Intel Extends Further Into AI Silicon Market with Habana Labs Acquisition

Intel Corp. is buying Habana Labs for approximately $2 billion.

Habana Labs is an Israel-based developer of programmable deep learning accelerators for the data center. Its Gaudi AI Training Processor, for instance, claims to deliver up to four times the throughput of systems built with the equivalent number of GPUs. That kind of compute boost is critical for training AI models, which use large amounts of data. Further, this kind of performance and flexibility is necessary as enterprise IT architecture extends from the cloud to the data center to the edge.

Intel’s Nevin Shenoy

“Habana turbocharges our AI offerings,” said Navin Shenoy, executive vice president and general manager of the Data Platforms Group at Intel.

Experts note that Intel’s acquisition gives enterprises flexibility in their cloud-to-edge strategies.

“In addition to using specialized AI hardware in the cloud,” said Dan Sullivan, principal architect at New Relic, a software analytics company, “with Intel’s acquisition of Habana Labs, enterprises now have more efficient options for training increasingly complex deep learning models on premises.”

The purchase also fast-tracks Intel’s forays in the AI silicon market, which is expected to grow at a clip. According to “The Artificial Intelligence Market” from Allied Market Research, the global AI chip market was valued at $6.64 billion in 2018. Intel estimates that it will be worth $25 billion by 2024. In 2019, Intel expects to generate more than $3.5 billion in AI-driven revenue.

The Habana acquisition complements other moves by Intel to secure its position in AI. Last month, Intel announced its first dedicated chips for powering artificial intelligence (AI) workloads in the cloud. Having AI chips for the cloud ensures better performance rather than having to rely on general compute.

Habana will remain an independent business unit and will continue to be led by its current management team. Habana will report to Intel’s Data Platforms Group.