Thursday, November 23, 2017

‘Revolutions do not happen on their own’. Why Lenovo has made its strongest bet on AI innovation data centres ever

Building on its largest data centre portfolio release in history, Lenovo is also planning to dive into the AI world to advance how supercomputers and data centres operate in the 21st century. João Marques Lima reports from Lenovo Transform in New York.

The use of artificial intelligence (AI) in data centres and supercomputing environments has been hailed by many as the next big revolution in computing.

The ability to automate, speed up and scale up IT environments to numbers never seen before, are making IT engineers, researchers and tech enthusiastic dream about the endless applications of AI.

Those dreams are being fuelled by an overall capital expenditure in AI technologies that are set to boom to nearly $37bn by 2025, up from $644m in 2017, according to Tractica.

This also includes figures based on the use of AI technology in data centres and supercomputers which will drive automation and speed of analytics across many fields, including finance, healthcare and research.

Trying to grab a large share of this new intelligent economy is Lenovo, which has set out a roadmap for the future around research and development of AI in supercomputing and data centres, which the company sees as a major pillar to its roadmap going forward.

Speaking at Lenovo Transform, the company’s chairman and CEO Yang Yuanqing spoke of the “astonishing growth” of data and advancements in AI.

“The computing power can now process the massive amount of data from your customers, suppliers, partners, competitors, and much more,” he said. “It can also give the insight you could not imagine before.”

Ultimately, for Yuanqing, this new age of computing capabilities when looked at form a business perspective are about harnessing what he calls “the intelligent revolution”.

“This is not just about making better business decisions; it is about making the world a better place.

“Could we build a new world? A world with no diseases, war and poverty.

“Revolutions do not happen on their own. Every industrial revolution has its leaders, its visionaries and heroes. In the fourth revolution, those heroes are the Massive Transformers of their age.

“The Massive Transformers of the next revolution are you; you are the designers and the builders of the network and the systems; you bring the benefits of intelligence to any corner of the enterprise.

“I challenge every single one of you to cease this opportunity and become a Master Transformer.”

Building on this, Peter Hortensius, SVP, CTO and chief strategist at Lenovo’s Data Centre Group, also spoke of how AI will change and make the world “a better place”.

“AI will deliver significant change not just in our industry but all industries. This is on the scale of the agricultural, industrial and digital revolutions,” he said.

“AI will accelerate large scale solutions, such as cure diseases, prevent crime and unlock dark data. It will unleash scientific discovery, enabling the exploration of new worlds, decode the brain and uncover things we have never imagined.

“This will also extend human capabilities, such as personalise learning experiences, enhance decisions and optimise time.

“And lastly, it will help us automate undesirable tasks, such as automate driving, save lives in danger and perform cores.”


Fighting AI barriers with global AI Innovation Centres

Despite the many promises of AI, the applications of the technology are only possible once the infrastructure that supports such workloads and applications is built.

Speaking of the barriers to adoption in any standard enterprise, Hortensius highlighted that data, either structured or unstructured is still one of the main challenges to any business because people have the data but they do not know how to use it.

“What makes the cloud powerful is not the compute power, is the data that sits in there,” he said.

This goes down to the second top barrier, which is expertise. “What they do not have is the AI expertise and consultants. AI is different, the language, the complexity, etc.

“Thirdly, they do not have the resources to do it as well because this is not cheap.”

With that in mind, Hortensius moved on to announce that Lenovo will be opening three AI Innovation Centres.

The centres will open in October 2017 and will be based in North America (Morrisville, US), Europe (Stuttgart, Germany) and Asia (Beijing, China).

The data centre facilities will run the company’s new infrastructure portfolio ThinkSystem and will be used by research institutions, technology partners and ecosystem partners.

Deep learning, open source and machine learning frameworks to be used on site include TensorFlow, Caffe2, Torch, Theano and mxnet.

“These innovation centres will give access to experts, infrastructure and tools for a fast start [in using the technology to solve business problems],” said Hortensius.

The executive did not reveal the capital expenditure associated with the establishment of such hubs, but said “they are not cheap”. Stuttgart is already home to a Lenovo data centre innovation centre.

“Our job in the Data Centre Group is not to invent AI, but we do know an awful lot about how to make things run faster.

“[However,] we are still far from [massive adoption of AI technologies]. From my experience, these things go slow, but then someone will build an application and this will just speed up.”