Machine learning startup Graphcore Ltd. has raised $50 million in new funding in a round led by Sequoia Capital, according to news published Sunday but not yet confirmed by the company itself.

It’s also not clear what the valuation was on the deal. A report from Bloomberg Nov. 3 said that the company had been chasing a valuation of $1 billion but “wasn’t able to reach that height.”

Founded in 2016, United Kingdom-based Graphcore is building “intelligence processing units,” chips that are specifically designed to assist programmers in creating machine learning computer systems that can be used in fields such as with autonomous cars or medical detection devices. The company claims its IPU accelerators and Poplar software framework deliver “the fastest and most flexible platform for current and future machine intelligence applications, lowering the cost of AI in the cloud and datacenter, improving performance and efficiency by between 10x to 100x.”

When the company last raised venture capital in July, it was clear that although building chips dedicated to artificial intelligence and machine learning isn’t a new concept, the market itself is best described as emerging. Others looking to capitalize on growing demand from machine learning processors include Google LLC through its “Tensor Processing Units” and Intel Corp. with its recently revealed NNP family of chips.

According to a report from Allied Market Research, the machine learning chip market is set to rise to $8.2 billion by 2022, offering ample opportunities to gain market share. Machine learning chips are widely used across the applications such as robotics, healthcare, automotive and consumer electronics, the researchers noted. “At present, rising demand for automated electronic devices and trending artificial intelligence are some factors that majorly drive the market,” they said. “Moreover, popularity of internet of things is expected to provide lucrative opportunities to market players.”

Including the new round, Graphcore has raised $110 million to date.

Image: Graphcore