What happened
Symatics, a Singapore-headquartered semiconductor startup, unveiled its Morphic Processing Unit (MPU) architecture at the Business & Philanthropy Forum 2026 in Singapore. The MPU reorganises AI computing around row-column parallelism, built for interactive AI, complex reasoning, and AI agents that depend on how fast a single request can be completed.
The MPU maps the parallel unfolding of a computation, and the dependencies between its steps, onto a row-column array, so data supply is matched precisely to where each calculation takes place.
Intermediate results stay on the chip and are used immediately nearby, while computing state is retained and continuously updated.
A compiler plans where computations take place, how information moves between them and the order in which operations are executed.
By determining these steps in advance, the architecture aims to reduce the time and resources spent coordinating tasks during execution.
The architecture also aims to improve the speed of individual AI requests, rather than focusing solely on the total number of requests a system can process.
Source details: us.acrofan.com
Why it matters
The Morphic Processing Unit (MPU) architecture is designed to address the underlying inefficiencies in AI computing and support a new generation of AI applications. By reducing unnecessary external memory reads and writes and data redistribution, the MPU aims to improve the speed of individual AI requests and give AI services a competitive edge where response time matters most.
The Morphic Processing Unit (MPU) architecture is designed to address the underlying inefficiencies in AI computing and support a new generation of AI applications.
By reducing unnecessary external memory reads and writes and data redistribution, the MPU aims to improve the speed of individual AI requests and give AI services a competitive edge where response time matters most.
Symatics will support partners in customising chips and systems, and share reference designs and implementation experience to build an open ecosystem.
The company will also build a Singapore implementation standard and demonstration projects that validate the path from model compilation to deployment.
Symatics will connect Asia-Pacific manufacturing, strengthen local design services and offer customised chip and system engineering for enterprises and national-level customers.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
A route planner searches possible journeys using explicit rules. What does this illustrate about AI?
What to watch next
Symatics will develop the architecture in collaboration with research and industry partners in Singapore, supporting the company's global ambitions while contributing to the region's AI computing capabilities.
Symatics will develop the architecture in collaboration with research and industry partners in Singapore.
The company will support partners in customising chips and systems, and share reference designs and implementation experience to build an open ecosystem.
Symatics will build a Singapore implementation standard and demonstration projects that validate the path from model compilation to deployment.
The company will connect Asia-Pacific manufacturing, strengthen local design services and offer customised chip and system engineering for enterprises and national-level customers.
Symatics will advance its architecture through research and engineering development, with ambitions to serve the global AI infrastructure market.