Intel CEO Lip Bu Tan Surpasses Wall Street Expectations on 1-Year Mark: We’re Embracing Our ‘Paranoid’ Origins

(SeaPRwire) –   In recent years, Intel has been focused on a transformation to demonstrate its continued significance in an AI-driven market largely led by Nvidia’s semiconductors.

However, on Thursday, as the company significantly surpassed Wall Street’s financial expectations, it conveyed a different message: there is strength in its legacy as a 58-year-old manufacturer of PC and server microprocessors.

“We are returning to our core identity as a data-driven, paranoid, and engineering-focused organization,” stated CEO Lip Bu Tan during the opening of the company’s Q1 earnings call, alluding to the well-known “only the paranoid survive” mantra of Intel’s late co-founder, Andy Grove.

Intel’s stock jumped over 22% in after-hours trading on Thursday following the release of its first-quarter earnings. Contrary to analyst projections of a 2% revenue decline for the quarter, Intel reported a 7% year-over-year revenue increase to $13.6 billion. The company also forecast current-quarter revenue to be between $13.8 billion and $14.8 billion, substantially higher than the $13.06 billion analysts had anticipated.

The company reported a surge in demand for its central processing unit (CPU) chips, which utilize its long-established x86 architecture. Intel noted that its revenue could have been even greater if it had the manufacturing capacity to produce more chips.

“Twelve months ago, the discussion about Intel centered on our potential for survival,” Tan remarked. “The focus today is on the speed at which we can expand our manufacturing capabilities and increase our supply to satisfy the tremendous demand for our products.”

This statement was not an overstatement regarding the company’s dire situation when Tan assumed the CEO role in March 2025, shortly after Pat Gelsinger was removed. During that period, many industry watchers and former board members questioned if Intel should be dismantled, with its manufacturing plants sold off or established as an independent entity. A few months into Tan’s tenure, the U.S. government acquired a 10% stake in Intel, a move the Trump administration stated was crucial for national security and U.S. industrial policy, providing the company with vital support.

CPUs are back, but is Intel?

The renewed demand for Intel’s CPUs marks a somewhat unexpected shift after a period where Nvidia’s graphics processing units (GPUs) seemed destined to lead the future due to their superior performance with AI models.

“We have observed unmistakable evidence in recent months that the CPU is reclaiming its position as the essential cornerstone of the AI era,” Tan stated on the call. He clarified that CPUs are more effective for operating AI services, known as inference, whereas GPUs maintain an advantage in the creation, or training, of AI models. In the initial phase of the generative AI surge, when firms like OpenAI, Anthropic, and Google were developing massive new models, GPUs were the dominant choice. Intel contends that as the market matures, the momentum is shifting back toward CPUs.

Intel’s CFO, Dave Zinsner, indicated that the GPU-to-CPU ratio within AI data centers is in flux. For training AI models, the ratio is typically seven or eight GPUs for each CPU. However, for inference, the ratio narrows to three or four GPUs per CPU. Zinsner added that as agentic AI becomes more prevalent, this ratio could reach parity or potentially become more favorable for Intel.

Despite this, significant hurdles remain. Nvidia has recently entered the market with its first independent CPU, intensifying the competition Intel already faces from arch-rival AMD and server chips using the ARM architecture, including an upcoming chip ARM is developing itself.

A more profound question is whether Intel’s recovery indicates a genuine corporate turnaround or is merely a byproduct of the expansive build-out of AI infrastructure, with data center operators aggressively purchasing available chips. Major uncertainties also persist regarding Intel’s foundry business, which produces chips for other companies and rivals industry leader TSMC. A key concern is whether Intel will maintain the enormous investments needed to advance to the next generation of chip manufacturing technology.

Tan has previously stated that Intel will not proceed with constructing factories for its most advanced 14A fabrication process, which can make chips with 1.4-nanometer circuits, without guaranteed customers. He provided no new information on this matter on Thursday, even though there was speculation that a recently announced partnership between Elon Musk’s Tesla and Intel, through Terafab, might identify Tesla as the anticipated 14A client.

When questioned about the Terafab agreement, Tan characterized it as a wide-ranging collaboration from which both firms will gain significant knowledge, but he offered little detail. “Elon and I share the view that the global supply chain is failing to match the swift increase in demand,” he commented.

Regarding 14A customers, Tan remained similarly reserved: “We are achieving excellent progress on yield and cycle time. We are actively and deeply involved with numerous potential customers. My approach is to underpromise and overdeliver. Therefore, we do not intend to make any customer announcements unless a customer chooses to do so themselves.”