Intel Soars 20% as CPUs Set to Take Back Lost Ground in the Agent Era

Last night, Intel's stock price briefly touched $70 during intraday trading, surging 20% after hours due to its latest earnings report beating all expectations.

Intel unveiled its performance for the first quarter of the 2026 fiscal year on Thursday, reporting a revenue of $13.6 billion, a 7% year-over-year increase, surpassing Wall Street's consensus estimate by 11%. The non-GAAP earnings per share were $0.29, compared to analysts' expectations of $0.01, exceeding expectations by 29 times, a remarkably rare deviation among large-cap stocks. Following the announcement, Intel's stock price surged 20% in after-hours trading.

The Q2 guidance also pointed to a more aggressive outlook, with a revenue range of $13.8 billion to $14.8 billion, higher than the consensus median estimate. Newly appointed CEO Lip-Bu Tan made a comment during the earnings call, suggesting that CPUs are re-establishing their essential position in the AI era.

This has been one of Intel's most discussed topics in the market over the past two years, as the company was believed to have completely missed the first wave of AI.

On one hand, Intel failed to produce a GPU that could rival NVIDIA, and on the other hand, its advanced manufacturing nodes could not keep up with TSMC. But in the past 12 months, as more AI deployments shifted from model training to inference and autonomous "intelligent agent" orchestration, CPUs, the once considered basic "brains" of computers, are once again in demand. Intel's rebound this quarter marks the first financial validation of this technological narrative.

Data Center Business Sees U-Shaped Rebound

Breaking down the $13.6 billion in Q1 revenue, the most significant change came from the Data Center and AI (DCAI) segment. According to Intel's earnings report, DCAI revenue for the quarter was $5.1 billion, a 22% year-over-year increase, reaching a historical high.

This was not a one-time surge. Looking back to 2025, DCAI achieved $4.1 billion in Q1, dropped to $3.9 billion in Q2, then rose back to $4.1 billion in Q3. This sideways trend in 2025 once made the market doubt the so-called "CPU resurgence" narrative. However, in Q4, based on Intel's disclosed figures compiled by Tom's Hardware, DCAI jumped from $4.1 billion in Q3 to $4.7 billion, a +15% sequential increase, marking the fastest quarterly sequential growth rate for the company in a decade.

Entering Q1 2026, the $5.1 billion figure painted a clear U-shaped curve, with the trough in mid-2025, the inflection point in Q4 2025, and confirmation in Q1 2026. The management's explanation was that the Xeon 6th Gen "Granite Rapids" processor began mass production, coupled with an AI infrastructure refresh cycle. The company even voluntarily sacrificed some client CPU capacity, allocating wafers to data centers, boosting the profitability of the entire DCAI segment. According to Intel's Q3 2025 financial report, the operating profit margin of this segment soared from 9.2% in Q3 2024 to 23.4%, nearly 2.5 times.

Same AI Narrative, Three Different Trends

Placing Intel's rebound into peer comparison reveals a more intriguing chart than mere price fluctuations.

Using January 2023 as a baseline to April 2026, Nvidia's stock index has surged to 1023, AMD to 406, while Intel stands at 245. Despite all starting from the same point, the endpoints differ by nearly fivefold. However, what's more noteworthy is the shape of Intel's blue line. It didn't climb steadily; instead, it first plummeted all the way to 64 in September 2024 (equivalent to a 36% drop from the starting point), then exhibited a V-shaped rebound before catching up to 245 in early 2026.

This chart essentially illustrates the market's two pricing events on "who truly profited from the AI capital cycle." From 2023 to 2024, money flowed to Nvidia because training required GPUs. AMD took a bite of the second cake with its MI300 series, reflected in its stock price. Intel, on the other hand, got systematically removed from AI transactions due to lower-than-expected Gaudi accelerator sales and delayed advanced process node production. According to a third-party estimate cited by Fortune in January 2025, Nvidia's share of the AI chip market surged from 25% in 2021 to 86% in 2024, while Intel dropped from 68% to 6%.

The second pricing event occurred from the second half of 2025 to early 2026, as the market began reconsidering a crucial question: if AI progresses from training to inference and agent stages, will the compute demand structure change? The answer to this question directly determines how far Intel's blue line can go.

As the Scene Nears the Agent, CPUs Return to the Center Stage

Breaking down the AI workflow into three scenarios, the weight of CPUs varies significantly. According to Deloitte's 2026 Tech Trends report, during the large model training phase, CPUs account for only about 8% of the workflow bottleneck, with the remaining 92% of computing power pressure on GPU cluster parallel synchronization, which is NVIDIA's stronghold. Moving into the large-scale inference phase, the CPU's weight rises to 25%, but the GPU's parallel throughput and memory bandwidth remain bottlenecks.

The real shift occurs in the Agent orchestration scenario. According to a joint study by the Georgia Institute of Technology and Intel published in November 2025, CPU processing for tool invocation in the Agent workflow accounts for 50% to 90% of the total process latency, depending on the tool type and orchestration complexity. In other words, when an AI agent is doing tasks such as "calling APIs, fetching data, coordinating subtasks, managing context memory," the bottleneck is not on the GPU but on the CPU.

This trend is of a magnitude. As estimated by Deloitte, the Inference workload will represent about one-third of the total AI compute in 2023, around half in 2025, and is expected to reach two-thirds by 2026. According to the Futurum Group, the server CPU market is projected to grow from $260 billion in 2025 to $600 billion in 2030, with a growth rate exceeding the long-term historical average. A more specific signal is OpenAI's disclosed compute roadmap, where the company plans to acquire "hundreds of thousands of state-of-the-art NVIDIA GPUs and compute power scalable to tens of millions of CPUs to support Agent workloads." While GPUs still dominate, the scale of CPUs is publicly put on the same level for the first time.

The Rebound Did Not Start in Q1 2026

Looking at Intel's stock price over the past five years and six key events, the 20% post-Q1 bounce was actually the culmination of earlier decisions.

In February 2021, Pat Gelsinger returned as CEO, unveiling the "IDM 2.0" strategy to transform Intel into both a chip design company and an open-foundry. When Gaudi 3 was released in April 2024, Intel set a $500 million sales target for AI accelerators in 2024.

On August 2, 2024, the Q2 2024 earnings report revealed a revenue decline of 12.8 billion year-on-year, a GAAP EPS of -$0.38, announced a 15% workforce reduction, suspended dividends, and saw a 26% single-day stock price drop, the worst since 1974. Intel disclosed at the time that the management subsequently acknowledged that Gaudi 3 would not achieve the $500 million target for the year and took a $300 million inventory write-down.

According to an official announcement by Intel, on December 1, 2024, Gelsinger stepped down, and the company entered a period with interim co-CEOs. In February 2025, the new management decided to cancel the standalone GPU project "Falcon Shores" aimed at NVIDIA, acknowledging that their in-house AI accelerator route could not compete with NVIDIA's ecosystem lock. On March 18, 2025, former Cadence CEO and semiconductor veteran Lip-Bu Tan officially took the helm as Intel's CEO. At this point in time, Intel's stock price was around $22, only up just over 20% from its low of $18 in September 2024.

From Lip-Bu Tan's appointment to this Q1 earnings report, Intel's stock price surged from $22 to $65 before the report, with an additional 20% in after-hours trading, briefly touching around $78. If the period from August 2024 to December 2024 was considered the company's darkest hour, the real inflection point was not in Q1 2026 but the moment Falcon Shores was canceled, and Tan was selected as CEO. The company abandoned the fantasy of competing with NVIDIA and returned to its core strength in CPUs.

An EPS that exceeded expectations by 29 times sends a financial signal, but behind that are actually two simultaneous events. The market began to reprice the position of CPUs in AI architecture, and Intel coincidentally completed a change in management and product line strategy. Both of these events did not occur in Q1.