Before the 9,000 attendees at the AWS Summit in San Francisco could find their seats at the conference’s morning keynote session today, Amazon Web Services Inc. had already announced the addition of image publishing service Shutterfly Inc. and vehicle services firm Cox Automotive Inc. to its expanding list of cloud customers.

“They are going all-in on AWS,” said Werner Vogels (pictured), chief technology officer at Amazon, who mentioned that Cox would be closing 40 data centers as part of the transition. “It’s a pattern that we see continuing.”

Part of the driver of those deals, as well as a string of earlier wins, is machine learning. Amazon’s efforts to provide its customers with an array of intelligent applications are being channeled through SageMaker, a relatively new cloud service designed to streamline the building and training of models for analyzing and getting value from enterprise data.

“Tens of thousands of customers are running machine learning on AWS,” said Matt Wood, director of artificial intelligence for AWS. “Machine learning is experiencing a renaissance in the cloud.”

Indeed, John Furrier, co-host of theCUBE, SiliconANGLE Media’s mobile livestreaming studio, said in an analysis of Vogels’ keynote today that the real story of the one-day conference was SageMaker. “My prediction is that SageMaker will surpass Aurora as the No. 1-shipping product,” he said.

Local mode for SageMaker

Wednesday’s announcements included the addition of a “local mode” tool for SageMaker that will allow developers to train machine learning models on their notebook computers. General availability was also announced for Amazon Translate and Amazon Transcribe, artificial intelligence-fueled language and text-to-speech translation services.

In addition, the company revealed that it would add the latest versions of the widely used deep learning software frameworks TensorFlow and MXNet to SageMaker and make them open-source. “The dirty secret of machine learning right now is that it’s like a primordial soup,” Wood said. “We’re trying to take machine learning and put it into the hands of every developer.”

That’s helping speed up deployment of machine learning models at some companies. Nhung Ho, data science manager for Intuit Inc.’s small business division, said in an interview that SageMaker allowed her team to create models in just a week, versus a minimum of six months previously, by making it easy to create the model on a laptop and scale it up to many computer servers.

“AWS allows us not to reinvent the wheel,” she said. “SageMaker enables us to be very fast.”

Stu Miniman, cohost of theCUBE, said SageMaker signals AWS’ broader ambitions in an era where AI and machine learning are driving countless new services from image and speech recognition to self-driving cars. “Amazon isn’t just about infrastructure and cloud anymore,” he said.

DeepLens now for sale

One of the more intriguing products highlighted by AWS on Wednesday was DeepLens, a deep learning-enabled video camera. The high-definition device, unveiled at AWS re:Invent last November, represents the coming wave of intelligent products capable of processing large streams of data at the edge of networks.

The company has given a number of the cameras away to developers for free and it is now available for preorder on Amazon for the list price of $249. Among the use cases cited by company executives on Wednesday was the identification of various dog species at a local park. “We’ve seen a lot of excitement from customers who are using DeepLens,” Wood said.

Marketplace reaches 170,000 customers

Aside from its focus on machine learning, AWS briefly highlighted a number of other cloud-related services. AWS Marketplace, the company’s digital catalog for finding and deploying software to run on the AWS cloud, now has 170,000 active customers and over 4,100 listings.

“If you’re a software vendor, you are running on AWS,” Vogels said. “Why? Because the customers are there.”

Amazon Aurora, a PostgreSQL- and MySQL-compatible relational database built for the cloud that was launched in 2014, is continuing to gain enterprise acceptance. “It’s the fastest-growing service in the history of AWS,” Vogels said. “We’re giving [customers] an enterprise grade database at an open-source cost.”

Essentially, AWS is gearing its massive and rapidly growing portfolio of services to be the one-stop shop for whatever the enterprise technology job demands.

Still, though AWS has so far dominated cloud computing, it will face more intense competition going forward, both from public cloud rivals such as Microsoft Corp. and Google LLC but also information technology stalwarts such as Oracle Corp. and IBM Corp.

“The major cloud players are going to fight tooth and nail for market share,” Furrier said. “You’re seeing Amazon for the first time dealing with competitive pressure that’s old-school tactics.” Indeed, Bloomberg reported today that Oracle co-Chief Executive Safra Catz had dinner with President Donald Trump Tuesday evening and expressed concerns about the government’s cloud contracts with AWS.

Furrier and Miniman provided more analysis of the key announcements today:

(* Disclosure: TheCUBE is a paid media partner for AWS Summit. Neither AWS nor other sponsors have editorial control over interviews or content on theCUBE or SiliconANGLE.)

With reporting from Robert Hof

Photo: Robert Hof/SiliconANGLE