What optoML does

optoML is a fabless semiconductor startup developing Palani, an AI inference silicon platform based on a hybrid analog-digital compute fabric that combines analog in-memory compute (AiMC) with optical interconnects on a 12nm FinFET TSMC process. The company claims its patented architecture delivers up to 50x greater energy efficiency than conventional digital AI accelerators by moving compute closer to where data lives and reducing costly data movement between memory and processing units. The Palani platform is positioned across three deployment classes — compact edge AI (cameras, gateways, embedded systems), multimodal sensor fusion (robotics, autonomous systems, industrial control), and scalable infrastructure inference — with a digital control plane for scheduling, orchestration, and model execution.

optoML is domiciled in Singapore with engineering headquarters in Bengaluru, India, and operates a fabless model with TSMC as its wafer manufacturing partner and Kaynes Semicon signed under MoU for assembly, testing, and post-silicon validation. The company is led by founder Saravana Maruthamuthu, who brings over 17 years of semiconductor experience from Intel, Qualcomm, and Continental.

Commercially, optoML is at a pre-revenue, silicon validation stage. Revenue mechanics are framed around hardware sales of the Palani SoC and selective IP licensing, but all commercial and technical engagements are conducted through NDA-based briefings with strategic investors, partners, and customers rather than open distribution. The company secured a $1.8 million Pre-Series A round in February 2026, co-led by Bluehill.VC and A99, with use of funds directed at hiring expansion, next-generation chip development, and TSMC wafer manufacturing follow-up.

Name

optoML

Legal name

optoML Semiconductors

Website

https://opto-ml.com

Company type

Private

Founded year

2024

Operating status

Operating

Headcount range

11–50 employees

Short description

optoML is a fabless semiconductor startup building Palani, a 12nm FinFET AI inference platform combining analog in-memory compute with optical interconnects for edge devices, sensor fusion systems, and scalable infrastructure, currently at the silicon validation stage.

Ownership category

akta.pro rank

Product category

AI Inference Semiconductors

NAICS

Computer Systems Design Services (541512)

SIC

Services-Computer Integrated Systems Design (7373)

akta.pro primary industry

AI/Neural Processing Units (NPUs) & ML Accelerators (HDAHAJAF)

akta.pro secondary industries

Edge AI Hardware & NPUs (SoCs, modules, accelerators) (HDAAAJAC), AI Application Enablement Platforms (Copilot/Agent Frameworks, SDKs) (HDAEANAJ)

Keywords

• AI inference silicon
• Analog in-memory compute
• Fabricated semiconductors
• Optical interconnect chips
• Edge AI accelerators

Headquarters

HQ city

Singapore

HQ country

Singapore

HQ region

Asia

Offices2 records

Markets served

GTM type

B2B

Offering type

Hardware or Manufacturing

Cost components

Technology or R&D, Supply Chain, Personnel, Operations, Infrastructure

Revenue model

1. AI Inference Semiconductor Sales: Fabless semiconductor company generating revenue through the sale of AI inference silicon — the Palani platform — as a complete SoC offering targeting edge devices, sensor fusion systems, and scalable infrastructure deployments.
2. IP Licensing: The company operates a protected semiconductor IP model with selective sharing under NDA, suggesting potential revenue from licensing its analog in-memory compute architecture and design assets to strategic partners and customers.

Go-to-market motion1 record

Distribution channels1 record

Marketing channels1 record

Core offering

optoML builds Palani, an AI inference silicon platform that combines analog in-memory compute with optical interconnects on a 12nm FinFET process, delivering up to 50x greater energy efficiency than conventional digital accelerators. The platform is offered as a complete SoC with a digital control plane, system interfaces, and an in-memory software compiler for model mapping and deployment, targeting edge AI devices, sensor fusion systems, and scalable infrastructure inference deployments.

Product overview

optoML offers Palani, a unified AI inference platform built on analog in-memory compute architecture with optical interconnects. Palani is designed as a complete SoC offering that provides efficient AI execution through a hybrid analog-digital compute fabric, targeting edge AI, sensor fusion, and infrastructure inference applications.

Differentiator

Problem solved

Functional benefit

Brands

• Palani: AI inference silicon platform using analog in-memory compute architecture with optical interconnects, designed for edge systems, sensor fusion applications, and scalable infrastructure.

Products and services

• Palani Palani is optoML's AI inference silicon platform that combines analog in-memory compute with optical interconnects on a 12nm FinFET process, delivering up to 50x greater energy efficiency than conventional digital AI accelerators. It is offered as a complete SoC with a digital control plane, system interfaces, and an in-memory software compiler, targeting edge devices, sensor fusion systems, and scalable infrastructure deployments. The platform supports vision, signal, and multimodal sensor data processing for industrial AI vision, robotics, autonomy, and large-scale inference applications.

Quantifiable outcome

• 50x greater energy efficiency compared to conventional digital AI accelerators
• +2 more outcomes

Segments3 records

Ideal customer profiles3 records

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

AI capability7 records

Feature6 records

Partnerships

Two partnerships are on record, tiered core.

• Kaynes Semicon (Kaynes Semiconductor)coreTechnology or Integration · 25 February 2026optoML has signed an MoU with Kaynes Semicon for assembly and testing services once wafers are delivered from TSMC. The partnership supports chip packaging, post-silicon validation, and the broader path to system readiness as part of optoML's India execution strategy.
• TSMCcoreTechnology or Integration · 24 February 2026optoML has completed its 12nm tapeout with TSMC for its core AiMC architecture and SoC integration path. TSMC serves as the primary wafer manufacturing partner, producing the silicon wafers that will subsequently undergo assembly and testing through Kaynes Semicon before product realization.

Scale indicators4 records

Recent moves5 records

Expansion highlights5 records

Broad incumbents

• Intel (Habana Labs / Movidius): Intel's AI silicon portfolio including Habana Gaudi data center accelerators and Movidius edge VPUs provides overlapping inference silicon competition for both edge and infrastructure deployment classes.
• Qualcomm: Large incumbent semiconductor company with AI inference capabilities across Hexagon NPU, Snapdragon, and Cloud AI 100 lines. Relevant as the broad incumbent optoML will compete against for edge and infrastructure inference design-wins.
• NVIDIA: Dominant AI accelerator supplier whose GPUs/TPUs set the energy-efficiency and software-toolchain benchmark optoML is competing against. Defines buyer expectations and qualification standards in the inference market.

Direct peers

• Blaize: AI inference chip company offering programmable graph-streaming processors for edge and data center workloads. Competes in the same AI inference silicon category as optoML.
• SambaNova Systems: AI accelerator company offering domain-specific RDU processors for enterprise and inference workloads. Comparable as a fabless AI silicon company pursuing enterprise inference deployments.
• Tenstorrent: AI accelerator startup building RISC-V-based scalable inference and training chips. Comparable as a fabless AI silicon startup targeting edge-to-data-center inference.
• Groq: AI inference accelerator company using a deterministic LPU architecture for low-latency, high-throughput inference. A relevant peer for optoML's infrastructure-inference deployment class.
• Syntiant: Fabless semiconductor company producing ultra-low-power edge AI NPUs for always-on sensing and audio/vision inference. Directly comparable in form factor, power envelope, and edge-AI focus.
• Hailo: Israeli fabless semiconductor company building edge AI accelerators for cameras, automotive, and industrial vision. Targets similar low-power, low-latency inference workloads at the edge as optoML.
• Mythic AI: U.S. fabless semiconductor startup developing analog in-memory compute (AiMC) AI inference chips. The closest architectural analog to optoML's Palani platform, competing in the same analog compute category for edge AI inference.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat3 records

Key risks6 records

Key highlights7 records

Customer concentration

Revenue estimate

Valuation estimate

Number of profiles

Profiles1 record

Funding overview

Funding rounds1 record

Investors2 records

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about optoML