| Service | Cost | Details | Source |
|---|---|---|---|
| Research Access | Custom enterprise contract | AI materials discovery platform for scientific research organizations | — |
| Partnership Access | Custom quote | Collaborative R&D partnerships with semiconductor manufacturers and research institutions | Company announcements and funding reports |
| Feature | Periodic Labs | Google DeepMind GNoME | Berkeley A-Lab | Materials Project |
|---|---|---|---|---|
| Core Functionality | Autonomous labs + AI scientists | Materials prediction | Autonomous materials synthesis | Materials database |
| Physical Experimentation | Yes (self-driving labs) | No | Yes | No |
| Proprietary Data Generation | Yes | No | Partial | No |
| Superconductor Discovery | Target focus | Demonstrated | Demonstrated | |
| Lab Automation | Yes | No | Yes | No |
| API Availability | N/A (early stage) | No | Research access | Yes (open) |
| Enterprise Partnerships | Yes (semiconductor firms) | Google internal | Academic | Open access |
| Pricing | Custom enterprise | Research only | Research only | Free/open source |
| Founded | 2024 | 2010 | 2022 | 2011 |
| Team Expertise | Ex-OpenAI/DeepMind | DeepMind researchers | Berkeley researchers | Academic consortium |
The difference in differentiation is based upon the type of labs in which each company conducts closed-loop autonomous labs to produce proprietary experimental data and in contrast, GNoME utilizes solely computational predictions without a laboratory for conducting experiments.
For enterprises that require experimental validation of materials science R&D, use Periodic; for pure computational screening of molecules, use DeepMind.
Although both companies are pursuing autonomous experimentation, Periodic has greater financial resources (as evidenced by its $300 million in funding vs. the grants received by A-Lab) and also has an industrial partner orientation as opposed to the more academic-oriented A-Lab.
Use Periodic when looking to deploy at scale commercially; for proof-of-concept research utilize A-Lab.
Leader in computational chemistry with revenues exceeding $500 million+ vs. the early-stage AI+robotics approach being pursued by Periodic. In addition to having a large established customer base for enterprise customers, Schrodinger does not have autonomous physical experimentation.
If you need validated simulation workflows, use Schrodinger; if you want to accelerate your experimental process using AI, use Periodic.
Leader in AI-based drug discovery vs. Periodic’s focus on materials science. Similar end-to-end AI approaches exist within both companies however they differ in their respective domains - Exscientia has progressed further in terms of clinical trials whereas Periodic has focused on validating materials.
Each of these companies are leaders in their specific domain: Exscientia is a leader in AI-based pharmaceuticals while Periodic is a leader in AI-based materials.
Cloud serverless serverless. What’s even better than a serverless?
Use the api keys provided in your members dashboard for this. “Add your api key in your request headers!
There is no server provided by periodic no need for you to sign up and sign up register. Instead, you can buy your computer from a store.
TabPFN is therefore a machine learning model for tabular data. Periodic Labs provides cloud hosted access to tabpf n cloud model via our api.
Yes. There is a quickstart guide you will find useful to get started. quickly through setup authenticating and use our documentation site itself.
Engineers and researchers building ML AI applications using tabpf n want to get started without managing the infrastructure themselves.
Periodic Labs acts as an API service offering cloud-based access to the TabPFN artificial intelligence model. There is limited publicly available information about the company beyond their API documentation and CB Insights entry regarding AI and data service operation.
Limited - publicly available information is sparse. Primary sources are official API documentation and minimal third-party references. Company financial data, pricing details, and customer information are not publicly disclosed.
Periodic Labs lists on CB Insights as operating in the categories of AI and data services. Periodic Labs offers API services which allow developers to use cloud based access to machine learning models.
The primary product from the company is a serverless API service that allows users to access the TabPFN, a very specialized machine learning model. Therefore they can be classified as a provider of inference-as-a-service, as opposed to a developer of frontier AI research lab technology.
Hypothetical AI systems that predict how the world may be and generate scientific hypothesis
Automated experimental design in conjunction with robotic laboratory systems
Concentration on the physical sciences including semiconductors and high temperature superconductors
Analysis of the Results of Experimental Data (Including Negative) to Speed-Up Research & Development (R&D)
Integration of AI Frontier Models with Simulation for Closed-Loop Learning
Advanced Reasoning for Chemistry, Physics and Material Science Applications
Chip Heat Dissipation Analysis and Iteration of Experimental Data
Development of New High-Temperature Superconductors and Novel Materials
Synthesis of Powders, Robotic Experimentation and Proprietary Generation of Data
Closed Loop Systems from Hypotheses Generation to Execution of Experiments
Partnerships for Advanced Materials in Mission Critical Domains
Materials R&D to Support Breakthrough Energy Applications
| Evaluation Dimension | Measurement Basis | Industry Standard | Assessment Frequency |
|---|---|---|---|
| Scientific Discovery Performance | Real-world experimentation success rate + materials discovered | Physical validation beyond digital benchmarks | Continuous via autonomous labs |
| Experimental Throughput | Experiments per day + data volume generated (GB/experiment) | Proprietary physical data generation | Real-time robotic lab monitoring |
| Data Moat Strength | Proprietary experimental data volume + negative results captured | Non-scrapable physical world data | Compounding via continuous experimentation |
| Industry Partnership Maturity | Semiconductor, space, defense deployments | Revenue-generating research applications | Per partnership deployment |
| Founding Team Pedigree | ChatGPT/GNoME/Operator/MatterGen creators + Nobel advisors | Elite AI + physical science convergence | Static (team composition) |
| Physical Infrastructure | Autonomous robotic laboratories operational status | Hardware-software closed loop validation | Continuous operation monitoring |
| R&D Timeline Compression | Decades to months for materials discovery | End-to-end hypothesis-to-validation measurement | Per scientific domain |
