ElastixAI, a Seattle, WA-based provider of a software platform that improves FPGA-based servers, has raised $18 million in seed funding.
The investors were not disclosed.
Read More:BreezeBio Closes $60 million in Series B Funding
The company plans to use the funds to grow its operations and advance its development work.
ElastixAI offers a direct replacement for older GPU workflows keeping them compatible while greatly improving efficiency. Key benefits include:
Up to 50x TCO Advantage: ElastixAI runs LLM operations up to 50x more efficiently than standard GPU kernels.
80% Lower Power Use: It powers only the circuits needed for inference, avoiding wasted energy.
Always Current: While custom chips take years to develop, AI evolves daily—ElastixAI lets you run the latest AI on existing hardware.
The AI inference market is projected to reach $255 billion by 2030, but current infrastructure isn’t well-suited for GenAI. LLM inference relies on memory while standard GPUs are built for compute heavy tasks like training, leading to low utilization and wasted capital and energy. Custom chips are designed years before production lagging behind the latest ML advances. For instance, 4 bit quantization, which should double performance, often only improves it by about 10% on hardware without native support.
“We’re moving beyond ‘one-size-fits-all’ hardware,” said Mohammad Rastegari. “By using our custom post-training optimizations on FPGAs, we let the hardware adapt to the model instead of making the model work around the hardware.”
“The industry is losing a huge amount of performance because hardware can’t keep up with ML advances,” said Mohammad Rastegari, PhD, co-founder of ElastixAI. “We’re moving beyond ‘one-size-fits-all’ hardware. Our proprietary post-training optimizations let FPGAs adapt to the model instead of making the model work around the hardware.”
About ElastixAI
Founded in 2025 by CEO Mohammad Rastegari, ElastixAI addresses the inefficiencies of GenAI inference with innovative software, ML and hardware co-design, providing the next generation of scalable, sustainable AI. The founding team has deep expertise in ML, software, and hardware from top tech companies and academic institutions, with over 24,000 combined citations.
Read More:Profitmind Raises $9M in Series A Funding
