Companies Use AI Outside the Cloud, Trimming Costs.

"Companies have started leveraging advances in networking, algorithms and edge computing to run artificial-intelligence workloads outside of data centers and closer to where applications are being put to use.

Artificial intelligence requires massive amounts of computing power, which is typically provided by stacks of servers and other equipment in large data-center facilities owned or operated by cloud vendors.

Corporate technology chiefs say emerging systems designed to distribute heavy workloads across networks of linked computers have the potential to lower cloud costs while reducing latency by feeding real-time data directly back into AI models at the source.

"Once a model is trained, subsets of it can be deployed closer to where new data is generated," said Dave McCarthy, research vice president at International Data Corp. "This distributed concept allows AI to scale effectively when put into use," McCarthy said.

Known as decentralized clouds, or distributed data centers, the networks of shared processing power are often coupled with Internet of Things and edge computing,where computer processing is done as close to the source of data as possible.

DHL Supply Chain, a unit of Germany-based courier Deutsche Post, is using a decentralized cloud system to run AI-powered computer-vision applications designed to enable warehouse robots to identify and handle thousands of packages, said Sally Miller, chief information officer of DHL's North American supply-chain business.

"We have also used AI solutions leveraging decentralized cloud tools for vision applications and augmented reality to aid warehouse workers in picking orders," Miller said.

Some of the company's warehouses exceed 1.6 million square feet and hold several million packages, each with unique shipping instructions, the company said. Rather than bouncing large amounts of data back and forth between an external cloud provider, the underlying AI software driving DHL's warehouse robots runs entirely on-site, Miller said.

Todd Florence, chief information officer of Estes Express Lines, a Richmond, Va.-based shipper, said the company is using a similar system to operate AI-enabled vision software in truck-mounted cameras designed to alert drivers to road hazards ahead. "We no longer have to wait for data to be streamed to a centralized service, either in the cloud or our own data center," he said. "This improves safety overall by nudging people in the moment."

Florence said the company's early investments in AI-powered computer-vision systems relied on centralized, cloud-based processing. Reducing the costs of added data storage and computing resources for AI tools was a key factor in switching to a distributed system, Florence said, adding that the bigger impact has been improved performance.

"Traditional computing infrastructure is no longer suited for how these organizations operate, especially as they continue to rely on real-time insights," said Evan Welbourne, head of AI and data at Samsara, a San Francisco-based Internet of Things software firm that helped develop Estes's onboard AI capabilities.

As the volume of data continues to increase, the ability to leverage AI at the edge has become critical, Welbourne said, citing applications in transportation, construction, manufacturing and other sectors.

Beyond running AI apps outside of the cloud, decentralized cloud developers are currently working on systems designed to enable businesses to train their own AI models -- a more power-intensive process." [1]

1. Business News: Companies Use AI Outside the Cloud, Trimming Costs. Loten, Angus. 
Wall Street Journal, Eastern edition; New York, N.Y. [New York, N.Y]. 17 Aug 2023: B.3.

 ,