Novyte Materials uses artificial intelligence to optimize research and development
This week, Mumbai-based advanced technology startup Novyte Materials revealed its advances in using artificial intelligence (AI) to revolutionize materials research and development (R&D), promising a significant transformation in the chemical industry.
The development of new materials is historically a slow and repetitive process. Months are often spent by scientists reading articles, formulating experiments, and testing compositions, often with no guarantee of a commercially viable product, as promising discoveries may fail to achieve economic production.
Novyte Materials is betting that AI can shorten industrial R&D time from years to a few months by automating repetitive scientific tasks. The startup focuses not only on discovering new materials, but also on how active AI can solve the complex problem of optimizing these materials in the real world, a crucial hurdle to commercialization.
Founded in 2025 by chemical engineer Ajaz Khan and incubated at Mumbai’s Institute of Chemical Technology (ICT), Novyte develops AI models that help chemical and materials companies optimize existing products, discover new formulations, and automate industrial R&D.
Materials science involves the creation of new materials with specific properties and functions to meet technological and industrial demands. The potential of AI in this sector is vast and disruptive.
According to research firm Market.us, the global AI market for materials discovery is expected to reach US$5.5 billion by 2034, with compound annual growth (CAGR) of 26.4%, starting from US$536 million in 2024.
More on this story: Laser microchip enables simple synthesis of gold nanoparticles in organic solvents
Globally, AI-driven materials discovery has attracted significant investment, with startups such as CuspAI, Orbital Materials and Citrine Informatics developing solutions for new materials design and enterprise R&D.
While many focus on discovery, Novyte prioritizes automating the formulation optimization process and enabling companies to commercialize materials, arguing that this area is equally crucial.
Materials Optimization: Novyte’s approach to artificial intelligence
Ajaz Khan, from Novyte, highlights that industrial R&D work goes far beyond the discovery of innovative materials, mainly involving the continuous improvement of existing formulations.
Most chemical manufacturers spend years perfecting formulations, changing ingredients, replacing hazardous substances or improving performance, a process that consumes time and resources.
Novyte’s platform, called Novyte Q, seeks to automate much of this process, using AI models with extensive chemistry knowledge to understand literature and materials properties.
Khan explains that the platform analyzes hundreds of articles, suggesting formulations and experimental paths. It uses reinforcement learning and density functional theory, a standard quantum chemistry technique, to evaluate the chemical stability of proposed materials.
Learn more: Researchers identify new quantum effect that drives energy transfer
According to the founder, Novyte’s AI agent examines between 500 and 1,000 scientific sources in each round, generating recommendations after several iterations.
The company states that the platform ranges from literature review (Technological Readiness Level – TRL 1) to real-time optimization of laboratory experiments (TRL 7). It is even possible to reverse engineer competing materials with just data sheets and properties, says Khan.
The platform is sold annually to companies, being implemented on their own GPU hardware infrastructure, which guarantees the privacy of research data.
A specialty chemicals manufacturer, using Novyte, replaced a hazardous additive and achieved desired specifications in 40 tests versus 200 internal tests, reducing lab time by 58%, Khan reported.
While Novyte Q is the primary focus, the company plans to expand into materials discovery with custom AI via the Psi platform. The long-term strategy is to create a “synthesis layer” to turn AI-generated candidates into manufacturable products, Khan said.
See Also
Learn how to identify intrusions on WhatsApp and avoid financial fraud
PlayStation 5 owners gain lifetime access to three great games
Russian Islam Makhachev dominates Ian Machado Garry and wins main fight at UFC 330
In June, Novyte reached an agreement with Chemvera Specialty Chemicals, a global specialty chemicals manufacturer, to develop, manufacture and commercialize a high-value specialty chemical for the polymer industry. Chemvera will be in charge of production and distribution.
Novyte also collaborates with leading companies in the chemical and materials sectors, such as Manipal Specialty Chemicals and Primacy Industries, applying its platform to solve formulation challenges.
Asked about the readiness of AI models, Khan explains that the challenge is not generating millions of “stable” structures, but the failure of these materials in real manufacturing conditions or the impossibility of synthesizing them.
On the same topic: Scientists manage to film the ultrafast movement of electrons in attoseconds, revealing secrets of the quantum world to humanity
The ability to synthesize is often overlooked; models deliver complex structures without viable production routes. Prediction of properties is superficial to industrial needs, such as mechanical, thermal and corrosion behavior, without a mechanism to integrate experimental failure results.
Novyte’s difference lies in its stability, integration of physical principles and ability to synthesize materials. Human intervention remains crucial to define problems, verify route viability and make final decisions, with AI accelerating the search.
Opportunities and challenges in the industrial R&D market
Six months after commercial launch, the company has gained a paying customer base in the “high single digits,” with trials underway, but has not released revenue figures.
Novyte takes about two weeks to implement, and customers outperform their own traditional, literature-based research processes in about six weeks.
Several companies have already expanded their deployments, increasing the number of licenses by 30% to 50%, starting second R&D projects in the first year and negotiating multi-year agreements.
In December 2025, early-stage investment fund Thiea Ventures, focused on AI and deeptech, led a ₹4.5 crore pre-seed round in the startup, with participation from Sandesh Paturi, co-founder of Venwiz, and Niharika Jain, director of Chemvera. Novyte also received a $40,000 grant for computing and access to NVIDIA’s GPU infrastructure via the Inception program.
Full coverage: Latest News (EN)
In India, direct competition is limited, unlike the global scenario with Aionics, Cusp AI and Orbital Materials. Other Indian startups operate in different niches, such as Whizzo (technological fabrics for fashion) and RF Nanocomposites (radar absorbent materials).
Global companies have capital advantage and supply chain ecosystems. CuspAI, for example, obtained US$400 million in funding, valuing it at US$2.6 billion. Others, such as Citrine Informatics, Lila Sciences ($200 million seed) and Periodic Labs ($200 million seed, $1 billion valuation), have also attracted massive investments, totaling more than $1.3 billion in the last two years for materials AI startups.
Large AI companies also enter the sector; Anthropic, for example, is investigating how its Claude model can analyze data from nuclear magnetic resonance (NMR) machines, used by chemists to study the composition and structure of new materials.
The weak competition in India, coupled with the advanced nature of the industry, creates a huge growth opportunity for Novyte, despite the inherent challenges.
Bryce Meredig, founder of Citrine, questions the financial return on widespread AI materials discovery. Novyte’s Khan responds that this is the “valley of death” for many programs, so the company built a three-tier structure: the AI Scientist (discovery and optimization), the quantum physics and chemistry engine (validation), and the synthesis workflows (translation to real procedures).
Finding promising material is just the beginning; the real challenge lies in successfully synthesizing and integrating it into existing manufacturing processes.
