Google Cloud uses Anthos and AutoML to differentiate itself

Google Cloud now uses Anthos and AutoML to distinguish itself from market leaders Amazon Web Services (AWS) and Microsoft Azure, chief AI scientist Andrew Moore, and product manager Rajen Sheth told reporters.

“If I run a small business or startup that depends on the technology of the cloud provider, if I go with Google, I can actually sell customers the ability to run their models on premise, or on GCP, or other clouds, so this huge flexibility helps,” Moore said Thursday.

Google Cloud made Anthos, a hybrid cloud management system that connects with AWS and Azure, a lot more flexible this week. Moore, who became Google Cloud’s head of AI in late 2018, also called BigQuery and Google’s major product differentiators.

In an increasingly competitive cloud market, Google pitched the world by releasing dozens of new products and services at this week’s Next Conference in San Francisco. Products introduced include new classes for AutoML, premade Retail and Contact Center AI services collection, and AI Platform, a collaborative model-making tool.

Developers with little coding experience can use AutoML, while AI Platform is for data scientists— part of Google’s attempt to deliver creator AI tools across a range of experiences. With AutoML, we can also extend this platform to a wider range of people, so you don’t have to be a data scientist who understands how to use a notebook.

You can go to the developer to the other end and connect the developer with a data scientist. I think Google’s key distinction, Sheth said. Because of Google Cloud’s AI services suite, Frank doesn’t find this AI push surprising. AI capabilities are a key reason companies consider adopting Google Cloud, so it’s important that the tech giant retains a strong position as it works to attract new businesses, he added.

Moor Insights analyst Patrick Moorhead agreed that Cloud Next’s BigQuery ML developments were noteworthy, but he said AI services are a declining distinction between Google and other public clouds. He added that this week, Anthos could be Google’s most important new product. Anthos was definitely the most impressive of the announcements, basically because the market leaders out there are not talking too much about multi-cloud, Moorhead told VentureBeat.

He noted that there are still a number of unanswered questions about Anthos, as clouds have different computing and storage instances and security models. The notion of doing it easily and arbitrating between three different public clouds seems more complex than shown. Another potential drawback for Anthos might be that it only works with Kubernetes containers, a format that can exclude most on-site workloads.

While Google has made open source projects— like TensorFlow and Kubernetes— that have become important to many AI developers, this has not translated into market share in cloud. Indeed, competition between major cloud providers resulted in each offering similar AI services. From Google TPUs to Microsoft FPGAs and AWS Inferentia, each firm offers hardware to accelerate AI inference and training.

Each has services to pull unstructured document data, recommendation, vision, and voice AI services. I think we’re going to enter a benchmarking phase. People generally hate benchmarking, but it becomes necessary now that so many services are similar.