"Having worked in the technology industry my whole life, I feel a bit saddened to see the current situation of intel," said Arm CEO Rene Haas in a recent interview with the USA tech media The Verge.

$Intel (INTC.US)$ Once a long-time "archenemy" of. $Arm Holdings (ARM.US)$ However, Arm currently has a market cap close to 150 billion dollars, far surpassing the troubled intel.

Haas noted that intel used to be an innovation giant. But the technology industry must constantly innovate. "Many great technology companies have disappeared because they failed to reinvent themselves."

Earlier this week, intel CEO Pat Gelsinger retired, allowing intel to reconsider some restructuring plans and trade options, including the spin-off of the chip manufacturing business previously rejected by Gelsinger.

Haas stated that the biggest challenge facing intel is how to integrate the integrated device manufacturing (IDM, where chip design, manufacturing, packaging, testing, and sales are all handled in-house) with the fabless model (only responsible for chip design, outsourcing the rest). "Over the past decade, intel has been indecisive."

"Pat Gelsinger's strategy is very clear, which is that IDM is the winning path. He began to implement this strategy in 2021, but in my opinion, three years is far from enough; this should be a long-term strategy of five to ten years. After Gelsinger's departure, intel's new CEO must make a choice."

Haas believes that IDM is a very good strategy. "If they can execute the IDM model well, I believe they will have a huge advantage. But the cost of IDM is very high, which could become a major challenge."

It was also reported that before Gelsinger's retirement, Haas had approached intel about wanting to buy a large share of the company. There were also rumors that Arm intended to expand into self-manufacturing chips, rather than just licensing designs.

However, Haas did not comment on the rumors. "If you are an IDM company and your strategic advantage lies in having products and fabs, then in terms of cost, you will have a huge potential advantage over competitors," he said.

While Gelsinger was CEO of intel, Haas repeatedly tried to persuade him to consider a partnership with Arm, "because if you have your own fabs, the key to those fabs is output, and we can provide that output. But I was not successful in convincing him to do so."

Regarding rumors that Arm will independently develop ai chips, Haas stated that designing computer architecture and future computing technology requires special attention to the relationship between hardware and software, deeply understanding where trade-offs need to be made to ultimately benefit consumers from the chips. Developing a specific chip offers a better perspective for weighing designs compared to licensing intellectual property and can come closer to achieving the goal. "Therefore, if we are to do something, that will be one of the reasons."

Regarding Sam Altman's prediction that general ai (AGI) will be realized by 2025, Haas said, "I know he has his own definition and reasoning for AGI. I don't really care about the distinction between AGI and ASI (super ai)."

What I am more concerned about is when these AI agents will start to think, reason, and create. It's like a moment of 'crossing the Rubicon.' If you had asked me this question a year ago, I would have said it's still a long way off. But now, if you ask me, I would say we are closer to that moment.

Editor/rice