BENGALURU | 26 September 2026: As the energy consumption of conventional digital processors threatens to bottleneck the global expansion of artificial intelligence, US-based semiconductor firm Mythic AI has established a new Centre of Excellence (CoE) in Bengaluru. Operating locally through Mythic India Private Limited in Koramangala, the facility is structured to serve a dual mandate: executing core pre-silicon and post-silicon chip design while anchoring the company’s commercial expansion across India and the broader Asian market.

Mythic is embedding the Bengaluru team directly into the development of its analog compute-in-memory (ACIM) architecture, a hardware approach that processes AI workloads directly inside memory arrays to curb the power and heat constraints of traditional GPUs.

KEY HIGHLIGHTS

  • Aggressive Headcount Expansion: Starting from an initial base of nearly 20 specialists, Mythic plans to triple to quadruple its Bengaluru engineering workforce exceeding 60 engineers by the end of Q1 2027.
  • High Share of Flagship Silicon: According to Chairman and CEO Dr. Taner Ozcelik, the India centre already executes roughly 25% of the engineering work on Vanguard, Mythic’s next-generation AI inference platform, with plans to scale that contribution to nearly 50% as local capabilities mature.
  • End-to-End Product Ambition: Corporate leadership has outlined a roadmap where the Bengaluru hub will eventually design and bring select semiconductor products to market entirely from India on an end-to-end basis.
  • Capital & IP Foundation: The Indian expansion follows Mythic’s $125 million oversubscribed funding round (announced in December 2025) and its acquisition of German processor-IP firm Videantis, which combined analog compute-in-memory with digital processing IP.

Founded in 2012 and dual-headquartered in Austin, Texas, and Palo Alto, California, Mythic AI is a deep-tech semiconductor company specializing in Analog Processing Units (APUs) for artificial intelligence inference.

Conventional digital AI chips consume vast amounts of electricity by repeatedly shuttling data back and forth between external memory (DRAM) and the processor a structural limitation known in computer architecture as the von Neumann bottleneck. Mythic bypasses this constraint through analog compute-in-memory, utilizing flash memory arrays as configurable resistors to perform massive matrix-multiply calculations directly inside memory. The company states that a single Vanguard rack can match the compute output of 650 conventional server racks, a density and power advantage that has already been validated for tactical-edge defense deployments through a strategic partnership with Lockheed Martin.

“

We intend to build capabilities in India, including the ability to design and take certain products to market entirely from the country, from end to end. India has an exceptional talent pool and a strong education ecosystem, including the IITs and other leading universities, which we are tapping into.

– Dr. Taner Ozcelik, Chairman and CEO of Mythic AI

SILICON SCOPE: INSIDE THE BENGALURU ENGINEERING MANDATE

Developing analog-digital hybrid processors requires tight coordination across physics, circuit layout, and system software. The Bengaluru centre is structured as a multi-disciplinary semiconductor hub spanning the full chip lifecycle:

  • Hybrid Circuit & Physical Design: Local engineers are actively executing VLSI design, front-end design and verification of core analog compute blocks, and analog-digital hybrid physical design (PD).
  • Automotive-Grade Design-for-Test (DFT): A critical focus of the Bengaluru team involves engineering specialized DFT techniques designed to meet rigorous automotive compliance and reliability standards, supporting workloads such as Mythic’s ongoing analog AI chip collaboration with Honda.
  • Post-Silicon & Intelligent Sensing: Beyond initial layout for the Vanguard platform, the facility’s scope is expanding into silicon testing and characterisation, systems engineering, applications engineering, and intelligent-sensing chip development.
  • Commercial & Asian Market Development: Alongside R&D, Mythic is utilizing its Indian footprint to evaluate commercial deployments across Asia, targeting domestic demand in data-centre inference, Advanced Driver-Assistance Systems (ADAS), industrial robotics, defence, and edge computing

Detailing the technical depth required for this roadmap, Manoj Kumar, Vice President of Mythic India, noted:

“

TBengaluru brings in an all-encompassing blend of semiconductor engineering talent-digital, analog, PD, DFT, SW, PE, TE and other specialized sub-disciplines in each of these complex domains. Mythic’s core AI semiconductor development would demand exceptional execution in each of these disciplines.

– Manoj Kumar, Vice President of Mythic India

Industry Insights

Mythic AI’s entry into Bengaluru highlights two major shifts within India’s semiconductor design landscape:

  1. Scarcity of Analog & Mixed-Signal Talent: While India houses tens of thousands of digital verification engineers, analog and mixed-signal design remains one of the hardest-to-staff disciplines in global microelectronics because it deals with continuous physical variables like voltage drift, thermal noise, and transistor physics. Mythic’s decision to build its hybrid analog-digital physical design team in Bengaluru underscores the city’s deep bench of veteran mixed-signal silicon architects
  2. From Sub-Block Verification to Full-Chip Ownership: Historically, multinational chipmakers used Indian centres to verify individual IP blocks designed in California or Austin. By assigning 25% to 50% of its flagship Vanguard platform to Bengaluru—and committing to future end-to-end product execution from the country—Mythic reflects a growing trend of venture-backed US silicon startups treating India as a co-primary R&D base.

FUTURE OUTLOOK

As global industries attempt to embed large AI models into power-constrained environments—such as autonomous vehicles, tactical defense drones, and factory robotics, the reliance on power-hungry data-centre GPUs is becoming economically and thermally unsustainable.

By establishing its Bengaluru Centre of Excellence, Mythic AI secures the multidisciplinary silicon engineering talent needed to commercialize low-power analog inference at scale. Simultaneously, positioning a combined engineering and business hub in South Asia places the company near emerging Asian automotive OEMs, robotics manufacturers, and hyperscale data-centre operators seeking energy-efficient alternatives to conventional silicon.

Curated by SSF Global

Tracking the shifts shaping GCCs, enterprise ecosystems, and the future of global business.

Share on      

SSF Global is a Global Community for Enterprise Function Leaders and serves as a research & advisory platform focused on Global Business Services (GBS), Global Capability Centres (GCCs), and the evolution of enterprise innovation in India and beyond. We track, publish, and partner in narratives that shape how capability centres transform into hubs of trust, intelligence, and sustainable growth. We also evaluate, assess and benchmark the GCCs for their performance, maturity and other parameters using our proprietary tools built from the knowledge gained from direct interaction with our members (GCCs & GBS).