A Pure-Play in the Age of Virtualization
This month’s issue of The George Gilder Report features a pure-play that’s currently redefining digital storage.
This company is going to continue to lead the pack in customer loyalty, subscription renewals, and strategic innovation — benefiting more than most from the market’s impressive growth.
Below you’ll find an excerpt from the October Issue…
AI works by iteratively processing millions and billions of data points. It requires massive floods of data drawn from the real-world. Even though the first ventures into Artificial Intelligence (AI) reach back deep into the last century, this is the time that AI remained essentially a toy, or a research project, was useful mostly for playing chess or the Chinese game Go or proliferating PhDs.
It was only with the dawn of the zettabyte era that AI progressed from amusing, to actionable, to indispensable. And with that progression, “storage” as we knew it died forever.
Reigning and relevant now is “store-width.” That is, the measure not of how much stuff you can jam into the basement… but of how quickly you can find it, and move it to where data can be turned into information. It’s by finding patterns, that artificial intelligence turns data into information. For AI to see patterns, data must flow at massive speeds and scale. And using old-school “storage” systems for AI tasks would require great — and ultimately unsustainable — effort.
Revving Up Store-Width for Next-Gen Entrepreneurs
Worldwide accumulated data will exceed 50 zettabytes within a year or so. In case you are counting, a zettabyte is 1021 or sextillion bytes. The very size of these numbers tells you that this is not data for “storage.” Nobody needs to store that much data, if by store we mean put somewhere in case we need it someday.
Data on that scale is collected only because it can yield the information we need to know right now — and might need again tomorrow, or next week when the AI wants to seek a pattern that can be discerned only across millions of iterations.
Just a few years ago an MRI brain scan might assemble 2,000 images to yield a diagnostically useful result. Today, it’s more likely to assemble 20,000 and at a higher — more data-rich — resolution.
MRI image creation alone is driving an explosion in data collection in the health care sector. And the data can’t be stuffed into the basement. It’s got to be available at speed for the AI to help your doctor compare it to a billion other images and keep you alive.
It’s not that disk drives couldn’t feed an AI. They can and do today. But to use data entrepreneurially, everything about processing that data has to be fast enough, easy enough, and cheap enough to make the machines (and all their arduous maintenance) disappear from the consciousness of the innovators.
The all too physical reality of massive, energy-sucking, heat-generating, mechanically siloed, protocol sensitive, generation bound, legacy anchored, balky, and vulnerable machines must not be allowed to dictate what is possible. We want the exploitation of terabytes, petabytes, and exabytes of data to be not just possible, but irresistibly easy and cheap. We want innovators of all kinds to go ahead and run one more analysis, or a thousand more, at the drop of a hat, without a second thought.
Do you want the AI processing your brain scan to be just a little slower, a little less thorough, run 10% fewer iterations — simply because the queue has gotten a little too long, and processing is now a little too expensive?
Probably not. So the market for the right systems will grow explosively over the next few years, presenting savvy investors and innovative entrepreneurs with great opportunities. And that what brought me to this month’s portfolio choice…
Luckily, there’s still time to invest!
Look for a more in-depth discussion of this in the October monthly issue of The George Gilder Report.
Editor, Gilder’s Daily Prophecy