The screen in a German foundry office shows a heatmap of potential defects, a probability cloud hovering over a digital twin of a casting mold. It’s not just monitoring. It’s recommending: adjust the pouring temperature here, modify the alloy composition there. The operator, whose expertise is in the feel of molten metal and the sound of a cooling mold, is being guided by a system that speaks his language of grain structure and thermal stress. This is the interface of TVARIT, a Frankfurt startup betting that the path to zero-waste manufacturing runs not through generic data dashboards, but through AI steeped in centuries of metallurgical craft [TVARIT, April 2023].
The metallurgical wedge
TVARIT’s product is industrial AI software, but its differentiation is a specific alloy of expertise. It combines machine learning models with what the company calls “metallurgical domain knowledge and heuristic process knowledge” [TVARIT, April 2023]. In practice, this means its systems are built to understand why a casting cracks or why energy is wasted in a heat treatment cycle, not just to flag that it happens. The output is prescriptive: actionable recommendations to adjust process parameters that traditionally relied on a veteran engineer’s intuition. The company claims this hybrid approach can reduce manufacturing scrap by 50% and energy use by up to 30% [TVARIT, April 2023]. An independent funding announcement cites specific customer results of 50% lower scrap and 18% lower energy consumption [EU-Startups, October 2023]. For an industry where material and energy costs are primary inputs, and sustainability mandates are tightening, these are existential percentages.
Traction across five continents
Founded in 2019, TVARIT has moved from thesis to deployment with notable speed. The company reports operating in nine countries across five continents and claims over 55 plants impacted, with annual customer savings exceeding 14 million euros [Startuprad.io] [EU-Startups, October 2023]. Its customer roster includes established names in metal components like Maxion Wheels (described as one of the world’s largest automotive wheel manufacturers), Kamax, KSM, and the Schunk Group [Futury Capital, May 2022]. The relationship with Maxion Wheels is particularly telling; TVARIT is cited as becoming the customer’s “sole AI partner for their world-wide implementation” [Futury Capital, May 2022]. This suggests a move beyond point solutions toward becoming a standardized, embedded layer of operational intelligence.
A key enabler of this expansion is a strategic partnership with PSI, a major industrial software provider. Through this integration, TVARIT’s metals-specific AI models are now embedded within the PSImetals Service Platform, giving the startup a direct channel into the digital infrastructure of large manufacturers [PSI, December 2024].
Funding and founding chemistry
The bet has attracted over $15 million in disclosed venture capital. A 2022 seed round of nearly $3 million was led by Matterwave Ventures, followed by a Series A in 2023 that totaled approximately $13.85 million led by Momenta and Futury Capital, with continued participation from Matterwave and new investor 4impact capital [Futury Capital, May 2022] [EU-Startups, October 2023] [Seedtable, October 2023].
2022 Seed | 2.996 | M USD
2023 Series A | 13.85 | M USD
The founding team pairs technical and commercial depth. Suhas Patel is the founder and CEO, while Rahul Prajapat, an IIT Bombay graduate, is the co-founder and was identified as CTO in earlier reporting [EU-Startups, October 2023] [EDI - European Data Incubator, March 2020]. The company’s leadership page now lists Deepak Singh as CTO, indicating an evolution in the technical leadership structure as the company scales [Startuprad.io]. The team has grown to an estimated 61 employees, with hiring focused on data science and engineering roles in Pune, India, reflecting a build-out of the core AI capability [RocketReach, Unknown] [Anjali Tiwari - TVARIT | LinkedIn, retrieved 2026].
The implementation frontier
For all its technical promise, TVARIT’s model faces the classic frontier of industrial software: implementation complexity. Selling to metal manufacturers means navigating long sales cycles, integrating with legacy control systems, and, most critically, earning the trust of plant-floor operators. The company’s wedge of domain expertise is designed to bridge this last gap. Yet, the scale of its claimed results,50% scrap reduction,sets a high bar for consistent, repeatable delivery across diverse foundries and forging plants. The partnership with PSI mitigates the integration challenge, but the cultural adoption of AI-driven recommendations in a hands-on industry remains a real-time experiment.
The competitive landscape is also opaque. While no direct competitors are named in available sources, the space for industrial process optimization is crowded with large incumbents like Siemens or SAP, and a host of specialist AI startups. TVARIT’s defensibility rests on the depth of its metallurgical models and the proprietary datasets gleaned from its growing project portfolio. Its early list of flagship customers, however, provides a strong beachhead.
What the screen is really asking
The interface in that foundry office, recommending a tweak to a process that has run a certain way for decades, poses a quiet cultural question. It asks how expertise is built and who gets to hold it. TVARIT is not just selling efficiency metrics; it’s selling a translation layer between the tacit knowledge of the master foundryman and the explicit logic of the algorithm. The product’s implicit bet is that the future of heavy industry belongs to hybrids,not just of metals, but of intelligence. The success of this Frankfurt startup will hinge on whether manufacturers believe that the deepest knowledge of their craft can be encoded, and whether they will trust the screen to help them pour a better casting.
Sources
- [TVARIT, April 2023] Tvarit raises €5m to drive net-zero metal manufacturing | https://tvarit.com/tvarit-raises-e5m-to-drive-net-zero-metal-manufacturing/
- [EU-Startups, October 2023] Frankfurt-based TVARIT bags €6.5 million to accelerate zero-waste manufacturing in metal industries | https://www.eu-startups.com/2023/10/frankfurt-based-tvarit-bags-e6-5-million-to-accelerate-zero-waste-manufacturing-in-metal-industries/
- [Futury Capital, May 2022] Deep-tech startup TVARIT successfully raised €2.8 million funding | https://futurycapital.vc/deep-tech-startup-tvarit-successfully-raised-e2-8-million-funding/
- [Seedtable, October 2023] TVARIT funding | https://seedtable.com
- [Startuprad.io] TVARIT | Startuprad.io | https://www.startuprad.io/entities/tvarit
- [PSI, December 2024] PSI and TVARIT announce strategic partnership | https://www.psi.de/en/press/press-releases/psi-and-tvarit-announce-strategic-partnership
- [EDI - European Data Incubator, March 2020] EDI startup profile | https://edincubator.eu
- [RocketReach, Unknown] TVARIT employee count | https://rocketreach.co
- [Anjali Tiwari - TVARIT | LinkedIn, retrieved 2026] LinkedIn profile | https://www.linkedin.com/in/anjali-tiwari-