*

Nvidia's ( NVDA ) Rubin AI chip to include GPU, CPU and networking

chips

*

Nvidia ( NVDA ) faces delay with current flagship Blackwell chip

due to

design flaw

*

AI startups claim competitive chatbots need fewer Nvidia ( NVDA )

chips

By Stephen Nellis and Max A. Cherney

SAN JOSE, California, March 18 (Reuters) - Nvidia ( NVDA ) CEO

Jensen Huang is expected on Tuesday to reveal fresh details

about the company's newest artificial intelligence chip at its

annual software developer conference.

Nvidia ( NVDA ) stock has more than quadrupled in value over

the past three years as the company powered the rise of advanced

AI systems such as ChatGPT, Claude and many others.

Much of that success stemmed from the decade that the Santa

Clara, California-based company spent building software tools to

woo AI researchers and developers - but it was Nvidia's ( NVDA ) data

center chips, which sell for tens of thousands of dollars each,

that accounted for the bulk of its $130.5 billion in sales last

year.

Huang hinted last year the new flagship offering will be named

Rubin and consist of a family of chips - including a graphics

processing unit, a central processing unit and networking chips

- all designed to work in huge data centers that train AI

systems. Analysts expect the chips to go into production this

year and roll out in high volumes starting next year.

Nvidia ( NVDA ) is trying to establish a new pattern of introducing a

flagship chip every year, but has so far hit both internal and

external obstacles.

The company's current flagship chip, called Blackwell, is coming

to market slower than expected after a design flaw caused

manufacturing problems. The broader AI industry last year

grappled with delays in which the prior methods of feeding

expanding troves of data into ever-larger data centers full of

Nvidia ( NVDA ) chips had started to show diminishing returns.

Nvidia ( NVDA ) shares tumbled this year when Chinese startup DeepSeek

alleged it could produce a competitive AI chatbot with far less

computing power - and thus fewer Nvidia ( NVDA ) chips - than earlier

generations of the model. Huang has fired back that newer AI

models that spend more time thinking through their answers will

make Nvidia's ( NVDA ) chips even more important, because they are the

fastest at generating "tokens," the fundamental unit of AI

programs.

"When ChatGPT first came out, the token generation rate only

had to be about as fast as you can read," Huang told Reuters

last month. "However, the token generation rate now is how fast

the AI can read itself, because it's thinking to itself. And the

AI can think to itself a lot faster than you and I can read and

because it has to generate so many future possibilities before

it presents the right answer to you."