Olix Raises $312M for AI Inference Chips
Olix raised $312M at a $3.3B valuation to build optical processors for AI inference, adding another signal to the specialised compute race.
The AI chip race is moving beyond GPUs and into more specialised ways of running models efficiently.
What happened
Olix raised a $312 million Series B at a $3.3 billion valuation. The startup is building optical processors for AI inference and plans to bring products to market next year.
Inference is the process of running AI models after they have been trained. As AI moves into everyday products and enterprise workflows, inference costs become a major operating expense.
Why it matters
This is a major deep-tech funding signal. Training models gets most of the attention, but the long-term economics of AI may depend more on how cheaply and efficiently models can be run at scale.
Specialised compute architectures are trying to reduce the energy, latency and cost burden of inference. If optical processors can deliver real efficiency gains, they could become an important alternative in the AI infrastructure stack.
The bigger picture
AI infrastructure is fragmenting into more specialised layers: GPUs, memory, networking, data centres, power and now alternative processors.
Olix’s round shows investors are still looking for ways to attack the cost side of AI, especially as enterprises move from model experiments to always-on AI systems.