The summer of AI is almost over and the bill is coming due.

That’s the thread running through the stories below, curated from Channel Dive and its Informa TechTarget sister publications between February and August. Together, these stories describe a year when enterprises stopped asking if AI works and began fretting about what it costs to keep it going. The bill arrived in dollars and megawatts, among other things.

In an August story, CIO Dive’s Paige Gross described how AI strategy shifted from the “tokenmaxxing” earlier this year to midsummer malaise. “Before CFOs could even get it on their radar, the bills, the damage had been done,” Nicholas Merizzi, principal at Deloitte Consulting, said. Gartner Senior Directory Analyst Will Sommer said, “The models produced by leading AI labs are getting more token-hungry faster than they are getting cheaper,”.

The stories below move through seven stages of reckoning. The first three describe the gap between what’s been spent and the value that hasn’t been harnessed yet. The next three stories attempt to price that value, from Flexera’s read on AI sprawl to the Linux Foundation’s push to standardize some AI costs. 

We also looked at the governance problem, where fewer than half of CISOs believe their bosses see AI security as a business enabler, and the physical limits of AI, from electrical power and construction to the political power small cities wield in big states. 

The roundup ends with stories about people — at trust and expertise gaps in the workforce, who is being paid to close those gaps and how investments in forward-deployed engineers change everything.

Every enterprise will handle AI differently and those variations can add up to a tidy business for channel partners. What we know is that slowing down doesn’t seem like an option. 

“With the pace of change as rapid as it is, FDEs are going to be critical in converting opportunities to revenue,” said Peter Bryant, GSI practice lead at Omdia, a Channel Dive sister company. “Enterprises don’t really have the patience at the moment to wait for the partners to get trained up, because by the time they do, the newest model comes out.”