AI Costs Enter the Conversation
Some Turn to China for Savings, but what About Security?
Over the past decade enterprise CIOs have been in a race to adopt public cloud as quickly as possible. Board and investor pressure to not fall behind a game-changing technology led the charge. But for a large portion of the business marketplace, spending freely on public cloud resulted in substantial technology cost overruns that went on, in some cases, for years.
Are early adopters again taking the financial arrows on the latest new technology?
While CIOs race to adopt artificial intelligence, their budget-minded C-suite colleagues are concerned. Like public cloud, AI is typically structured with a consumption-based pricing model, which brings back bad memories for some who have suffered financial overruns in the public cloud. Only 26% of CFOs say that their companies have a comprehensive view of their AI costs according to a KPMG survey of financial executives.
As a result, users may be pivoting towards cost control more quickly with AI than they did with public cloud. Microsoft is scaling back on their investment in OpenAI and Anthropic’s more advanced models in favor of the company’s homegrown offerings. Microsoft AI Chief Executive Mustafa Suleyman told Bloomberg, “Anthropic is extremely expensive and I think many people are urgently looking for alternatives.”
At a June AI event, Sam Altman said something quite honest given the Open AI CEO’s role in the growth of AI.
“People are really saying, you know, it’s kind of a meme now, but ‘My company spent my entire 2026 budget in Q1. Can you make this more efficient? We are continuing to push on that more with models. I think we’ll have a lot of ways we can help people get more value for less spend. But that went from, at the beginning of this year, an issue that never came up (people were totally happy with the amount they were spending) to, all of a sudden, a huge issue.”
Kudos to Altman for addressing the issue with such frankness. Nobody in a similar role with a major public cloud provider addressed runaway cloud costs in the early innings of cloud adoption with that level of honesty.
In an effort to mitigate the sticker shock of their AI consumption, some companies are turning to cheaper Chinese AI models to meet their needs more economically. While China has earned a raised eyebrow when it comes to security, some find US-based AI providers’ focus on that issue to be a ploy to protect their market share and have chosen to experiment with the Chinese models, while others are finding their way to more affordable US companies like Thinking Machines, a well-backed startup American AI firm founded by Mira Murati, the former OpenAI chief technology officer. CIOs are trying to find the best solution, balancing security, features, and cost, but there’s no clear winner.
The Benefit of Predictable Compute Cost: a healthy dose of AI is definitely in our computing future, but the present is messy. As we saw with cloud, rabid early adopters with top-level tech staffs seem to be willing to pay a premium to gain first mover advantage. It may behoove the typical enterprise to continue to evaluate the playing field and wait until there is more proof of concept for massive AI adoption. Meanwhile we’re hearing success stories from our customers who are focusing on individual AI projects that are less ambitious but more affordable and achievable.
With AI in your future, you may be looking for budget certainty in other areas of your IT. If so, a form of hybrid colocation may be worth considering. Direct LTx offers a dozen carriers, direct access to multiple clouds, and meaningful experience with high-power-density computing. For a discussion on reducing the variance in your IT infrastructure costs contact us at strategy@DirectLTx.com.