Language can be frustratingly ambiguous. Or delightfully ambiguous. When you read the title of this blog, did you parse it as Google is a machine that is learning to compete? Or that machine learning will be how "the Google" competes? Both work, and both are true.
First meaning: there is clear evidence Google is making huge progress in cloud services to better compete against its rivals. Executives at the Google Next 17 conference cited a competitive win rate of 60% in the last quarter, with best results when the company gets a fair shot and customers dig deep into the technical differentiation. Sure, Microsoft is entrenched in most enterprises, and AWS has ridiculous momentum, but Google has invested $29 billion over the last three years to innovate in its own way. Many of the services' advantages are subtle but impactful, such as more granular billing for data warehouse consumption with BigQuery, custom configured compute instances, or the potential for API access to data services already within Google's domain. These have real benefits in reducing costs and increasing value of data. Machine learning even helps Google be more efficient, like finding ways to reduce data center cooling costs by 50%. As ESG research shows the financial cost/benefit equation is still the top perceived advantage for cloud-based databases, then Google should win simply on price efficiency for compute and storage resources. See a past comparison of costs here. Assuming buyers take the time to understand this and don't default to their Microsoft sales teams or Amazon's DevOps audience dominancy.
Second meaning: there is clear evidence Google is making huge progress in machine learning to better compete against its rivals. Winning a storage and compute resource price battle by changing the rules of pricing is typical Google-clever, but it's still a never-ending price battle, especially as most companies will end up arbitraging a hybrid and multi-cloud environment. Winning the long game will be done by getting the most insights out of the data, and this is where machine learning, deep learning, and artificial intelligence come into play. Here Google is taking a "have it your way," multi-pronged approach to meet mass-market customers where they are but help them go further than they thought possible, as CMO Alison Wagonfeld nicely describes it. APIs make it easy to treat AI as-a-service (AIaaS perhaps?) for your applications around popular general purpose functions like voice, text, translation, image, and video recognition. Anyone can use these. Let me say that again - anyone can use ultra-powerful AI capabilities for their apps today. That's stunning. For those who know what they're doing here, the Google Cloud Machine Learning service gives a bit more control and flexibility. For those who REALLY know what they're doing here, Tensorflow can be made to do anything. This all operates in a rich ecosystem including Dataproc, Dataflow, Dataprep for the management, BigQuery, Datalab, and Datastudio for discovery and analytics, and building from repositories like Cloud Storage, Cloud SQL, Cloud Datastore, and Cloud Bigtable. Oh, did I mention Cloud Spanner as a fully synchronized, globally distributed database? That too. The net result of this portfolio is that you don't have to be a prototypical Google genius to get value from machine learning and AI on your data. But if you do happen to be or become a Google genius, then there is no limit to what you can build.
IT market landscapes can also be frustratingly ambiguous or delightfully ambiguous. While many rush to crown AWS the winner in the cloud(s) and argue only over 2nd or 3rd place, Google Cloud SVP Diane Greene is right to point out the game has just started for most enterprises beginning digital transformation initiatives. Google is going to keep innovating and changing the rules to their competitive advantage. And their customers' competitive advantages, too. Machine learning and AI will be the most strategic play.