We are creating a novel entrepreneurial engine to generate semiconductor technologies that drive optimal AI processing exploiting IP across a diverse range of applications and markets.
The Edinburgh Venture builder for AI-hardware, EVA, will co-locate researchers, innovators and investors creating critical mass in a thematically cohesive business. This will translate academic research that has clear market fit, adopting an agile approach to commercialisation that capitalises on successes and failures, and upskills a dedicated resource to deliver viable technologies underpinned by world-leading research.
Partners will benefit from:
Challenge:
Land and resource usage for data centres and Telecomms is unsustainable
Solution:
Support up to 500x more computations per second
Challenge:
Large and slow AI models mean more servers serving fewer customers
Solution:
Reach up to 100x faster access to data
Challenge:
Datacentre energy demands are expected to reach up to 20% of global consumption if unchecked
Solution:
Use up to 1000x less energy for computing
Challenge:
Cloud computing is vulnerable to cybersecurity, privacy and trust breaches
Solution:
Deliver training capabilities at the edge making it 100x cheaper to access AI capabilities
Challenge:
Miniaturising electronics is increasingly unsustainable demanding new methods to process data
Solution:
Deploy AI at 100x smaller form-factors
Regius Chair of Engineering
RAEng Chair in Emerging Technologies
