The Austin-based startup has opened an engineering hub in India’s tech capital, and it’s not just a support office — Bengaluru is already doing a quarter of the work on its next flagship chip.
The Mythic AI Bengaluru Centre is now officially open, and it’s arriving with more substance than the usual “global capability centre” press release. The US semiconductor startup, known for chips that do AI math inside memory instead of shuttling data back and forth to a processor, launched its Centre of Excellence (CoE) in the city on September 23. Its India team, currently around 20 engineers, is set to more than triple to over 60 people by the end of Q1 2027 — and it’s being handed real engineering ownership from day one, not just testing and support work.
That distinction matters. Chairman and CEO Taner Ozcelik, a former Nvidia executive, told Indian media this week that the Bengaluru unit already contributes close to a quarter of the engineering work on Vanguard, Mythic’s next-generation AI inference platform — and that share could grow to nearly half as the centre matures. For a global chipmaker’s first India centre, that’s an unusually fast ramp into core product work.
Why the Mythic AI Bengaluru Centre Is a Bigger Deal Than It Looks
Founded in 2012 and headquartered in Austin, Texas, Mythic builds analog compute-in-memory chips — a design where computation happens directly inside the memory array rather than on a separate logic core. The company says this approach can deliver dramatically better energy efficiency for AI inference than conventional GPU-based systems, since it avoids the constant back-and-forth of moving data between memory and processor that eats up power in traditional chips.
The Bengaluru CoE isn’t starting from scratch on side projects. It’s plugged straight into Vanguard, Mythic’s flagship next-gen platform, working across front-end chip architecture, physical design, and design-for-test (DFT) — the unglamorous but critical disciplines that decide whether a chip actually works at scale. From there, the mandate expands into product engineering, systems engineering, testing and characterisation, and intelligent sensing.
Ozcelik has framed it plainly: he wants Mythic chips eventually powering autonomous systems on Bengaluru’s own roads — tying the centre’s engineering work directly to the market it’s meant to serve.
Data Centres, Cars, Robots, Drones: Mythic’s Four-Market India Playbook
Mythic has publicly named four target industries — a list it first laid out alongside its December 2025 funding round — and all four map onto specific India opportunities:
Where Mythic wants to sell in India
- Data centres — positioned as the first wedge, as rising AI compute demand collides with India’s power-infrastructure constraints.
- Automotive — Mythic already has a joint development agreement with Honda for an automotive AI system-on-chip, aimed at software-defined vehicles by the late 2020s/early 2030s.
- Robotics & factory automation — Ozcelik has flagged India’s growing use of mobile robots and industrial automation as a near-term opening.
- Defence & intelligent sensing — low-power edge chips for drones and sensor systems, aligning with India’s push for indigenous defence electronics.
There’s also a longer-horizon piece: Mythic currently manufactures its chips in Germany and Arizona, with packaging done elsewhere in Asia, and has not announced any India manufacturing plans. But Ozcelik has said chip production in India becomes more plausible as the country’s fabrication ecosystem matures — and that Mythic’s die-to-die advanced packaging work, which it positions as a rival to TSMC’s CoWoS technology, could eventually see part of its development driven out of India.
The Reality Check: What’s Confirmed vs. What’s a Company Claim
What It Means for Indian Chip Engineers and the AI Ecosystem
Mythic isn’t the first global chip company circling Bengaluru this year — Marvell has separately outlined plans to grow its roughly 1,700-strong India workforce by 15% annually to chase AI infrastructure demand. What makes Mythic’s move notable is the seniority of the work: VLSI design, physical design, and DFT are core silicon-engineering roles, not adjacent support functions. For engineers with analog design, RTL, or physical-design backgrounds, this adds another serious employer to Bengaluru’s semiconductor hiring pool — one explicitly building toward defence and automotive-grade chips rather than just software layers on top of someone else’s silicon.
It also feeds a broader pattern worth watching: India’s AI chip story in 2026 isn’t just about being a low-cost design outsourcing hub anymore. Companies are starting to hand Indian teams real product ownership on flagship chips before those chips even ship — a shift that, if it holds, changes what “AI chip jobs in India” actually means over the next two to three years.
FAQs
What does Mythic AI’s Bengaluru centre actually do?
It works on core engineering for Mythic’s next-gen Vanguard AI chip — chip architecture, physical design, and testing — not just support work.
How many people will Mythic hire in India?
The team is around 20 people now and is expected to grow to 60+ by the end of Q1 2027, per CEO Taner Ozcelik.
Will Mythic manufacture chips in India?
Not yet — chips are currently made in Germany and Arizona. India manufacturing is a stated future possibility, not an announced plan. </p>
Disclaimer: This article is based on statements made by Mythic AI’s leadership to the above publications. Figures such as efficiency claims, hiring targets, and manufacturing plans are company-stated and have not been independently verified by aitechnews.in.