SYSTEM NOTICE

Auto translation by AI. Be sure, accuracy, nuances and authorial intent may not be fully reflected.
見出し画像

04/24-2 | Intel up 12% after-hours despite $2.4B loss. Google, NVIDIA, and Tesla customers all on board | Trex IT NEWS

🔑 Key takeaways from this news

Intel's stock rose 12% after-hours despite reporting a $2.4 billion operating loss in its foundry business for Q1 2026. In the same earnings report, the company announced a multi-year contract with Google, the adoption of DGX Rubin by NVIDIA, and Musk's participation in Terafab, which the market interpreted as a turning point for a business structure that previously had virtually zero external customers.

🎯 The main point

Tesla, Google, and NVIDIA were all named in the same quarter for Intel, a company that was even rumored to be exiting the foundry business.

*Diagram created by Trex based on data

📌 Intel up 12% after-hours despite $2.4B loss. A quarter where external customer names were listed


Q1 earnings figures. Revenue of $13.5 billion and non-GAAP EPS of $0.29 exceeded market expectations across the board

On April 23, 2026 (US time), Intel announced its financial results for the first quarter of fiscal year 2026. The figures released by official IR are as follows:

  • Revenue:$13.58 billion (7% increase from $12.7 billion in the same period last year, exceeding the LSEG consensus of $12.42 billion)

  • GAAP EPS:-$0.73 (worsened from -$0.19 in the same period last year)

  • Non-GAAP EPS:$0.29 (more than double the $0.13 from the same period last year, significantly exceeding the consensus of $0.01)

  • GAAP Operating Loss:$3.14 billion (including $4.07 billion in restructuring costs)

  • Q2 Guidance:Revenue of $13.8–$14.8 billion (exceeding the consensus of $13.07 billion)

CEO Lip-Bu Tan explained that they have exceeded guidance for six consecutive quarters on a non-GAAP basis. According to CNBC reports, the stock price rose by up to 15% in after-hours trading following the earnings announcement, marking an unusual trend of approximately 80% growth year-to-date and about 48% growth in April alone.

The details of the $2.4 billion foundry loss. Revenue grew 16% despite being on par with the same period last year

On the other hand, Intel's foundry business (contract manufacturing division) still shows tough numbers. According to official earnings documents, the operating loss for the foundry alone was $2.44 billion. This is a slight increase from $2.32 billion in the same period last year, a business for which management had even publicly mentioned the possibility of withdrawal.

However, looking at the revenue composition, a subtle change can be read. Foundry revenue was $5.42 billion (a 16% increase from $4.67 billion in the same period last year), the majority of which is internal transfer revenue. According to TIKR's analysis,revenue from external customers was only $307 million for the full year 2025, and even looking at it by quarter, the $222 million in Q4 2025 was the company's highest quarterly value. In other words, while the external customer business exists, it is small in scale when viewed as a business. The structure of having a $2.4 billion loss on top of that has been the source of investor anxiety until now.

Intel's number of employees was 78,500 as of the end of March. This is a decrease of about 20,000 from 97,600 a year ago, and a further decrease of 1,600 from the previous quarter. Recording the $4.07 billion in restructuring costs in Q1 is an extension of this downsizing. In terms of costs, thorough restructuring is underway.

It's not just Terafab. The significance of Google, NVIDIA, and SambaNova all appearing in the same quarter

The reason the market pushed the stock price up by 12% lies more in the names of the external customers listed in the same release than in the earnings figures themselves. Organizing the official Intel release, the following four major partnerships were announced in this quarter alone.

  • Google: Deploying Intel Xeon 6 processors for Google Cloud's C4 and N4 instances under a multi-year agreement. Additionally, aiming to improve AI workload efficiency through the joint development of custom ASIC-based IPUs (Infrastructure Processing Units).

  • NVIDIA: Intel Xeon 6 selected as the host CPU for the next-generation AI server, "DGX Rubin NVL8."

  • SambaNova: Announced a blueprint for a heterogeneous configuration combining GPUs, SambaNova RDUs, and Intel Xeon 6.

  • Terafab participation: Announced on April 7, participating as a strategic partner in the $25 billion fab project funded by Musk's SpaceX, xAI, and Tesla. Tesla confirmed the adoption of the Intel 14A process during its April 22 earnings call.

What is particularly significant is that all of this has happened in the last quarter alone.Customers from three different directions—hyperscalers (Google), AI semiconductor leaders (NVIDIA), and automotive/space big tech (Tesla/SpaceX)—have stepped forward at the same time.This composition fundamentally shifts the premise of the argument that "they are withdrawing because they cannot secure external customers."

14A mass production is in 2028. Why the stock still jumped 12% in after-hours trading

Looking at it calmly, the degree to which these customer acquisitions will contribute to short-term performance is limited. The Intel 14A process used in Terafab is expected to enter mass production from 2028 onwards. While Xeon 6 for Google and NVIDIA is already in mass production, the joint development of custom ASICs has only just entered the design phase. As TradingKey points out, there is almost no immediate impact on revenue.

Even so, the reason the stock price jumped 12% in after-hours trading lies in the risk scenario the market was trying to price in. According to TIKR's analysis, the bears' rationale was based on three points: "the stock is trading at 126 times future earnings," "the 2025 foundry operating loss is $10.3 billion," and "external customers are still just names and have not translated into contracts." With this earnings report, at least the first and third points have been provided with offsetting factors.

In addition, the second-quarter revenue guidance of $13.8–$14.8 billion reported by US media is well above the consensus of $13.07 billion. This also reassured investors as evidence that the CPU market remains tight. According to Keybanc's research, the 2026 server CPU production capacity for Intel and AMD is effectively sold out, lead times for some high-end models are 6 months, and prices offered for the year are rising by about 10%. The supply-side narrative that CPUs are also becoming scarce due to AI server demand has been backed up by numbers.

Another point often overlooked is that Intel holds a unique position in advanced packaging technology. The business highlights in the earnings report explicitly mention a collaboration with NVIDIA to "package CPUs and integrate them with NVIDIA GPUs." With TSMC's advanced process nodes becoming tight, the reality is that the "process of assembling finished chips into final products" is becoming the industry bottleneck, and the expansion of assembly and testing capacity at the Penang, Malaysia plant was also announced at this time. It is a strategy for Intel, which is more than two years behind in the pure process node race, to regain its presence through packaging.

🔍 Trex's Perspective

From the perspective of someone with 15 years of IT management experience and an AI bridge engineer, the essence of this deal is not "improvement in financial figures" but "improvement in the customer pipeline." Because the foundry business is front-loaded with depreciation of built fabs and labor costs, even if contracts are signed, it will be several years before they become profitable. The reason the market is moving despite this is because it is in a phase where "whether or not there are customer names" precedes earnings figures as an evaluation axis within the industry. In the semiconductor contract manufacturing market where TSMC is the sole dominant player, the geopolitical need to secure options for foundries within the US is the motivation for putting names on the customer list. For the Japanese market, whether Intel remains or disappears as a 2nm-class procurement source is a context related to Rapidus's business plan.

💡 Today's Action

  1. Review CPU procurement plans on a 1-2 year basis: The shortage of Xeon 6 is expected to continue throughout 2026. Inventory the number of units needed by early 2027 in internal server renewal plans this week, and share estimates that incorporate longer lead times and price increases of around 10% with the procurement department.

  2. Watch for moves toward foundry diversification: The trend of Google and NVIDIA moving away from relying solely on TSMC and diversifying to US-based foundries will also spread to the Japanese equipment and automotive industries that incorporate semiconductors into their products. Update your supply chain map for the semiconductors your company uses.

  3. Treat "customer name lineups" as a management decision indicator: Intel's stock price reaction is a typical example of modern market behavior that evaluates customer pipelines ahead of earnings figures. In your company's pitch decks and IR materials, try creating one slide on how to present "announced customer names" alongside earnings figures.

Reference Links

Related Sites

HAKUTAKA Information Shop | A trend media where Fixer logically explains the background of hot topics

いいなと思ったら応援しよう!

この記事は noteマネー にピックアップされました

noteマネーのバナー