Upcoming Cloud Compute Regulations — Winners and Losers

To recap, the U.S. CLOUD Act opens up U.S. warrants to data stored anywhere in the world (in past years, it was common to use cloud “endpoints” in specific countries not just for data locality but also for regulatory protection.) The European Union followed by introducing a similar bill applying to the entire 28 member bloc. Further, bilateral agreements might open up data stored in the U.S. to foreign warrants (possibly authoritarian governments.)

Data volumes have grown from gigabytes to terabytes and now to petabytes, exabyte, zettabytes, and yottabytes. Growing private cloud capital infrastructure to keep up with data demands remains possible for only a few sophisticated players (Amazon, Google, Facebook, Uber, Baidu, Tencent, Alibaba, and some major banks and hedge funds.) For most others, the obvious solution has been to grow elastically via public cloud infrastructure. A recent walk around Washington D.C.’s “Metro Center” station stop revealed no less than a dozen cloud computing advertisements — it is clear that even more conservative players like Government are embracing the public cloud. 

One of several Wall Street Journal articles, “EU Calls for Access to Foreign Server” measured a paltry fifty five words. Yet the impacts are enormous and encumber hyper-growth prospects for the four major cloud providers (Amazon Web Services, Microsoft Azure, IBM SoftLayer, and Google Cloud) — all of which are American and would fall under the regulations. This is significant.

Analysts see the market segment approaching a half trillion dollars in the next five years. Unlike hardware and other industries, there are no strong Asian challengers (or anywhere else for that matter.) This is a market with U.S. dominance so vast it would make Zbigniew Brzezinski jealous, as the grand chessboard of strategic imperatives is seen through an orthogonal dimension of cloud supremacy — the bridges and canals of our generation.

There are three potential ways regulations might shift the landscape:

1. Challengers in favorable regulatory geographies grab market share and fracture U.S. firms’ current capture. This might be China (for their local market) or a host of newer players catering to specific regions (e.g., the Gulf nations.)
2. Free-for-all storage of marginally-usable data (often kept for future value in feeding ML/AI models) might soften. To be sure, growth will continue to be orders of magnitude, just not as high as it could be. This would be a pity since the real value in data growth is all the models which could be trained with an ever-growing breadth of features and volume of records.
3. An increasing move to private clouds, and especially, an increasing move away from structured data storage such as Google BigQuery, towards in-house solutions built atop vendor packages. 

Route three is the most interesting because it introduces as many winners as losers. This would shift investment from public cloud providers towards those offering private cloud infrastructure and software. One can imagine a move away from AWS S3 towards the likes of Ceph. Away from AWS EC2 towards commodity racks abstracted away with Kubernetes or dense compute appliances such as NVIDIA’s DGX. Away from BigQuery/AzureSQL/AirTable/AmazonRDS and Managed SaaS solutions and towards in-house Oracle Exadata appliances or a Kinetica database running atop NVIDIA DGX.

Ultimately, customers might be the biggest winners, as larger private clouds encourage more liberal storage of private data, ever-increasing enterprise feature catalogs, and more valuable AI/ML models being trained.

Given double-digit annual hyper-growth in cloud computing, the reality is continued growth no matter what happens. We just won’t know what it could have been. One indicator may be enterprise value shifts towards the private cloud sub-sector (both hardware and software). The next eighteen months will be revealing.

Disclosures:

• I’m an outspoken fan and shareholder of NVIDIA (NASDAQ: NVDA). I’m also a shareholder in cloud giants Amazon Web Services (NASDAQ: AMZN), Google Cloud (NASDAQ: GOOG), and Baidu (NASDAQ: BIDU).
• I’m an AI practitioner and desperately want the industry to inexpensively stockpile data of even marginal current use, as it can help train the models of tomorrow.
• I’m so convinced industry will move away from Google BigQuery towards in-house Oracle Exadata and Kinetica that I’ve bet my career on it — I am the Machine Learning Product Owner here at Kinetica.
• Most importantly, I’m a proud American and selfishly hope for our continued market preeminence.