In response to growing rumors of Intel potentially being sold or split into separate businesses, Jim Keller, a legendary chip designer and former Intel executive, voiced his opinion on Twitter. Keller, now the CEO of AI chip startup Tenstorrent, argued that although Intel's profits may have decreased compared to its historic heights, selling or splitting the company would not unlock shareholder value. Instead, he believes it would be a "fire sale" of a company with vast potential for growth.

"You build value by having a great goal and a team that loves working towards that goal," Keller wrote. "Intel built the fastest CPUs on the best process technologies. Turning Intel's businesses over to third parties doesn't unlock shareholder value, it's a fire sale. This makes me sad."

Keller went further, stating, "I think a great Intel is worth $1 trillion. It seems a little careless to throw it away."

The remarks come amid reports suggesting that Broadcom might be interested in acquiring Intel's product division, while its foundry unit could potentially form a joint venture with TSMC or other major players, including Qualcomm, to create a new independent chipmaker. Keller, however, expressed skepticism about such moves, suggesting they would undermine Intel's potential for future growth.

A commenter on Keller's Twitter thread proposed that Intel could instead become a privately held company with the help of American investors, restructure its leadership, and focus on revitalizing the company. Keller responded, acknowledging that it would be challenging but doable, stating, "Humans are amazing when they have a great goal and a team they believe in."

Intel's current situation is complicated by the potential loss of its manufacturing edge. If the company were to abandon its own fabrication facilities (a hallmark of its integrated device manufacturing model), it would no longer control the manufacturing of its products. This would likely diminish its competitive advantage, as it would lose direct control over production and innovation.

Additionally, there is a broader geopolitical angle to consider. While the U.S. government has supported domestic manufacturing, especially under the Trump administration, it's unlikely to back Intel's fabs being controlled by foreign companies like TSMC. Moreover, Intel's foundry unit, which has yet to secure significant customer orders, is unlikely to attract major investments without substantial support.

Keller's concerns are not without merit. Companies like Broadcom, with substantial financial resources, could benefit from acquiring Intel's product division to strengthen their CPU capabilities and product design teams. However, such a move would trigger the automatic termination of Intel's extensive cross-licensing agreement with AMD, which could have serious consequences for the new entity's access to shared innovations unless a new agreement is negotiated.

In the end, Keller believes that Intel still holds significant untapped potential and urges stakeholders to consider its long-term future rather than selling off parts of the company.