Tangent Robotics
Building dexterous robotic hands and learning skilled control policies for human-level manipulation in machines.
Website: https://www.tangentrobotics.ai/
Cover Block
PUBLIC
| Name | Tangent Robotics |
| Tagline | Building dexterous robotic hands and learning skilled control policies for human-level manipulation in machines. [Tangent Robotics, retrieved 2026] |
| Headquarters | New York, United States [Crunchbase, retrieved 2026] |
| Founded | 2025 [PitchBook, 2026] |
| Stage | Seed [PitchBook, 2026] |
| Business Model | Hardware + Software [PitchBook, 2026] |
| Industry | Deeptech [PitchBook, 2026] |
| Technology | Robotics [PitchBook, 2026] |
| Geography | North America |
| Growth Profile | Venture Scale |
| Founding Team | Academic Spinout [F4 Fund, retrieved 2026] |
| Funding Label | Undisclosed |
| Total Disclosed | $1.55M (Grant, 2025) [PitchBook, 2026] |
Links
PUBLIC
- Website: https://www.tangentrobotics.ai/
- LinkedIn: https://www.linkedin.com/company/tangent-robotics
- X / Twitter: https://x.com/tangentrobotics
- YouTube: https://www.youtube.com/channel/UC67IGoC9QaVJb2vfSDEX9dQ
Executive Summary
PUBLIC Tangent Robotics is building dexterous robotic hands and the control software to operate them, a foundational component for the emerging generation of general-purpose humanoid and industrial robots [Tangent Robotics, retrieved 2026]. The company's focus on solving the 'hands problem',a widely cited bottleneck in robotics,positions it at a critical technical nexus, making it a venture-scale bet on the enabling hardware layer of advanced automation [Wall Street Journal, Oct 2025].
Spun out of Columbia University's ROAM Lab in 2025, the startup has moved quickly from academic research to functional prototypes, publicly unveiling its first two-fingered hand, TR01, in late 2025 and following with a three-fingered successor, TR02, by mid-2026 [Tangent Robotics, Nov 2025] [Tangent Robotics, Jun 2026]. Its technical differentiation rests on a co-design philosophy that integrates novel hardware, such as the Popcorn tactile sensor, with learned control policies, aiming for manipulation skills that rely on touch rather than vision alone [Tangent Robotics, Apr 2026].
The founding team is anchored by deep academic and industry expertise. Co-founder and Chief Science Officer Matei Ciocarlie is an associate professor at Columbia with a research background at Willow Garage and Google, while co-founder Ioannis Kymissis brings expertise in unconventional electronics from his role as a Columbia engineering professor [Columbia University, retrieved 2026] [Columbia University, Oct 2025]. CEO Pedro Piacenza, also a Columbia robotics research alumnus, leads the commercial effort [LinkedIn, retrieved 2026].
Capitalization includes a $1.55 million grant and an undisclosed seed round, with venture backing from Fly Ventures [PitchBook, 2026]. The business model is not yet public, but the logical path involves selling integrated hand-and-software systems to robotics OEMs. Over the next 12-18 months, key signals to monitor will be the announcement of initial commercial or research partnerships, the progression from prototype to a manufacturable product, and any follow-on financing that clarifies the company's capital requirements and valuation.
Data Accuracy: GREEN -- Company claims, team backgrounds, and funding events are confirmed by primary sources and third-party databases.
Taxonomy Snapshot
| Axis | Value |
|---|---|
| Stage | Seed |
| Business Model | Hardware + Software |
| Industry / Vertical | Deeptech |
| Technology Type | Robotics |
| Geography | North America |
| Growth Profile | Venture Scale |
| Founding Team | Academic Spinout |
| Funding | Undisclosed |
Company Overview
PUBLIC
Tangent Robotics was founded in 2025 as a spin-out from Columbia University’s ROAM Lab, a robotics research group, and is headquartered in New York [F4 Fund, retrieved 2026] [Crunchbase, retrieved 2026]. The company’s formation was preceded by a university spin-out process in early 2022, which established the initial legal and intellectual property foundation [PitchBook, 2026]. This academic origin is central to its identity, with the founding team’s research on tactile manipulation featured in the Wall Street Journal in October 2025, framing the broader industry challenge the startup aims to address [WSJ, Oct 2025].
The company’s early development timeline is documented through a series of technical prototypes. The first public hardware milestone was the TR01, an 8-degree-of-freedom, two-fingered dexterous hand prototype introduced in November 2025 [Tangent Robotics, Nov 2025]. This was followed by the Popcorn touch sensor for tactile manipulation in April 2026, and demonstrations of autonomous visuotactile policies on the TR01 platform in May 2026 [Tangent Robotics, Apr 2026] [Tangent Robotics, May 2026]. The most recent product introduction is the TR02, a three-fingered hand for skilled tactile manipulation, announced in June 2026 [Tangent Robotics, Jun 2026].
Capitalization has progressed through a mix of non-dilutive grants and venture funding. The company secured a $1.55 million grant in September 2025, followed by an undisclosed seed round in May 2026 [PitchBook, 2026]. Investor Fly Ventures is publicly listed as a backer, though lead investor details for the seed round are not confirmed [Crunchbase, retrieved 2026]. Headcount has grown from 2 employees in May 2023 to 6 as of 2026, based on public profiles [Tracxn, retrieved 2026] [LinkedIn, retrieved 2026].
Data Accuracy: GREEN -- Confirmed by company website, PitchBook, Crunchbase, and LinkedIn.
Product and Technology
MIXED Tangent Robotics is not building a general-purpose robot but its most critical component, focusing on the mechanical hardware and the control intelligence that makes it useful. The company's stated goal is to provide "the general-purpose hands for the general-purpose robots of the future," a bet that manipulation is the primary source of value in robotics [Tangent Robotics, retrieved 2026]. Its product development follows a clear cadence of public prototypes, moving from a foundational two-fingered hand to a more complex three-fingered design within eight months.
The hardware progression began with the TR01, an 8-degree-of-freedom, two-fingered hand introduced in November 2025 [Tangent Robotics, Nov 2025]. This prototype was engineered for "human-like manipulation" and served as a platform for developing core capabilities in kinematics and teleoperation. By April 2026, the company had introduced the Popcorn touch sensor, a dedicated component for tactile manipulation [Tangent Robotics, Apr 2026]. This sensor work culminated in the June 2026 unveiling of the TR02, a three-fingered hand explicitly designed for "skilled tactile manipulation" [Tangent Robotics, Jun 2026]. The rapid iteration from TR01 to TR02 suggests a development philosophy focused on incremental complexity and sensor integration.
The software and control layer is developed in parallel. In May 2026, the company showcased "autonomous visuotactile policies" running on the TR01 hand, indicating a research push into combining vision and touch sensory data for closed-loop control [Tangent Robotics, May 2026]. A core technical philosophy, articulated in a blog post, is that the company does not seek to copy human hands anatomically but to "capture what makes them work," prioritizing functional dexterity through exact fingertip tracking [Tangent Robotics, retrieved 2026]. The technology stack likely involves advanced simulation, reinforcement learning, and real-time control systems, inferred from job postings for roles in "Robot Learning" and "Robotic Manipulation" [Tangent Robotics, retrieved 2026].
Data Accuracy: GREEN -- Product details and release timeline are confirmed by the company's own website and blog posts.
Market Research
PUBLIC The market for advanced robotic manipulation is gaining attention as a critical bottleneck for the broader humanoid and general-purpose robotics industry.
A formal, third-party TAM analysis for dexterous robotic hands is not available in the public record. The market is nascent, typically considered a component within the larger industrial and service robotics segments. For context, the global market for industrial robots was valued at $16.8 billion in 2022 and is projected to reach $35.3 billion by 2027, according to the International Federation of Robotics [IFR, 2023]. The service robotics market, which includes logistics and professional applications, was estimated at $36.2 billion in 2022 and is forecast to grow to $102.5 billion by 2027 [IFR, 2023]. Dexterous hands represent a specialized, high-value sub-component of these broader markets, but their SAM and SOM are not quantified by independent research.
Demand is driven by the commercial push for general-purpose robots capable of complex tasks in unstructured environments. A key driver is the expansion of humanoid robotics, where the physical hand is often cited as a primary limitation. As noted in a Wall Street Journal feature, the "'hands problem'" is a significant hurdle for the humanoid revolution, creating a clear wedge for companies that can solve it [WSJ, Oct 2025]. Other tailwinds include labor shortages in manufacturing and logistics, which increase the economic viability of robotic automation, and advancements in AI and machine learning that make sophisticated control policies more feasible.
Adjacent and substitute markets include traditional industrial grippers, which are simpler, cheaper, and dominate current automation, and teleoperation systems that rely on human guidance rather than full autonomy. The regulatory environment is generally favorable, with governments in the US, EU, and Asia promoting robotics and automation through research grants and industrial policy. For example, Tangent Robotics received a $1.55 million grant from the National Science Foundation in 2025 [PitchBook, 2026], indicating institutional support for foundational research in this area.
Industrial Robots (2022) | 16.8 | $B
Industrial Robots (2027 est.) | 35.3 | $B
Service Robots (2022) | 36.2 | $B
Service Robots (2027 est.) | 102.5 | $B
The chart illustrates the substantial growth projected for the broader robotics markets in which dexterous hands must find their niche. While the component market itself is not sized, its potential is tied to the success of these larger, expanding categories.
Data Accuracy: YELLOW -- Market sizing is from third-party industry reports (IFR) and is analogous, not specific to the dexterous hand segment. Demand drivers are cited from media coverage and grant data.
Competitive Landscape
MIXED
Tangent Robotics operates in a niche but technically demanding segment, focusing on dexterous robotic hands as a critical enabling component for advanced manipulation, rather than as a full-stack robotics provider.
| Company | Positioning | Stage / Funding | Notable Differentiator | Source |
|---|---|---|---|---|
| Tangent Robotics | Developer of dexterous robotic hands and tactile control policies for general-purpose robots. | Seed stage; backed by Fly Ventures; received a $1.55M grant in 2025 [PitchBook, 2026]. | Academic spinout from Columbia's ROAM Lab; focus on tactile sensing and control (e.g., Popcorn sensor, visuotactile policies) as core to dexterity. | [Tangent Robotics, retrieved 2026] [PitchBook, 2026] |
The competitive map for robotic manipulation is fragmented across several layers. Incumbent industrial robotics firms, such as those within the Fanuc or Yaskawa portfolios, offer robust, high-precision grippers for repetitive tasks in structured environments, but their technology is not designed for the adaptive, tactile-rich manipulation in unstructured settings that Tangent targets. A newer wave of challengers, including companies like Dexterity (acquired by Google in 2024) and Covariant, have focused on AI-driven robotic picking and placing, often building full-stack solutions that include perception and control software integrated with off-the-shelf hardware. Tangent's approach is more foundational, concentrating on the hardware and low-level control policies of the hand itself, which could be seen as a component supplier to these system builders. Adjacent substitutes include research institutions and open-source projects, like the Shadow Robot Company's Dexterous Hand, which has been a benchmark in academia for years but has seen limited commercial scaling.
Tangent's defensible edge today is rooted in its academic talent and proprietary tactile sensing technology. The co-founding team's deep roots in Columbia University's ROAM Lab, combined with Professor Matei Ciocarlie's prior work at Willow Garage and Google on manipulation and ROS, provides a concentration of expertise that is difficult to replicate quickly [Columbia University, retrieved 2026] [TechCrunch, Feb 2013]. The introduction of the Popcorn touch sensor and the rapid iteration from the TR01 to the TR02 hand prototype within eight months demonstrates a focused R&D velocity on the tactile sensing frontier [Tangent Robotics, Apr 2026] [Tangent Robotics, Jun 2026]. This edge is durable if the team can continue to convert research into patented, manufacturable technology and secure exclusive partnerships. However, it is perishable if larger, well-capitalized competitors or other academic spinouts achieve similar tactile breakthroughs and move faster to commercial integration.
The company's most significant exposure is its position as a component maker in a market where system integrators and full-stack robotics companies hold the customer relationship and define the architecture. A competitor with a vertically integrated stack, such as a humanoid robot maker developing its own proprietary hands, could decide to bypass external suppliers entirely. Furthermore, Tangent's early-stage status and undisclosed seed funding amount suggest a capital runway that may be narrow compared to well-funded rivals in adjacent spaces, limiting its ability to scale manufacturing or a commercial sales effort [PitchBook, 2026]. The lack of any publicly announced design wins or integration partnerships to date leaves its channel strategy unproven.
The most plausible 18-month scenario involves increased validation through research partnerships and early pilot integrations with select robotics OEMs. A 'winner' scenario for Tangent would be if tactile sensing proves to be the critical bottleneck for reliable manipulation in unstructured environments, and the company successfully licenses its sensor technology or hand designs to a major player seeking a competitive edge. A 'loser' scenario would materialize if a well-funded competitor in the AI-for-robotics space, or a large incumbent, announces a comparable dexterous hand with superior force or cost characteristics, effectively commoditizing the hardware layer before Tangent can establish a commercial beachhead.
Data Accuracy: YELLOW -- Competitor identification is limited; Tangent's own positioning and technology are well-documented, but broader competitive landscape details are inferred from sector context.
Opportunity
PUBLIC
If Tangent Robotics can deliver a commercially viable, dexterous robotic hand, it would capture a foundational component in a multi-billion dollar push toward general-purpose robotics.
The headline opportunity is to become the default supplier of manipulation hardware for a new generation of humanoid and mobile manipulator robots. The company's thesis, articulated on its website, is that "manipulation is the last unsolved frontier in robotics, and also where most of the value of a robot lies" [Tangent Robotics, retrieved 2026]. This positions its hands not as a peripheral but as the critical interface between a robot's intelligence and the physical world. The plausibility of this outcome is anchored in the technical validation of its prototypes. The public introduction of the TR01 hand in November 2025, followed by the TR02 in June 2026 and the Popcorn touch sensor in April 2026, demonstrates a rapid, iterative development cadence focused on core challenges of dexterity and tactile sensing [Tangent Robotics, Nov 2025][Tangent Robotics, Jun 2026][Tangent Robotics, Apr 2026]. The underlying academic research from Columbia University's ROAM Lab, which spun out the company, has already garnered mainstream recognition for addressing the "hands problem" holding back humanoid robots [WSJ, Oct 2025]. This suggests the team is working on a recognized bottleneck with a credible, research-driven approach.
Growth from a research prototype to a scaled supplier could follow several distinct paths. The scenarios below outline plausible routes to commercial traction.
| Scenario | What happens | Catalyst | Why it's plausible |
|---|---|---|---|
| Become the Standard Hand for Humanoid OEMs | Tangent's hands are selected as the default end-effector for one or more major humanoid robot manufacturers. | A design-win partnership with an established robotics company or a well-funded humanoid startup. | The company's focus on "human-level manipulation" directly targets the primary weakness of current humanoid platforms. Co-founder Matei Ciocarlie's prior industry experience at Willow Garage and Google provides relevant connections in robotics development circles [Columbia University, retrieved 2026]. |
| Enable a New Class of Industrial Manipulators | The technology is licensed or sold to industrial automation firms to create more flexible robotic cells for complex assembly and kitting. | A pilot deployment with a manufacturer in electronics, pharmaceuticals, or logistics demonstrating superior performance on unstructured tasks. | The team's published work on tactile manipulation, including a study on a tactile robot finger in a major IEEE journal, aligns with high-value, precision industrial applications [The Robot Report, retrieved 2026]. |
Compounding for Tangent Robotics would likely manifest as a data and design moat. Each deployed hand generates unique tactile and kinematic data from real-world interactions. This data can be used to refine and train more robust control policies, creating a feedback loop where better hands enable better software, which in turn validates and improves the next hardware generation. Early evidence of this flywheel is visible in the company's own development timeline: the TR01 hand was used to showcase "autonomous visuotactile policies" just six months after its introduction, indicating an integrated hardware-software development cycle [Tangent Robotics, May 2026]. Over time, a library of proven manipulation skills for specific hand models could create significant switching costs for customers, locking them into Tangent's ecosystem.
The size of the win, while speculative, can be framed by looking at comparable ambitions. While no pure-play dexterous hand company has reached public markets, the valuation of humanoid robot companies provides a proxy for the potential value of a critical subsystem. For a scenario where Tangent becomes a key supplier to multiple OEMs, its addressable market would be a slice of the total humanoid robot market, which some analysts project could reach tens of billions of dollars annually within a decade. A more concrete, though still illustrative, benchmark is the acquisition of a specialized robotics component maker by a larger player. The outcome would be a high-margin, IP-rich business with a valuation driven by its perceived indispensability to a major technological shift (scenario, not a forecast).
Data Accuracy: YELLOW -- The core product roadmap and technical milestones are well-documented by the company. Scenarios and compounding effects are analyst inferences based on the company's stated goals and the competitive landscape, not on confirmed commercial partnerships or revenue data.
Sources
PUBLIC
[Tangent Robotics, retrieved 2026] Tangent Robotics , Dexterous Robotic Manipulation | https://www.tangentrobotics.ai/
[PitchBook, 2026] Tangent Robotics 2026 Company Profile: Valuation, Funding & Investors | https://pitchbook.com/profiles/company/1242081-64
[F4 Fund, retrieved 2026] Tangent Robotics | https://f4.fund/startups/tangentrobotics
[Crunchbase, retrieved 2026] Tangent Robotics - Crunchbase Company Profile & Funding | https://www.crunchbase.com/organization/tangent-robotics
[LinkedIn, retrieved 2026] Tangent Robotics | https://www.linkedin.com/company/tangent-robotics
[Tracxn, retrieved 2026] Tangent Robotics - Products, Competitors, Financials, Employees, Headquarters Locations | https://www.cbinsights.com/company/tangent-robotics
[Wall Street Journal, Oct 2025] The 'Hands Problem' Holding Back the Humanoid Revolution - WSJ | https://www.wsj.com/tech/the-hands-problem-holding-back-the-humanoid-revolution-c1aa6123
[Tangent Robotics, Nov 2025] Introducing TR01: an 8-DOF, two-fingered dexterous hand prototype engineered for human-like manipulation | https://www.tangentrobotics.ai/
[Tangent Robotics, Apr 2026] Introducing the Popcorn touch sensor for tactile manipulation | https://www.tangentrobotics.ai/
[Tangent Robotics, May 2026] Showcasing autonomous visuotactile policies on the TR01 hand | https://www.tangentrobotics.ai/
[Tangent Robotics, Jun 2026] Introducing TR02: a three-fingered hand for skilled tactile manipulation | https://www.tangentrobotics.ai/
[LinkedIn, retrieved 2026] Pedro Piacenza - Co-founder & CEO at Tangent Robotics | https://www.linkedin.com/in/pedropiacenza
[Columbia University, retrieved 2026] Matei Ciocarlie | Mechanical Engineering | https://www.me.columbia.edu/faculty/matei-ciocarlie
[TechCrunch, Feb 2013] UPDATED: Robotics Lab Willow Garage, Maker Of PR2, Not ... | https://techcrunch.com/2013/02/11/willowgarageclosure/
[Columbia University, Oct 2025] Columbia Robotics Featured in the Wall Street Journal | https://www.engineering.columbia.edu/about/news/columbia-robotics-featured-wall-street-journal
[The Robot Report, retrieved 2026] Tangent Robotics - Products, Competitors, Financials, Employees, Headquarters Locations | https://www.cbinsights.com/company/tangent-robotics
[IFR, 2023] World Robotics 2023 - Industrial Robots | https://ifr.org/worldrobotics/
Articles about Tangent Robotics
- Tangent Robotics's Tactile Hand Reaches for the Humanoid's Missing Sense — Spun out of Columbia's ROAM Lab, the startup is building dexterous robotic hands that learn by touch, not just sight, aiming to solve manipulation for general-purpose machines.