Perceptyne
Last reviewed
May 9, 2026
Sources
No citations yet
Review status
Needs citations
Revision
v4 · 2,814 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
May 9, 2026
Sources
No citations yet
Review status
Needs citations
Revision
v4 · 2,814 words
Add missing citations, update stale details, or suggest a clearer explanation.
| Perceptyne | |
|---|---|
| General information | |
| Full name | Perceptyne Robots Pvt. Ltd. |
| Founded | 2021 |
| Founders | Raviteja Chivukula, Mrutyunjaya Nadiminti, Jagga Raju Nadimpalli |
| Headquarters | Hyderabad, Telangana, India |
| Industry | Robotics, Artificial intelligence, industrial automation |
| Products | PR-34D, PR-9D, PR-Omni, PR-Uno, PR-Duo, PR-PhI |
| Total funding | Approximately $4.31 million (as of 2025) |
| Employees | Approximately 22 to 30+ (2025) |
| CEO | Raviteja Chivukula |
| Website | perceptyne.com |
Perceptyne is an Indian deep-tech robotics startup based in Hyderabad, Telangana, that designs and builds AI-first semi-humanoid robots for industrial manufacturing. Founded in 2021 by Raviteja Chivukula, Mrutyunjaya Nadiminti, and Jagga Raju Nadimpalli, the company develops vertically integrated robotic platforms that combine artificial intelligence, multimodal sensing, and dexterous end-effectors aimed at automating fine-motor assembly tasks on factory floors. Its flagship products are the dual-arm PR-34D and the single-arm PR-9D, both targeted at automotive and electronics manufacturing customers. The company markets itself as building robots that are "as flexible and adaptable as humans" and positions its systems between rigid industrial robots and full superhumanoid platforms.[1][2][3]
Perceptyne came to wider notice in October 2024, when it closed a $3 million seed round co-led by Endiya Partners and Yali Capital, with participation from Whiteboard Capital and a group of angel investors that included founders of Indian deep-tech companies such as Ather Energy and Skyroot Aerospace. The PR-34D was later named a finalist at the 2025 Humanoid Robotics Industry Awards in the Groundbreaking Technology category, alongside global names such as NVIDIA and AgiBot.[3][4][5]
Perceptyne was incorporated in 2021 as Perceptyne Robots Pvt. Ltd. According to founder interviews, the original concept was sketched out by Raviteja Chivukula nearly a decade earlier, when he was an intern at a vending machine company and began thinking about intelligent robotic arms that could pick and dispense products without rigid mechanical fixtures. The company's first prototypes followed in 2022 and 2023, and operations began at scale once the founding team came together in Hyderabad.[1][6]
The early years focused on developing in-house hardware and software stacks rather than integrating off-the-shelf industrial robot arms. According to investor disclosures, the team designed its own actuators, motor controllers, gearboxes, and firmware so that the platform would not depend on imported components. This vertical integration is a recurring theme in how Perceptyne describes itself and how investors describe the company. Endiya Partners has stated that the founders' "engineering depth and domain experience across aerospace, automotive, and electronics" was a key factor in the firm's investment decision.[3][7]
By 2025, Perceptyne had grown to a team of roughly 22 employees according to data aggregator Tracxn, with the company itself reporting 30 or more team members across mechanical and embedded systems engineering, AI and robotics research, product design, firmware, and operations. The startup operates from Hyderabad and reported revenue of approximately INR 2.07 crore for the financial year ending 31 March 2024.[8][9]
| Name | Role | Background |
|---|---|---|
| Raviteja Chivukula | Co-founder, CEO, and Co-CTO | IIT Madras and BITS Pilani alumnus; deep-tech product experience across aerospace, automotive, and electronics; originator of the Perceptyne concept[2][6][7] |
| Mrutyunjaya Nadiminti | Co-founder, Chief Business Officer, and Co-CTO | IIT Madras and BITS Pilani alumnus; product development background in deep-tech hardware[2][7] |
| Jagga Raju Nadimpalli | Co-founder and Chief Operating Officer | IIT Madras Engineering Design graduate; SPJIMR (with a stint at Cornell Johnson); previously held roles at Caterpillar, Ashok Leyland, Ola, and was Chief Technology Officer at Boson Motors[10] |
Public profiles describe Mrutyunjaya Nadiminti's surname inconsistently, with some early outlets calling him "Mrutyunjaya Sastry". Perceptyne's own website and the company's filings list him as Mrutyunjaya Nadiminti.[2][9][11]
Perceptyne initially launched with two semi-humanoid robotic platforms, the PR-34D and the PR-9D, both aimed at fixed assembly stations rather than mobile humanoid form factors. Through 2025, the company expanded its product lineup with a family of platforms that share core technology but differ in form factor and target task. Perceptyne's website describes the following products:
| Product | Form factor | Status (2025) | Notes |
|---|---|---|---|
| PR-9D | Single-arm semi-humanoid | Shipping | Compact configuration for tasks where one manipulator suffices[2][3] |
| PR-34D | Dual-arm semi-humanoid | Shipping | Flagship dual-arm platform; finalist at 2025 Humanoid Robotics Industry Awards[2][5] |
| PR-Omni | Mobile semi-humanoid with dual 7-DOF arms and holonomic base | In development | Combines dual-arm manipulation with a wheeled mobile platform; total payload around 10 kg[12] |
| PR-Uno | Single articulated robotic arm | In development | Modular arm system for specialised tasks[12] |
| PR-Duo | Dual articulated arm system | In development | Twin-arm variant of the modular arm family[12] |
| PR-PhI | Industrial arm with modular architecture | In development | Visual servo tracking, multimodal perception, teleoperation[12] |
| PR-Sync | Not disclosed | Coming soon[12] |
Perceptyne refers to its broader platform family using the names Omni, Duo, Uno, PhI, and Sync on its corporate website, while public funding announcements and trade-press coverage continue to reference the original PR-34D and PR-9D.[2][12]
Perceptyne's robots are designed around what the company calls a "full-stack, AI-first robotics platform". The platform combines computer vision, force and tactile sensing, articulate manipulators, and AI software for perception, motion planning, and skill execution. Rather than program specific motions, customers train the robot on new tasks through teleoperation, with the AI system generalising those demonstrations into repeatable skills.[2][3][7][13]
The PR-34D and PR-9D integrate three sensing modalities into a single perception stack:
The combination of these modalities allows the robot to operate in unstructured assembly environments where parts may be slightly mispresented, where lighting may vary, and where touch is necessary to succeed at tasks like cable mating, snap fitting, or screw starting.[2][13]
The PR-34D uses two seven-degree-of-freedom arms, with three degrees of freedom at the shoulder, one at the elbow, and three at the wrist. Each arm is fitted with a three-fingered hand that has 10 degrees of freedom and integrates tactile sensing. Trade press has reported the dual-arm platform has a combined payload capacity of around 6 kg, making it suitable for precision assembly rather than heavy lifting.[2][13]
Perceptyne's hand design uses three fingers rather than the five-finger anthropomorphic hands common on full humanoid platforms. The company has stated this choice reflects the practical demands of factory assembly tasks: most production-line manipulations can be accomplished with three opposable digits, and reducing finger count simplifies kinematics, reduces failure modes, and lowers cost.[1][13]
A distinctive feature of Perceptyne's stack is the use of AI-led teleoperation as a training method. Operators teleoperate the robot through new tasks, and the system records the demonstrations as training data for its skill models. Once a skill is learned, the robot performs it autonomously and improves through continued practice. This approach is sometimes referred to in the company's marketing as low-code or no-code trainability, in the sense that customers do not need to write robot programs in a traditional sense to deploy a new task.[1][2][7]
The company has also stated that it uses NVIDIA Isaac Sim for simulation and pre-deployment testing of skills, allowing engineers to validate behaviour in a virtual factory before installation. This pairs the in-house hardware with a widely used third-party simulation environment from NVIDIA.[12]
Perceptyne builds the majority of its robotics stack in house. Public statements from the company and its investors describe in-house design and manufacturing of:
The company markets this vertical integration as a way to control cost, iterate quickly on the platform, and reduce dependence on imported deep-tech components. Endiya Partners has cited the vertical integration explicitly as part of its investment thesis.[7][12]
Perceptyne emphasises drop-in deployment as a commercial differentiator. According to the company and its investors, robots can be installed at existing assembly stations without modifying the surrounding line, reducing integration time "from quarters to weeks". One press release went further, claiming the company aims to compress automation implementation timelines by up to 90 percent and to install robots without halting production.[3][7][11][12]
The target customer is mid-scale and large manufacturers in electronics manufacturing, automotive components, and consumer goods, where labour shortages, rising wages, and high mix and low volume product runs make traditional fixed automation impractical. Perceptyne has named potential applications such as torchlight assembly, headlight assembly, brake component assembly, and mobile phone assembly.[1][3]
As of 2025, Perceptyne has raised approximately $4.31 million across at least three funding rounds, according to data aggregator Tracxn. The company has not disclosed its valuation. Reported funding events are summarised below.
| Date | Round | Amount | Lead investors | Other participants |
|---|---|---|---|---|
| April 2023 | Initial round (per Tracxn) | Not disclosed | Not disclosed | Early backers including Z21 Ventures, T-Hub[8][14] |
| 1 May 2024 | Pre-seed | Not disclosed publicly (part of approximately $1.31 million prior to the seed round) | Venture Catalysts | T-Hub, Z21 Ventures, Tarun Mehta and Swapnil Jain (Ather Energy), Pawan Chandana and Bharat Daka (Skyroot Aerospace), Aditya Damani (Credit Fair), and other angels[14][15] |
| 14 October 2024 | Seed | $3 million (approximately INR 25 crore) | Endiya Partners, Yali Capital | Whiteboard Capital and angel investors[3][4][16] |
The seed round announced on 14 October 2024 was co-led by Endiya Partners and Yali Capital. According to public coverage, Endiya was represented by Raghav Gupta, who described Perceptyne as building "a drop-in, generic automation solution that learns through AI-led tele-operation". Karthik Madathil, partner at Yali Capital, described Perceptyne as having "unique multi-faceted technology to deliver robotic solutions precisely tuned to sunrise sectors of India's growing economy". The company stated that proceeds would be used to ramp up product development, customer acquisition, and growth.[3][4][11]
Perceptyne's pre-seed round in May 2024 was led by Venture Catalysts and notably included founders of other Indian deep-tech hardware companies as angels, including Tarun Mehta and Swapnil Jain of Ather Energy and Pawan Chandana and Bharat Daka of Skyroot Aerospace. The participation of T-Hub, the Hyderabad-based startup incubator, has tied Perceptyne to the broader Telangana deep-tech ecosystem.[14][15]
Perceptyne has not publicly named individual customers, but it has disclosed several details about its commercial traction:
In parallel, Perceptyne has continued to develop later prototype generations of its platform. Public coverage in 2025 referred to the company building a "Version 4" prototype in-house for internal testing and pilot deployment.[13]
| Year | Recognition | Details |
|---|---|---|
| 2025 | Finalist, Humanoid Robotics Industry Awards | The PR-34D was named a finalist in the Groundbreaking Technology category, alongside global leaders including NVIDIA and AgiBot. Co-founder Jagga Raju Nadimpalli described the recognition as "a moment of pride for Indian deep-tech innovation".[5] |
| 2025 | Forbes India DGEM 2025 Select 200 (per company website) | Perceptyne lists this recognition on its corporate website as part of a cohort of 200 companies with global business potential.[12] |
Perceptyne operates in a segment of the robotics market that sits between traditional industrial robot arms and full humanoid robots such as Tesla Optimus and Figure AI platforms. Investors have framed the company's opportunity in terms of two trends:
Perceptyne positions itself as part of an emerging cohort of Indian deep-tech robotics companies competing in segments where Western and Chinese players have historically led. Coverage in trade press has framed the company as a test case for whether the Indian market is ready to absorb domestically built collaborative robots and semi-humanoid platforms at scale.[1][13]
The PR-34D is the company's flagship dual-arm semi-humanoid robot, designed for tasks requiring precision and dexterity such as component assembly, quality inspection, and handling intricate objects.
| Specification | Value |
|---|---|
| Form factor | Stationary dual-arm semi-humanoid |
| Number of arms | 2 |
| Arm degrees of freedom | 7 per arm (3 shoulder, 1 elbow, 3 wrist) |
| Hand degrees of freedom | 10 per hand (three-fingered) |
| Sensing | Multimodal: vision, force, and tactile |
| Combined payload | Approximately 6 kg |
| Training method | AI-led teleoperation, low-code skill training |
| Target tasks | Assembly, quality checks, intricate object handling, fine motor tasks |
| Target industries | Automotive, electronics, consumer goods, pharmaceuticals[2][13][16] |
The PR-9D is a single-arm variant of the platform aimed at applications where one manipulator is sufficient, offering a more compact and cost-effective option for specific manufacturing tasks.
| Specification | Value |
|---|---|
| Form factor | Single-arm semi-humanoid |
| Number of arms | 1 |
| Arm degrees of freedom | 7 |
| Hand degrees of freedom | 10 (three-fingered) |
| Sensing | Multimodal: vision, force, and tactile |
| Training method | AI-led teleoperation |
| Target tasks | Single-handed assembly, pick and place, inspection |
| Target industries | Automotive, electronics[2][13] |
PR-Omni is described on the company website as a mobile semi-humanoid platform that combines dual 7-DOF arms with a holonomic mobile base. According to the website, the platform offers a combined payload of around 10 kg and is intended for applications that require both manipulation and intra-facility mobility.[12]