Oracle has finally announced its ambitions to be a competing force in the cloud artificial intelligence(AI) services market. At its annual Open World Conference, the software giant announced a new suite of tools and services to enable the implementation of aI solutions. Typical Oracle, the announcement was full of marketing non-sense already claiming a leadership position in the meerging market. However, if we read between the marketing red-tape, there are some interesting implications of the new Oracle offering for the AI market.
What Did Oracle Announced?
It’s been a few days since Open World which means that the marketing hype has settled down a bit and we can take a more pragmatic view at Oracle’s AI offering. The core of Oracle’s release is the new AI Platform Cloud Service which can be seen as a workbench for the implementation of cloud data science applications.
The Oracle AI Platform Cloud Service includes a large GPU-based computation runtime for the execution of AI programs. The runtime is already integrated with popular AI technologies such as TensorFlow, Keras, Scikit-Learn and Jupyter Notebooks. The integration with Jupyter is particularly interesting because it provides data scientists with an interactive environment for the implementation of data science workloads. Until that point, Oracle’s AI Platform Cloud Service doesn’t’t look very different from the current AI stacks from Google Cloud, Azure, Bluemix or AWS. From the purely technical standpoint, you can make the case that the AI Platform Cloud Service is relatively limited in terms of capabilities compared to the cloud incumbents. However, Oracle doesn’t bring some unique features to the table such as the new Adaptive Intelligent Apps that define AI-powered business processes in vertical areas such as finance, marketing, manufacturing, supply-chain, sales and several others. The Adaptive Intelligent Apps follow a similar strategy to what Salesforce has been doing with its Einstein platforms and definitely play to Oracle’s strengths in the industry-solution space.
3 Points About Oracle’s AI Strategy
When evaluating Oracle’s AI strategy from the market standpoint, there are several key points that deserve some attention.
1 — Visible Technical Limitations
Oracle AI Platform Cloud Service is certainly a welcomed addition to the market but is far from being a leader in the space. Compared to other AI cloud suites in platforms such as Google Cloud, Azure, AWS or Bluemix, the AI Platform Cloud Service has key technical limitations in areas such as cognitive APIs (ex: Watson Developer Cloud, Microsoft Cognitive Services…), a unique machine learning runtime( ex: Azure ML, Google Cloud ML….), support for a broader set of deep learning frameworks(ex: Theano, Torch, Caffe2, MxNet…) or data science tools(ex: Azure ML Experimentation Service).
2 — Unique Industry-AI Differentiator
The Oracle AI Platform Cloud Service Adaptive Intelligent Apps represent a unique differentiator compared to the AI offerings of Google, IBM, Microsoft and Amazon. Salesforce.com seems to be the company that can compete with Oracle in the SaaS-AI area but the AI Platform Cloud Services seems to be more flexible than the Salesforce Einstein stack in terms of the implementation of custom AI programs.
3 — Monetization Path Within Existing Customers
Similarly to IBM and Microsoft (and differently from Amazon and Google), Oracle can find a quick path to increase the adoption of the AI Platform Cloud Service within its existing customer based. From that perspective, making it easy for WebLogic and Oracle Apps customers to adopt the AI Platform Cloud Service will be a key element in the success of Oracle’s AI Strategy.