Machine learning technologies and techniques are giving organizations powerful new ways to utilize the vast amounts of data they’re collecting. According to several reports, ML spending is increasing at a compound annual growth rate (CAGR) of around 25%. That’s benefitting vendors providing ML solutions, which appears to be mostly cloud vendors outside of the HPC segment.
According to Zion Market Research’s July report, the global market for ML was valued at $1.6 billion in 2017 and is expected to account for $20.8 billion in spending by 2024, which translates into a rather healthy 44% compound annual growth rate (CAGR). That was the outlier in a recent roundup of ML market reports.
Market Reports World came up with a similar number in its global tally on ML spending. The company said that $2.5 billion was spent in 2017, and with a 22% CAGR, global spending on ML should reach $12.3 billion by 2026. In its 2018 report, BCC Research concluded that global spending on ML reached $1.4 billion in 2017. And with a CAGR of 44%, it would hit $8.8 billion by 2022.
Intersect360 Research sliced the figures a slightly different way, and looked at the global spending on infrastructure in support for ML and AI for its recently released report, “Worldwide AI and Machine Learning Training Market Model: 2018 Spending and Future Outlook.” The Silicon Valley firm concluded that AI and ML infrastructure spending grew 50% over the past two years and would soon surpass $10 billion in annual spending.
“Machine learning has been in a very high growth stage,” Intersect360 Research CEO Addison Snell said in a press release. “In addition to that $10 billion, many systems not one hundred percent dedicated to machine learning are serving training needs as part of their total workloads, increasing the influence that machine learning has on spending and configuration.”
These figures may sound off to anyone who has heard the hype about AI generating trillions in value in the near future. We even have IDC, one of the biggest and most respected analyst firm, predicting spending on cognitive and AI systems will reach $77 billion by 2022 and drive a 37% CAGR through 2022. So what gives?
The problem, as Snell explains, is that AI does not constitute a “market” in the traditional sense. That makes it hard to assess the size of the market and how quickly it’s growing. It should also serve as a word of caution to vendors who think they can get rich quick by selling AI or ML solutions.
“While many organizations are investing resources in the training of machine learning models, in most cases these efforts overlap with IT initiatives already in place,” Snell writes in an October blog post. “The training of AI models might involve systems, storage, or networks that were already in place or already budgeted.”
AI and ML workloads are growing quickly in the high performance computing (HPC) segment, and in some cases are displacing modeling and simulation workloads running on large supercomputers. While some groups are buying new systems dedicated to AI, there’s a lot of overlap with other workloads.
In the non-HPC segment, most of the ML and AI investment is occurring in the cloud, Snell says. “Our research suggests that as of 2018, this work—where it exists—is done predominantly through cloud resources…” he writes.
With that in mind, one can comfortably conclude that the ML “market” will double in about three years and will be driving tens of billions of dollars in new spending in just a few years. That’s not a bad place to be.
Sorting AI Hype from Reality
Gartner Sees AI Democratized in Latest ‘Hype Cycle’
Credit: Google News