Execs from Nvidia and Equinix on Scaling AI

Why Many Digital Transformation Efforts Fail

If you have been around information technology investments long enough, you know that outcomes often are negative: the hoped-for benefits do not emerge at all, or prove to be less than expected. Possibly 30 percent of major IT investments actually produce the expected results. 

The same continues to be true of digital transformation efforts and investments. A new survey of IT professionals suggests DX has the same outcomes. 

source: Walkme 

The study also had respondents estimating that, of all IT projects, fully 41 percent failed to meet objectives, in the form of key performance indicators. And the larger the organization, the more likely it is that DX or any other IT initiative will fail. Larger enterprise respondents reported failure rates at six times the rate of small entities. 

source: Walkme 

To be sure, the more-important KPIs deal with financial performance: productivity, equity value growth, higher income or lower costs. 

Such “failures” are not hard to understand. Most large enterprises rely significantly on third-party integrators to attempt complex initiatives. That means lots of organizational complexity. And the number of persons who must not only agree to support, but actively ande usefully support such initiatives is therefore quite high. 

source: Walkme 

And that runs squarely up against an immutable law: the more permissions one requires to get anything done, the lower the chances of success. Consider any project with a number of key influencers. Assume that at each stage, the chances of getting a “yes” are 50 percent.

If only one “yes” is required, odds of success are 50 percent. If two “yes” hurdles are required, success rates drop to 25 percent. If three “yes” hurdles exist, odds drop to 13 percent. Each successive hurdle reduces success rates further. 

As you can see, even a small number of hurdles, each with a 50-50 chance of success, quickly reduces odds of success to an impossible level. After seven hurdles, odds of success are one in 100. 

Obviously, any complex IT project has many more gates than seven. With sufficiently vigorous top-level support, the odds of success at each gate are likely higher than 50 percent. But there are so many more hurdles or gates that odds of success are low; odds of failure high. 

The conventional wisdom is that about 30 percent of big IT projects actually succeed as expected.

Sustainable Competitive Advantage from Advanced Technology?

Perhaps every firm dreams of gaining sustainable advantage over its key competitors. Perhaps no firm ever achieves it. Gartner now seemingly argues that digital transformation will wind up the same way: becoming a way firms compete, but only one way among many, and without permanent leadership outcomes. 

Competitive advantage from “digital transformation” will be as transitory as all earlier applications of computing and information technology, one might conclude from the Gartner assessment. 

As digital networks, “always on” connectivity, and smart devices become ubiquitous and commonplace, computing will simply fade into the background, becoming as unobtrusive as the dumb thermostat or light switch, says Ed Gung, Gartner Research Board managing VP.

As digital technology becomes embedded into the way all parts of the business operate, “digital” will cease to be a useful modifier, he also notes. In the end, digital technology will become just one more dimension on which companies compete. 

Distribution networks, capital assets, exploration rights, customer relationships and content are other levers companies can pull. But sustainable advantage will be difficult to maintain in those realms as well.

Hyperscalers to Lift Data Center Capex to $377 Billion in 2026, According to Dell'Oro Group

Will #EdgeComputing capex be 8% of data center capex by 2026? http://dlvr.it/SWKNPN
http://dlvr.it/SWKNPN

What Does 1,000X Compute Architecture Look Like?

Some believe the next-generation internet could require a three-order-of-magnitude (1,000 times) increase in computing power, to support lots of artificial intelligence, 3D rendering, metaverse and distributed applications. 

What that will entail depends on how fast the new infrastructure has to be built. If we are able to upgrade infrastructure roughly on the past timetable, we would expect to see a 1,000-fold improvement in computation support perhaps every couple of decades. 

That assumes we have pulled a number of levers beyond expected advances in processor power, processor architectures and declines in cost per unit of cycle. Network architectures and appliances also have to change. Quite often, so do applications and end user demand. 

The mobile business, for example, has taken about three decades to achieve 1,000 times change in data speeds, for example. We can assume raw compute changes faster, but even then, based strictly on Moore’s Law rates of improvement in computing power alone, it might still require two decades to achieve a 1,000 times change. 

source: Springer 

And that all assumes underlying demand driving the pace of innovation. 

For digital infrastructure, a 1,000-fold increase in supplied computing capability might well require any number of changes. Chip density probably has to change in different ways. More use of application-specific processors seems likely. 

A revamping of cloud computing architecture towards the edge, to minimize latency, is almost certainly required. 

Rack density likely must change as well, as it is hard to envision a 1,000-fold increase in rack real estate over the next couple of decades. Nor does it seem likely that cooling and power requirements can simply scale linearly by 1,000 times. 

Persistent 3D virtual worlds would seem to be the driver for such demand.  

Low-latency apps such as persistent environments also should increase pressure to prioritize traffic, move computing closer to the actual end user location and possibly lead to new forms of content handling and computation to support such content. 

Compared to today, where content delivery networks operate to reduce latency, content computation networks would also be necessary to do all the local and fast processing to support immersive 3D experiences that also are persistent. 

How we supply enough fast compute to handle rendering, for example, could be a combination fo device and edge computing architecture. 

Among the other issues are whether chip capabilities can scale fast enough to support such levels of compute intensity. 

So long as we have enough levers to pull, a 1,000-fold increase in computing availability within two or three decades is possible. Moore's Law suggests it is possible, assuming we can keep up the rate of change in a variety of ways, even if, at the physical level, Moore’s Law ceases to operate.  

But that also means fully-immersive internet experiences, used by everybody, all the time, also would be accompanied by business models to match. 

So in practical terms, perhaps some users and supported experiences will use 1,000 times more computational support. But it is unlikely that the full internet will have evolved to do so.

Metaverse for Teaching

Meta Immersive Learning project is creating 10 “metaversities,"  digital campuses affiliated with real universities in the United States. In the last academic year, nine classes were offered to about 400 students in the metaverse. They ranged from journalism to biology, English to history. Next falll, the metaverse course offerings will increase to 15.

For Most of Us, AI is a Feature of a Product

For suppliers of artificial intelligence infrastructure, AI is a direct revenue source. For most of the rest of us, AI is a feature of some other product we buy or use. Amazon, for example, uses AI to drive its $222 billion e-commerce business plus its $103 billion third-party merchant revenues.

AI also underpins Amazon Web Services, subscriptions and advertising. 

source: Mozilla