Y Combinator-Backed Archgen Labs Develops AI Agent to Automate Advanced Semiconductor Design

Archgen Labs, a semiconductor technology startup founded by Hariharan Ayappane and Naveen Venkat, has introduced an artificial intelligence-powered engineering agent designed to automate complex tasks involved in semiconductor chip development. The technology aims to accelerate physical chip design by identifying engineering problems, implementing corrective changes and reducing the repetitive manual work traditionally required during semiconductor development.

The San Francisco-based startup, backed by Y Combinator’s Fall 2026 programme, has developed an AI agent called Newton that can analyse semiconductor design workflows and identify problems associated with circuit placement, routing, timing and manufacturing constraints. The company is also developing advanced tools for designing three-dimensional semiconductor architectures, in which multiple layers of computing components are vertically integrated to improve performance and energy efficiency.

The development was highlighted during the company’s November 2026 launch activities, with details emerging in reports published on November 9 and 10. Archgen is also working with semiconductor manufacturer GlobalFoundries on three-dimensional chip development, targeting its first semiconductor tapeout during 2027.

Newton AI Agent Automates Complex Semiconductor Engineering

Semiconductor design is among the most demanding areas of modern engineering, requiring intricate coordination between electronic circuit architecture, physical component placement, electrical connections, power distribution and thermal management.

Before a semiconductor chip can enter manufacturing, engineers must ensure that millions or billions of circuit elements operate correctly while satisfying strict physical and electrical requirements.

This process involves numerous specialised software tools collectively known as Electronic Design Automation (EDA). Engineers repeatedly analyse the results of these tools, identify problems and modify chip layouts to resolve performance or manufacturing constraints.

Archgen’s Newton AI agent is designed to automate parts of this iterative engineering process.

According to the company, Newton can examine the results of physical design operations, identify the underlying causes of failures, modify relevant design parameters and initiate subsequent verification runs.

The system is intended to address problems involving circuit placement, routing congestion, design-rule violations and timing closure.

Timing closure is particularly important because electrical signals must travel between components within specified time limits. If signal delays exceed these limits, the chip may fail to operate reliably at its intended frequency.

By automating the identification and correction of such problems, Newton aims to reduce the engineering effort required to move semiconductor designs towards manufacturing readiness.

Artificial Intelligence Moves Beyond Conventional Design Automation

Traditional semiconductor design tools automate many individual engineering operations, but substantial human involvement is still required to coordinate the overall workflow.

Engineers must interpret the results of successive software runs, understand the relationships between different design constraints and determine how changes in one part of a circuit affect other components.

For example, moving a circuit block may improve signal timing but increase routing congestion or create new power-distribution problems elsewhere in the design.

Archgen’s approach seeks to introduce greater autonomy into these interconnected engineering processes.

Rather than simply identifying an error, Newton is intended to determine why the error occurred, implement an appropriate correction and evaluate the consequences through additional design runs.

The company also describes a learning mechanism through which information obtained from previous engineering operations can inform subsequent optimisation tasks.

This approach could allow AI agents to assist semiconductor engineers across multiple stages of physical design rather than performing isolated operations.

The technology remains subject to the practical requirements of semiconductor manufacturing, where automated changes must ultimately satisfy rigorous verification and reliability standards.

Developing Next-Generation Three-Dimensional Semiconductor Chips

A major focus of Archgen’s research is the development of design technologies for three-dimensional integrated circuits.

Conventional semiconductor chips generally arrange their principal computing components across a two-dimensional silicon surface. Improvements in processing capability have historically depended heavily on reducing transistor dimensions and increasing the density of components.

However, continued transistor miniaturisation has become increasingly difficult because of manufacturing complexity, power consumption and physical limitations.

Three-dimensional semiconductor architectures provide an alternative approach by stacking multiple semiconductor dies vertically.

These architectures can bring computing, memory and input-output components closer together, reducing the distance that electrical signals must travel.

Shorter electrical connections can potentially improve performance and reduce energy consumption, particularly in applications where large quantities of data must move between processing and memory components.

The approach is increasingly relevant to artificial intelligence accelerators, high-performance computing and advanced data-centre processors.

However, vertically integrated architectures introduce additional engineering challenges involving heat dissipation, power delivery, electrical connections and interactions between different semiconductor layers.

Archgen is developing software intended to address these challenges by treating multiple semiconductor dies as components of a single integrated design.

LayerTherm Targets Faster Thermal Simulation

One of the technologies developed by Archgen is LayerTherm, a thermal-analysis tool designed to predict temperature distribution within complex semiconductor structures.

Heat management is a major challenge in advanced processors because high concentrations of transistors generate substantial thermal energy during operation.

The problem becomes more complicated in three-dimensional chips, where vertically stacked semiconductor layers can restrict the movement of heat away from internal components.

Excessive temperatures can affect electrical performance, reliability and the operating life of semiconductor devices.

Engineers therefore conduct thermal simulations to identify potential hotspots and evaluate different component arrangements before manufacturing.

Archgen claims that LayerTherm can predict semiconductor temperature distributions approximately 5,000 times faster than conventional simulation approaches in its internal benchmarks.

If this performance can be reproduced across practical industrial workloads, faster thermal analysis could allow engineers to evaluate substantially more design configurations within the same development period.

The company aims to combine thermal predictions with physical design optimisation, enabling engineering decisions that account for electrical performance, power consumption and heat distribution simultaneously.

The reported acceleration remains a company benchmark rather than an independently established industry-wide performance result.

GPU-Accelerated Placement Engine Supports Chip Optimisation

Archgen has also developed a GPU-accelerated placement engine designed to optimise the arrangement of major circuit components within a semiconductor chip.

Component placement is a critical stage of chip development because the physical location of processing elements, memory blocks and other circuits influences electrical performance, power requirements and manufacturing feasibility.

Poor placement decisions can increase the length of electrical connections, produce routing congestion and contribute to timing or thermal problems.

The startup reports that its placement engine achieved first place in a macro-placement challenge organised by Hudson River Trading and Partcl in June 2026.

The result provides an indication of the company’s work in computational optimisation, although performance in a specialised benchmark does not automatically establish superiority across all commercial semiconductor design environments.

Archgen intends to combine its placement technology with Newton’s AI-driven workflow automation and LayerTherm’s thermal-analysis capabilities.

Together, these tools are intended to support a more integrated approach to designing advanced semiconductor systems.

GlobalFoundries Collaboration Targets First Tapeout in 2027

Archgen has disclosed that it is working with GlobalFoundries on the development of three-dimensional semiconductor technology.

GlobalFoundries is an established international semiconductor manufacturer that provides chip fabrication services to companies developing a wide range of electronic products.

The collaboration is intended to support the practical development of advanced semiconductor designs using Archgen’s engineering technologies.

According to the company’s Y Combinator profile, the startup is targeting its first semiconductor tapeout in 2027, with contemporary reporting identifying April 2027 as a more specific objective.

Tapeout is an important milestone in semiconductor development, marking the point at which a completed chip design is finalised and submitted for fabrication.

Reaching tapeout requires the design to satisfy applicable electrical, physical and manufacturing requirements.

However, tapeout does not mean that a manufactured chip has been tested successfully or entered commercial production.

Archgen’s proposed 2027 milestone therefore represents a significant development objective rather than an already completed manufacturing achievement.

Emerging Opportunities in AI-Powered Semiconductor Design

The introduction of artificial intelligence into semiconductor engineering is becoming increasingly important as chip architectures grow more complex.

Advanced processors used for artificial intelligence, data centres, telecommunications and scientific computing require significant engineering resources during design and verification.

The growing complexity of these devices has increased interest in software capable of automating repetitive engineering tasks and supporting more sophisticated design optimisation.

AI-assisted semiconductor design has the potential to reduce development time, improve engineering productivity and expand the number of design alternatives that can be evaluated before manufacturing.

For emerging semiconductor companies, such technologies could help reduce some of the engineering barriers associated with developing specialised computing hardware.

Nevertheless, commercially successful semiconductor design automation requires high levels of reliability, compatibility with established EDA workflows and extensive validation against actual manufacturing requirements.

Archgen’s future progress will therefore depend on demonstrating that its AI-powered tools can deliver consistent improvements across practical semiconductor projects.

Indian-Origin Entrepreneurship in Advanced Semiconductor Technology

Founded by Hariharan Ayappane and Naveen Venkat, Archgen represents an emerging venture in the rapidly developing field of AI-assisted semiconductor engineering.

Although the company is based in San Francisco, its founders’ participation in advanced semiconductor technology reflects the wider presence of Indian-origin entrepreneurs in international deep-technology innovation.

The company’s selection for Y Combinator’s Fall 2026 programme provides access to a startup ecosystem that has supported numerous technology companies through early product development and commercial expansion.

Archgen’s focus on semiconductor design is particularly relevant at a time when the global computing industry is exploring new architectures to meet the increasing demands of artificial intelligence and high-performance computing.

Its development of Newton, LayerTherm and GPU-accelerated placement technology illustrates how artificial intelligence is beginning to influence not only software applications but also the engineering processes used to create the underlying computing hardware.

Advancing the Future of Semiconductor Engineering

Archgen Labs’ introduction of Newton represents an effort to bring greater automation to one of the most technically demanding stages of semiconductor development.

By combining AI-driven design analysis, accelerated thermal simulation and computational placement optimisation, the company is seeking to address engineering challenges that become more complex as semiconductor architectures evolve.

Its emphasis on three-dimensional chip design is especially significant because vertical integration is emerging as an important approach to improving computing performance while managing the limitations of conventional semiconductor scaling.

The company’s proposed collaboration with GlobalFoundries and its targeted 2027 tapeout will provide important opportunities to evaluate the practical application of these technologies.

While further industrial validation remains necessary, Archgen’s development highlights the expanding role of artificial intelligence in advanced hardware engineering and the growing opportunities for startups working at the intersection of semiconductor technology, computational physics and automated design.

---