Nvidia (Nvidia Co., NVDA) are accelerating the expansion of business scope from computer graphics into planning in the field of artificial intelligence, to this end, the company introduced a different processor, and a configuration of the processor and the computer to solve the problem of science at high speed.

The company said on Tuesday, new Tesla P100 chip designed for enterprise data center, the chip has a very high performance, integrated with 15 billion transistors, the number is about nvidia previous generation of high-end graphics processors and Intel (Intel Co., INTC) part of the new server chip twice last week.

Nvidia CEO jen-hsun huang at the company's annual technology conference, said during a speech it is by far the largest chip.

He predicts that the original cloud computing services to buy the chip, chamber of commerce to next year will start configuration sold in other companies in the server. Meanwhile, nvidia also plans to launch its carrying eight Tesla P100 chip computer and artificial intelligence application software, at a price of $129000.

Huang said, this new computer can quickly handle called DGX - 1 artificial intelligence task, speed and 250 by the British thermal production by your company general chip driver server, cost much less. A standard server, he says, need 150 hours to complete the task, DGX - 1 only two hours.

Headquartered in California, Clara, nvidia has for years been trying to for graphics processor (GPU) chip electronic games outside of the user. The chip integrates hundreds of simple processor, has been used in many large super computer science problem.

Huang has recently been focused on artificial intelligence, especially a technique known as machine learning, the technology of image recognition and voice language is very helpful. In face recognition, the traditional image recognition for programmers define facial features, and machine learning software can make the computer by browsing a huge amount of so-called training pictures to learn to recognize faces.