GraphcoreGraphcore, a Bristol-based computer chip startup just raised more than $30 million to take on Intel and Nvidia, two of the biggest industries in the rapid growing and promising market of Artificial Intelligence (AI) hardware.

The young U.K. startup hopes to design hardware to make machine learning faster and easier, by reducing the costs for faster AI applications in the cloud and bringing AI to robotics, driverless cars and to consumer devices.

With processors 100 times more powerful than today’s chips, that require large amounts of data to teach AI systems, a goal which Graphcore hopes to achieve by 2017.

The AI industry is expected to skyrocket from $14.9 billion to $70 billion according to a research conducted by the IDC as reported by the Bank of America Merrill Lynch, with the hardware spending on deep-learning to be at $41.5 billion by 2024, according to Tractica.

Hence, it is not surprising indeed, that some of its backers include Samsung, German engineer Bosch and Hermann Hauser and the founder of Cambridge-based chip designer Arm, hoping to leverage off this lucrative industry.

Nigel Toon, CEO and co-founder of Graphcore added on their financial backers’ motives by saying

“We have Bosch as a strategic investor, we have Samsung as an investor and Bosch is interested in autonomous vehicles and the next generation of transportation, while Samsung is interested in missing word edge of network devices,”

 

 

Graphcore
Nigel Toon, CEO and co-founder of Graphcore

 

 

One thing Graphcore hopes to move away from, are the currently used Graphic Processing Units (GPUs).

These are used to run machine-learning applications but Toon said in an interview with CNBC, that

“GPUs have been built to run programs that completely describe the algorithm. Machine learning is different. You are trying to teach the system using data and that requires a different style of compute,”

 

This is a giant leap for a two year old chip startup with just 40 employees; however, they are certainly not lacking in determination to emerge as the industry leaders as Toon added that

“We are not lacking in ambition,” he said. “The energy in our team as we start our journey is remarkable. The innovation they are unlocking is groundbreaking. We look forward to sharing more about our technology and helping you to create the next generation of machine intelligent applications, services, products and devices.”

 

 

The big challenge for Graphcore, currently, is facing the transition from conventional pocessors (the GPUs), which are widely used currently in AI algorithms and programs that learns to recognize objects, understand human language and drive cars, to its new and upgraded chips.

The new upgraded chips by Graphcore, are capable of exponentially powering up power machine learning systems. Toon further added

 

 “With standard chips, you end up using an awful lot of servers, so we can reduce [numbers of] machines, save a huge amount of power and make the system a lot more efficient, which can generate a 5-10 times advantage over existing technology,”

“For new and emerging applications like driverless cars, where we think today’s GPUs will struggle, we think our advantage grows to 50-100 times.”

 

 

On the road to change the AI industry

There is no denying that, Graphcore is indeed on the right path to becoming an AI chip industry giant, and with the additional financial backing, Nigel Toon said about Graphcore, that they could

 

“innovate faster and execute at scale with backing from some of the biggest names in the technology industry”.

“Machine learning is becoming a major market and Graphcore has the technology to lead this next wave of computing.”

 

 

Follow me on Snapchat, and I’ll keep you updated with what’s going on every now and then. 🙂