Intel reported first-quarter results and second-quarter guidance that significantly exceeded market expectations, indicating the long-struggling chipmaker is benefiting from the massive artificial intelligence infrastructure build-out. Financial results show Intel's Q1 revenue grew 7% year-over-year to $13.6 billion, substantially above analysts' average estimate of $12.4 billion. Non-GAAP net income reached $1.5 billion, up 156% year-over-year, while adjusted earnings per share were $0.29, representing 123% growth and far surpassing the consensus estimate of $0.01. Gross margin improved to 41.0% from 39.2% in the same period last year.
By business segment, Client Computing Group revenue reached $7.7 billion, up 1% year-over-year; Data Center and AI business revenue grew 22% to $5.1 billion; and foundry operations generated $5.4 billion in revenue, a 16% increase. Meanwhile, Intel projected Q2 revenue between $13.8 billion and $14.8 billion, with the midpoint of $14.3 billion well above the average analyst estimate of $13 billion. The company expects adjusted EPS of $0.20 for the second quarter, also significantly higher than the $0.09 consensus, with an anticipated gross margin of 39%.
Bolstered by the strong results, Intel shares surged over 19% in after-hours trading Thursday. Before this rally, the stock had already gained nearly 81% year-to-date. The optimistic outlook suggests CEO Patrick Gelsinger is making progress with his challenging turnaround plan. After securing substantial investment last year to strengthen the company's balance sheet, he is now delivering on operational improvements.
"Everyone is starting to shift orders to Intel, and I believe we're still in the early stages," said Thomas Hayes, Chairman of Great Hill Capital. "Market sentiment has shifted from frustration to enthusiasm in a very short time."
Gelsinger stated that Intel delivered a "solid report card" with performance exceeding expectations. He anticipates continued expansion in demand for processors powering AI systems and noted the company is "highly focused" on increasing factory output, though it still cannot fulfill all orders. "Demand is extremely strong," Gelsinger said. "We're working very hard to ensure delivery and meet demand, but there remains a gap as customer requirements continue to grow."
Beyond production progress, Gelsinger has repaired Intel's balance sheet through external investments, enabling the company to repurchase partial ownership of its Irish fab that was previously sold to raise capital. This buyback is seen by investors as a vote of confidence in the future.
Gelsinger stated that compared to when he became CEO last year, Intel today is a "fundamentally different company." He noted, "A year ago, the discussion about Intel was whether we could survive. Today, the conversation is about how quickly we can expand manufacturing capacity and scale supply to meet the enormous demand for our products."
Intel CFO David Zinsner indicated that Intel's spending on new production equipment will exceed initial budgets. The company has ample factory space and will add equipment to utilize it. He stated capital expenditures are now expected to be roughly flat with last year, whereas the company had previously signaled it would reduce spending.
Despite the strong performance, Intel still faces challenges in restoring its former dominance in the chip industry. The company's annual revenue last year was $53 billion, approximately $25 billion below its 2021 peak. Wall Street projects only 3% revenue growth by 2026.
The AI Agent Era: CPU Returns to Center Stage Financial results indicate that demand for data center chips to support massive AI infrastructure development is boosting sales of Intel's flagship Xeon server processors. These general-purpose semiconductors—central processing units (CPUs)—are regaining enterprise focus as they help convert AI software into revenue-generating services.
Gelsinger stated that AI is entering the real world through physical AI, robotics, and edge AI, with CPUs once again becoming the foundation for AI. Over the past two years, the AI narrative has been dominated by GPUs, while CPUs have had minimal presence in AI servers. The reason lies in the training phase, where the core bottleneck is parallel computing capability—GPUs handle the heaviest matrix operations while CPUs manage general control and basic scheduling tasks.
However, with the explosive growth of AI agent and reinforcement learning workloads, CPUs are undergoing structural revaluation in data centers. The essence of AI agents isn't about generating longer answers but breaking down a single request into an entire workflow. Models no longer just produce an answer but execute a process. Once AI shifts from "single computation" to "running processes," system dependence on CPUs increases significantly.
The reason is that many critical workloads aren't suitable for GPUs. Tasks like job orchestration, thread scheduling, process management, sandbox execution, pre/post-processing, cache coordination, and state maintenance are typical CPU workloads. Particularly in multi-agent collaboration scenarios, where multiple agents run concurrently, call tools mutually, and share states, higher demands are placed on CPU core count, thread count, single-core performance, and memory management capabilities.
Dylan Patel, Chief Analyst at semiconductor analysis firm SemiAnalysis, stated in an April 8 interview that AI workloads are evolving from simple text generation to complex agents and reinforcement learning, with CPUs facing "extremely severe capacity shortages."
A recent TrendForce report confirms this assessment—current AI data centers maintain CPU-to-GPU ratios of approximately 1:4 to 1:8, but in the AI agent era, this ratio is expected to evolve to 1:1 to 1:2. Regarding market size, Creative Strategies predicts the data center CPU market will grow from $25 billion in 2026 to $60 billion by 2030; with additional agent-related demand, the market could approach $100 billion.
This structural shift has triggered chain reactions on both supply and demand sides. Intel and AMD increased prices for certain CPU product lines by the end of Q1 2026. Meanwhile, both Nvidia and Arm announced entries into the server CPU market in March 2026—a GPU giant and an IP licensor making identical moves in the same month represents concentrated market signaling rather than coincidence.
Intel's Xeon processors once held over 95% market share in data center CPUs. However, this dominance began weakening in 2021 when yield issues with the Intel 7 process caused nearly two years of delays for Xeon Sapphire Rapids, creating market openings for AMD's EPYC Milan.
Intel plans to launch two flagship products in 2026. One is Xeon 6+ (Clearwater Forest) using the Darkmont architecture, featuring 288 cores/288 threads with approximately 450W TDP. The other is Xeon 7 (Diamond Rapids) using Panther Cove-X architecture, with up to 256 cores/256 threads and 650W TDP. Both products utilize Intel's most advanced 18A process and introduce Foveros Direct hybrid bonding technology for the first time.
However, TrendForce notes that persistent yield issues with the 18A process could delay mass production of both products until 2027. In contrast, competitor AMD maintains steadier momentum—its 2026 flagship EPYC Venice will use TSMC's N2 process with Zen 6 architecture, featuring CoWoS-L and SoIC advanced packaging, achieving 256 cores/512 threads through simultaneous multithreading technology—the highest thread count currently available. TrendForce expects AMD to continue gaining market share from Intel throughout 2026.
Beyond the two industry giants, non-traditional players are entering the server CPU arena at unprecedented speed, seeking to fundamentally reshape the competitive landscape. In March, Nvidia announced it would sell Vera CPU as a standalone product to meet customer demand for more flexible CPU:GPU configurations. Vera uses Nvidia's proprietary Olympus architecture based on TSMC's N3 process and CoWoS-R packaging, offering 88 cores/176 threads with 1.8 TB/s NVLink-C2C interconnects for memory sharing with Nvidia GPUs. Nvidia also introduced Vera CPU racks integrating 256 CPUs per rack, totaling 22,528 cores/45,056 threads with 400 TB total memory.
Also in March, Arm announced its first self-developed CPU product, Arm AGI CPU, ending its 35-year history as a pure licensor. The product uses TSMC's N3 process with Neoverse V3 architecture, providing 136 cores/136 threads at 300W TDP, supporting DDR5-8800 memory and PCIe Gen6. Arm simultaneously launched two rack configurations: air-cooled versions integrating 60 AGI CPUs (8,160 cores, ~180 TB memory) and liquid-cooled versions supporting 336 CPUs (45,696 cores, 1 PB memory).
Major cloud service providers are also accelerating proprietary CPU development. Amazon AWS released Graviton5 (192 cores/192 threads) based on TSMC's N3 process in December 2025, deploying it alongside proprietary Trainium 3 AI ASICs to reduce AI computing costs. Microsoft launched Cobalt 200 (N3 process, 132 cores/132 threads) in November 2025. Google plans to introduce Axion C4A.metal bare-metal versions and next-generation Axion N4A in 2026, focusing on maximum cost-effectiveness.
Intel Scores Major Victory: 14A Process Lands First Major Customer Further boosting market sentiment, Tesla CEO Elon Musk announced Wednesday that the company plans to use Intel's advanced 14A manufacturing process for its Terafab chip project, making Tesla the first major customer for Intel's 14A technology. This represents a significant breakthrough for Intel as the company works to transform its business to attract external customers for its chip manufacturing technology.
Intel's upcoming 14A chip manufacturing process aims to compete with industry leader TSMC, though no significant external customers had been disclosed previously. According to previously released data, the 14A process delivers 15-20% performance improvement, 30% density increase, and 25-35% power reduction compared to the 18A process.
Earlier this month, Intel announced it would join Musk's TeraFab chip project, collaborating with SpaceX, xAI, and Tesla to help redefine silicon wafer manufacturing technology. Musk stated during Tesla's earnings call, "Given that 14A will likely be mature and ready for mass deployment when Terafab reaches scale, adopting 14A seems the right choice, and we maintain a good cooperative relationship with Intel."
Following Intel's previous chip yield challenges, this collaboration could boost investor confidence in its next-generation manufacturing process. Ben Bajarin, Principal at tech consulting firm Creative Strategies, stated that Intel's 14A technology "could be more important for Intel than people realize." He added, "Bringing multiple partners in early as design partners is crucial to streamline process flows and accumulate experience needed for advanced nodes. Tesla undoubtedly has mass production capability and represents an ideal first external customer for Intel."
Although many Terafab project details—such as who funds expensive chip manufacturing equipment, who operates the factories, and when production begins—remain unknown, Jay Goldberg of Seaport Research Partners noted that Musk's vote of confidence in Intel's technology outweighs these uncertainties. Musk has stated Terafab will eventually achieve 1 terawatt of annual computing output. Bernstein estimates building chip capacity supporting 1 terawatt annually would require $5-13 trillion in capital expenditure.
Goldberg commented that whether Musk's grand predictions about future robotics chip demand materialize remains uncertain, but even producing chips for Tesla's existing business alone represents a major victory for Intel. "In terms of chip volume, it doesn't compare to Apple or Nvidia yet," he said. "But this is a real customer that brings actual volume."
