As more people use their smartphones for tasks like watching movies, and as more devices like wind turbines are connected to the Internet, the behind-the-scenes-infrastructure required to funnel all that data is struggling to keep up with the demand.
That’s partly why telecommunications giant AT&T (T) has been exploring cutting-edge technologies to improve its giant data center network and help it stream data to customers without any hiccups.
On Wednesday at Intel’s (INTC) annual developer conference, Intel and AT&T said they were deepening an existing partnership that involves AT&T getting early access to new Intel technologies and jointly working on networking infrastructure problems specific to the carrier’s needs.
The two companies have already created a technology team of 50 engineers that is attempting to fine-tune Intel’s chip technology to work more efficiently with AT&T’s custom networking software that it plans to make publicly available for free.
Diane Bryant, executive vice president of Intel’s data center business, said Intel is modeling its AT&T partnership on current relationships it already has with its so-called Super 7 cloud-computing customers. These giants include Amazon (AMZN), Microsoft (MSFT), Google (GOOG), and Facebook (FB), along with Chinese tech titans Alibaba, Baidu, and Tencent.
These companies are considered leaders when it comes to designing massive cutting-edge data centers and experimenting with new technology, Bryant explained.
“If you take the servers and storage away from Google and Amazon, there is no Google or Amazon,” she said about the importance of data center infrastructure.
In these partnerships, Intel provides chips and designs to the companies before they are publicly available, and often builds specialized equipment for each business. For example, in 2014 Intel built a custom chip for Amazon that the retail giant uses to help power its AWS cloud-computing business.
Get Data Sheet, Fortune’s technology newsletter.
Over 50% of the chips used by the cloud-computing giants are custom built, Bryant said, and these companies are constantly experimenting with new technologies so that everything in their data centers is always top notch.
AT&T, however, is a much older company than these newer more web-oriented cloud businesses that are part of the Super 7. And it has decades of older technology operating behind the scenes. Therefore, its needs and requirements are different from the younger companies.
“With AT&T, you are in the middle of a massive transformation,” Bryant said.
AT&T chief strategy officer John Donovan said that AT&T is retrofitting a lot of its older data center technology that’s “very rigid and hard.” He said that his staff found a 42-year-old software bug in its system, which highlights the elderly technology the company must deal with.
“All those old switches, they leak green gloop,” Donovan joked. “They kind of creek.”
At the heart of AT&T’s data center upgrade is a reliance on two types of newer networking technologies called “software-defined networking” and “network functions virtualization.” These wonky names generally refer to the idea of using generic servers, or so-called white boxes, that are bundled with custom-built and open-source networking software.
For more about Intel, watch:
In recent years, however, these networking companies have begun to shift to a more software-oriented line of products and are restructuring their own businesses to accommodate the needs of business customers that want more flexibility.
Donovan said one reason AT&T built its own networking software was because proprietary networking gear often contained numerous features AT&T didn’t need. However, like AT&T CEO Randall Stephenson told Fortune last year, his company will still work with Cisco on some of its data center technology.
Bryant said that it’s likely that Intel will forge similar partnerships with other carrier giants as well.