorithms for enterprise software delivery.
Traditional cybersecurity vendors (Palo Alto Networks, CrowdStrike)
Legacy signature-based security; SandboxAQ's AQtive Guard uses AI for cryptography management and post-quantum threat preparation.
Why this matters: SandboxAQ represents a rare combination: a Alphabet spinout backed by $950M+ from top-tier investors (including Google, NVIDIA, Ray Dalio, Marc Benioff, and Yann LeCun), led by a serial entrepreneur (EarthWeb IPO, Vista Research acquisition), with defensible IP from six years of Google R&D. The company's focus on quantitative AI for measurable real-world problems in drug discovery and security—rather than chasing general-purpose LLMs—positions it as a contrarian AI play with clear enterprise ROI.
Best for: Large enterprises in pharma, biotech, telecom, and defense needing AI-powered solutions for drug discovery acceleration, cryptographic modernization, and GPS-denied navigation.
Use cases
Drug Discovery Acceleration
Pharmaceutical companies use AQBioSim and AQAffinity to simulate molecular interactions and reduce drug development timelines from years to months. The physics-based models predict binding affinity and chemical behavior with measurable accuracy, enabling faster candidate screening and reducing costly late-stage failures.
Post-Quantum Cryptography Readiness
Enterprises use AQtive Guard to discover, inventory, and modernize cryptographic systems before quantum threats materialize. The AI-driven platform identifies vulnerable algorithms across infrastructure and recommends migration strategies, preparing organizations for the post-quantum era.
GPS-Denied Navigation
Military, defense, and autonomous vehicle operators deploy AQNav for resilient positioning in environments where GPS is jammed or unavailable. The sensing technology provides reliable localization in tunnels, urban canyons, and contested environments without external signals.
Alternatives
Schrodinger
Narrower focus on computational chemistry and drug discovery; SandboxAQ spans multiple quantitative domains including cybersecurity and navigation.
Recursion Pharmaceuticals
Uses high-throughput screening and machine learning; SandboxAQ emphasizes physics-aware quantitative models with more interpretable equations-based reasoning.
Traditional quantum computing platforms (IBM, IonQ)
Hardware-focused quantum platforms requiring specialized expertise; SandboxAQ delivers pre-built AI models via accessible cloud software for specific enterprise problems.
FAQ
What does SandboxAQ do? +
SandboxAQ develops Large Quantitative Models (LQMs)—AI systems that incorporate physics and mathematical reasoning for complex problem-solving. The company serves enterprises through three main products: drug discovery acceleration (AQBioSim, AQAffinity, AQChemSim), cryptography management (AQtive Guard), and GPS-denied navigation (AQNav). Models are delivered as cloud-based subscriptions.
How much does SandboxAQ cost? +
SandboxAQ uses subscription-based pricing with usage components. Customers pay monthly or annual subscription fees plus compute charges based on how much cloud computing they use. Enterprise pricing is custom and not publicly disclosed; contact sales for quotes.
What are alternatives to SandboxAQ? +
Schrodinger (computational chemistry), Recursion Pharmaceuticals (ML-driven drug discovery), quantum platforms like IBM Quantum and IonQ (for quantum computing), and traditional cybersecurity vendors. None offer the same breadth across drug discovery, cryptography, and navigation with physics-aware quantitative models.
Who uses SandboxAQ? +
Global 1000 enterprises in pharma, biotech, telecom, healthcare, and defense. Named customers include AstraZeneca, Sanofi, Vodafone Business, SoftBank Mobile, Mt. Sinai Health System, and Wix. The company targets organizations needing advanced computational solutions for regulated, calculation-intensive domains.
How does SandboxAQ compare to OpenAI/ChatGPT? +
ChatGPT and similar LLMs excel at language, reasoning, and general knowledge but lack physics-aware, equation-based reasoning. SandboxAQ's Large Quantitative Models are purpose-built for quantitative domains like chemistry, physics, and cryptography where measurable accuracy and mathematical reasoning matter. LLMs cannot reliably predict molecular binding affinities or design quantum-resistant algorithms.
Why did Google and NVIDIA invest in SandboxAQ's Series E? +
Google spun out SandboxAQ from Alphabet in 2022 and continues investing, recognizing the value of enterprise AI for regulated domains. NVIDIA's participation signals alignment with the company's compute-intensive large quantitative models delivered through cloud infrastructure, where NVIDIA chips power the underlying calculations.
Tags
quantum AI
large quantitative models
drug discovery
cryptography
cybersecurity
physics-aware AI
enterprise software