Musk predicts: severe shortages of AI chips and memory; Terafab to be built at Giga Texas, spending $3 billion to run Intel 14A for SpaceX delivery

📄Full Article· Automatically extracted by trafilaturaGemini 翻譯1718 words

Tesla CEO Elon Musk warned during the Q1 2026 earnings call: both logic and memory chips will face severe shortages. If Tesla does not manufacture them in-house, the company will inevitably hit production bottlenecks—this is the very reason for the birth of Terafab. During the earnings call, he disclosed for the first time that the research facility will be located at the Giga Texas campus, with a budget of approximately $3 billion, utilizing the Intel 14A process, and producing thousands of wafers per month. He also revealed the core logic of the integrated architecture: mask creation, logic, memory, and packaging will all be housed within the same building, while the "scaling phase" from research facility to mass production will be undertaken by SpaceX. (Context: What is Terafab? How does Musk, who claims global chip supply meets only 2% of demand, plan to build a factory "larger than TSMC"?) (Background: Musk responds to C.C. Wei: Terafab will always be TSMC's largest customer, not a competitor.) "In terms of industry growth speed, we anticipate that if we do not manufacture chips ourselves, we will encounter bottlenecks for logic chips, and even more so for memory chips. This is why Terafab was born." On April 22, during the Tesla Q1 2026 earnings call, this statement by Musk provided the most straightforward definition of the Terafab project: it is not a bargaining chip to pressure suppliers, but a realization that Tesla sees no path to rely on external supply amid the explosion of AI compute—in-house production is the only solution. During the earnings call, Musk disclosed the specific outline of the research facility for the first time: the site is confirmed at the Giga Texas campus, with a budget of approximately $3 billion, a monthly capacity of several thousand wafers, and the use of the Intel 14A process. Musk explained that Terafab is planned to start as a research facility and gradually scale up to mass production; by the time Terafab enters the "initial scaling phase," the Intel 14A process is expected to be ready for mature mass production, aligning with the timeline. Notably, the "scaling phase" from research facility to mass production will not be handled by Tesla alone—Musk explicitly stated that SpaceX will take over this execution. The mass production expansion and system integration experience accumulated by SpaceX in rocket manufacturing are seen as the key leverage for Terafab to bridge the gap from laboratory scale to industrial scale. The most disruptive design of the research facility is the ultimate vertical integration: lithography mask creation, logic chips, memory chips, and packaging are all concentrated in the same building. Musk stated bluntly during the earnings call: "You won't find another place in the world that puts these four things under the same roof." There is a clear engineering logic behind this integrated architecture—a quick iteration loop. In the traditional chip development process, mask creation, logic design, memory integration, and packaging/testing are dispersed among different suppliers and locations, with every round of back-and-forth costing weeks or months. Terafab compresses the entire chain into one building, aiming to allow Tesla to iterate AI chip designs at several times the industry speed, supporting the rapid evolution of product lines such as the Optimus robot, FSD, and xAI compute. The new information from the Q1 earnings call is the latest node in a series of recent Terafab actions. Earlier, Intel officially announced its participation in the Terafab project, joining SpaceX and Tesla in building the terawatt-level AI compute infrastructure. On the talent front, Tesla has actively recruited in Taiwan, offering triple the salary to attract engineers familiar with 2nm processes and CoWoS advanced packaging, directly targeting TSMC's core technical talent pool. Addressing external concerns about competition with TSMC, Musk made his stance clear: "Terafab will always be TSMC's largest customer, not a competitor." This statement echoes the positioning logic of Terafab—the goal is to fill the absolute gap in AI compute, not to replace the existing foundry system. He had previously suggested that global AI compute annual capacity is about 20 GW, which only meets 2% of his long-term demand estimates; the scale of the gap is unimaginable for the existing supply chain. Looking at the financial figures, the $3 billion budget for the Terafab research facility is just the beginning. Analysts at Bernstein estimate that to truly fill the compute gap described by Musk, the total construction cost could reach as high as $5 trillion, while Tesla's official total budget range is $25 billion to $40 billion. The first phase is scheduled to begin production in the second half of 2027, reach mass production in 2028, and be fully completed by 2030. When the world's most high-profile tech entrepreneur decides to build

Data Status✓ Full text extractedRead Original (動區 BlockTempo)

🔍Historical Similar Events· Keyword + Asset Matching1 items

2026-04-23

Musk warns: AI chips and memory are in severe shortage; Terafab is being built at Giga Texas, burning $3 billion to run Intel 14A for SpaceX delivery.

Similarity 330%關鍵字記憶體嚴重缺貨/texas/ai晶片

💡 Currently matching via keywords + symbols (MVP) · Will be upgraded to embedding semantic search later

Raw Information

ID:82aa807601

Source:動區 BlockTempo

Published:2026-04-23 02:53:44

Category:zh_news · Export Category zh

Symbols:Unspecified

Community Votes:+0 / −0 · ⭐ 0 Important · 💬 0 Comments