Enterprises are still struggling with the business case for generative AI projects more than a year after the craze started, and we may have to wait until the end of 2025 to see if they're seen through to completion.
Datacenter operators are in a key position when it comes to AI because their facilities are the physical homes for much of the infrastructure used for AI training. Global datacenter and colocation provider Equinix says many customers are still trying to figure out if there is a commercial need for GenAI projects.
"From my conversations with our customers, many of them are working through the business case, the actual business value that taking this proof of concept into production will release for them," Equinix president and CEO Adaire Fox-Martin said at Citi's recent 2024 Global Technology Media and Telecom Conference.
"And they're not always easy business cases to resolve. My take is that it will be a 12 to 18-month journey before we see the significant impact of that," the former Google Cloud EMEA exec added.
That time frame chimes with another recent prediction from analyst house Gartner that at least 30 percent of GenAI projects are set to be abandoned after proof of concept by the end of 2025.
The reasons for this are various; poor data quality, inadequate risk controls, escalating costs or unclear business value, the consultant said.
Gartner warned that while many organizations are looking to GenAI to transform their business models and create new opportunities, getting into the game requires substantial investment, with upfront costs alone ranging from $5 million to $20 million.
The same company warned this week: "It is really easy to waste money on generative AI," noting that "500 to 1,000 percent errors of AI cost estimates are possible."
Yet there is money to be made by tech giants while businesses are still afraid of being left behind in the AI stakes, as Equinix reports.
"Our partnership with Nvidia is playing out beautifully in the context of the enterprise space," Fox-Martin told the gathering of investors at the Citi event.
"The opportunity with Equinix and Nvidia is that Equinix takes all that complexity, those three pieces (GPUs, software, and networking), it takes the implementation and the systemization of those and implements them on behalf of a customer so that a customer can simply execute," she added.
- Microsoft exec warns of business functions being sacrificed on the altar of AI
- OpenAI co-founder's Safe Superintelligence startup inhales $1B in funding
- If every PC is going to be an AI PC, they better be as good at all the things trad PCs can do
- Fintech outfit Klarna swaps humans for AI by not replacing departing workers
"A customer can choose to do that either on a capex or on an opex basis, but we take the pain away of making that available. We have a number of customers up and running already on that environment because, essentially, the Equinix-Nvidia partnership is providing the opportunity to have private AI, private learning, private LLMs operating in our environment," Fox-Martin explained.
However the signals about AI, and GenAI in particular, can seem confusing. In June, a study found that businesses have become more cautious about investing in AI due to various concerns, including that "the financial benefits of implemented projects have been dismal."
In contrast, a report commissioned by Google – a company with a vested interest in seeing AI succeed – last month found that early adopters of GenAI were already seeing boosts to revenue of more than six percent.
More recently, enterprise software provider ServiceNow tried to warn clients that they should consider GenAI as a long-term investment, and that they shouldn't expect to get huge returns tomorrow or next week.
Whether organizations are prepared to keep pumping cash into AI projects is a different matter, especially when many are experiencing constrained IT budgets and looking for cost savings. There have also been warnings this week and last month that AI could be diverting badly needed investment from other IT functions.
How many projects are set to be abandoned after proof of concept by the end of next year? The million dollar question. Why not ask ChatGPT. ®