Intel is investing heavily into deep learning with the acquisition of a startup called Nervana Systems, as it bids to play catch-up in the accelerating artificial intelligence space.

Nervana, founded in 2014, operates a hosted platform for deep learning that’s geared towards solving “machine-learning problems,” the company says. The platform is optimized “from algorithms down to silicon,” and has been used for a wide range of tasks, including analyzing seismic data, studying plant genomes to find new hybrids and even searching for places to drill for oil.

The most common use case however, is businesses looking to use Nervana’s cloud service to build and deploy applications that make use of deep learning, which is a branch of AI that’s used to uncover patterns in large pools of data, and perform voice and image recognition.

More importantly, perhaps, Nervana is also said to be building a specialized processor, called an ASIC, that’s customized for deep learning.

Intel didn’t disclose terms of the deal, but Re/Code reported that the figure is likely north of $400 million.

“We will apply Nervana’s software expertise to further optimize the Intel Math Kernel Library and its integration into industry standard frameworks,” said Diane Bryant, head of Intel’s Data Center Group, in a blog post.

In addition, Bryant said Nervana’s expertise will help “advance Intel’s AI portfolio and enhance the deep-learning performance and TCO of our Intel Xeon and Intel Xeon Phi processors.”

Intel adds AI “inside” its chips

Computing has only just transitioned from mainframes and PCs to servers and cloud-based data center, but already it’s on the verge of a second transition, with the emergence of artificial intelligence technologies capable of injecting smarts into almost any object.

Deep learning is at the core of this transition, responsible for new technological wonders such as self-driving cars and instant language translation on our phones. And now, chip makers like Intel are set to accelerate this transition by embedding deep learning software directly into the microprocessors that power everything from smartphones and connected cars to drones and much, much more.

Intel, which has seen its PC market cash cow slowly but surely erode in the last few years, is desperate to be among the leaders of this brave new world. To date, most of its efforts in the AI space have focused on adapting its latest generation Xeon chips for use in AI and machine learning. Just last year it released a special version of that chip, called Xeon Phi, which came with more cores for parallel processing.

With Nervana on board, Intel gains expertise on the software side that it can put to use right away. It also gains access to Nervana’s designs for specialized deep learning chips, so we can expect to see the development of these accelerate rapidly.

Naveen Rao, who helped develop neural computer networks at Qualcomm Inc. before founding Nervana, said the company’s chip designs made it a natural fit for Intel.

“Nervana’s AI expertise combined with Intel’s capabilities and huge market reach will allow us to realize our vision and create something truly special,” Rao said in a blog post announcing the acquisition.

One of the chip’s key capabilities is said to be a much higher rate of data flow when compared with regular chips in use today. Rao told Fortune that the new chips are designed to mimic the parallel operation of neurons and synapses in the human brain in order to read and write data from storage, and share that data with other chips extremely rapidly.

Nervana’s chips will be able to transfer data at speeds of 2.4 terabytes a second, with an ultra low latency of around five to ten times faster than current chips, Rao said.

Image credit: geralt via pixabay