Does the Terafab project plan to make memory chips?

Yes. Per Bloomberg's coverage of Tesla's January 28, 2026 earnings call, Elon Musk told investors that Terafab would be a very big fab that includes logic, memory, and packaging, domestically. Memory production is part of the original Terafab blueprint, alongside logic chip fabrication and advanced packaging. The full scale facility is targeting 100 to 200 billion AI and memory chips per year.

Is Micron a current chip supplier to Tesla?

Yes. At the March 21, 2026 Terafab announcement, Elon Musk explicitly thanked Samsung, TSMC, Micron, and others as existing chip partners for Tesla, SpaceX, and xAI. Musk added that Tesla will continue buying all the chips its current suppliers can make, while Terafab is built out as supplemental capacity.

What is the bull case for Micron from the Terafab project?

The bull case is that Micron gets added as the American HBM partner to Terafab, similar to how Intel was added on the logic side on April 7, 2026. This would unlock a multi-year volume commitment from one of the largest single buyers in the AI buildout without surrendering pricing power, because Micron is the only American HBM producer and the political backdrop favors onshore semiconductor partners.

What is the bear case for Micron from the Terafab project?

The bear case is that Terafab proceeds without a memory partner at all, brute forces its way into DRAM and HBM production over the next decade, and eventually displaces Micron as a meaningful Tesla customer. This is the lowest probability scenario because building HBM from scratch is genuinely one of the hardest things in semiconductors. Even Samsung, with decades of memory experience and unlimited capital, has struggled to keep up with SK Hynix on HBM yields.

What should Micron holders watch for next regarding Terafab?

Watch for any announcement attaching SK Hynix, Samsung, or Micron to the Terafab project as a memory manufacturing partner. The Intel partnership announced on April 7, 2026 is the template, where the project quietly added a manufacturing partner whose institutional knowledge filled a gap Tesla and SpaceX could not fill on their own. A memory partner is the most likely next addition, and Micron is the most strategically interesting candidate because it is the only American HBM producer.

What Terafab Means for the Memory Trade and Micron Holders

Q: Will Terafab hurt Micron in the near term?

Almost certainly not. Terafab pilot production is targeted for 2026 in small batches and 2027 for volume, which means realistic chip output is likely 2028 or later. HBM memory production specifically is even further out because high bandwidth memory is harder to manufacture than logic chips, constrained by yield, stacking, and known good die selection. Tesla's chip needs from Optimus, AI5, Robotaxi, and Dojo are growing faster than Terafab can plausibly ramp, so the next 2 to 3 years look like more Tesla revenue for Micron, not less.

Q: Who produces HBM memory for the AI market?

There are three companies in the world that produce meaningful volumes of HBM (high bandwidth memory): SK Hynix, Samsung, and Micron. SK Hynix currently leads in HBM share and has the deepest stacking expertise. Samsung is the second largest. Micron is the smallest of the three but has been gaining ground aggressively, particularly through HBM3E shipments to Nvidia. Micron is also the only American HBM producer.

Rebellionaire Staff
46 minutes ago
7 min read

Futuristic scene with glowing circuit chips linked by blue energy streams on a planet-like surface. Background: stars, planets, TERAFAB text.

When Elon Musk unveiled the Terafab project on March 21, 2026, most of the headlines focused on the logic side. The 2nm process node, the AI5 chip, the orbital D3 processor, the wafer start ambitions. What got buried in the press coverage was that Terafab's stated scope is not just logic. It's also memory, which makes this a story Micron (NASDAQ: MU) holders should be sitting with, particularly after Intel joined the project on April 7.

What is the Terafab project?

Terafab is a planned semiconductor fabrication venture jointly owned by Tesla, SpaceX, and xAI, with Intel joining as a manufacturing partner on April 7, 2026. The project was first floated by Musk on Tesla's January 28, 2026 earnings call and formally announced on March 21, 2026 in Austin, Texas. Its stated goal is producing one terawatt per year of AI compute capacity at full build-out, across 100 to 200 billion AI and memory chips per year.

The official Terafab pitch is a vertically integrated facility doing logic, memory, packaging, and testing under one roof. On the January 28 earnings call where Musk first floated the project, his exact words to investors, per Bloomberg's coverage, were that Terafab would be "a very big fab that includes logic, memory and packaging, domestically." Memory is in the pitch from the original blueprint, not bolted on later.

In other words, the Terafab pitch is not just a Tesla-branded TSMC clone. It's a Tesla-branded TSMC plus Micron plus ASE clone, which is something nobody on Earth currently does at scale. Even TSMC, which is the closest thing the industry has to a vertically integrated leader, does not produce DRAM or HBM.

So when you read "Terafab will make AI chips," the more accurate translation is "Terafab will make AI logic chips and the high bandwidth memory those chips need to actually do anything useful." Those are two different industries, and the second one is the one Micron lives in.

Where Micron sits in the AI memory market today

Micron is one of three companies in the world that produces meaningful HBM (high bandwidth memory) volumes. The other two are SK Hynix and Samsung. That is the entire competitive set. Micron is the smallest of the three by HBM share but has been gaining ground aggressively, particularly on HBM3E shipments to Nvidia that came on faster than analysts modeled.

Micron is also a current supplier to Tesla. At the Terafab announcement, Musk explicitly named Micron in his thank-you list of existing partners. His exact words were "We're very grateful to Samsung, TSMC, Micron, and others, and we would like them to expand as quickly as they can. We will buy all of their chips." That's the part the news cycle skipped over, and it matters for how you frame the next 24 months.

Near term, Terafab is a tailwind for Micron, not a headwind

If you only read the Terafab pitch, you'd assume Musk is gunning for Micron's lunch. That's not the right read for the next two to three years. Here's the math.

Terafab pilot production is targeted for 2026 in small batches and 2027 for volume. Those are Musk timelines, which means the realistic version is 2028 or later for anything resembling meaningful chip output, and the realistic version for memory specifically is even further out. HBM is harder than logic. It's constrained by yield, stacking, advanced packaging, and known good die selection. Tom's Hardware ran the analysis and concluded that HBM may become the actual bottleneck for Terafab's stated ambitions, and that the front end wafer cost alone for the HBM portion would require something like twelve fabs at twenty billion each.

In the meantime, Tesla's chip needs are growing aggressively. Optimus deployment, AI5 ramp, Robotaxi, Dojo, the orbital data center fantasy. Every one of those programs needs more memory than Tesla currently buys. Until Terafab actually ships HBM in volume, Tesla is going to be buying everything Micron and its peers can make. Probably at premium pricing, because the AI memory market is structurally tight.

Translation for a Micron holder. The next few years look like more revenue from Tesla, not less. The Terafab announcement is the equivalent of a major customer telling their supplier "we'd like to expand faster than you can grow, so we're going to start trying to do this ourselves, but we'd also like to keep buying everything you can make in the meantime." That's not a threat. That's a procurement department saying the quiet part out loud.

Medium term, watch for memory partners joining Terafab

The Intel announcement on April 7 is the template for what to look for next. Terafab launched as a pure DIY pitch and has now quietly added a manufacturing partner whose entire institutional knowledge base is the thing the project was missing. The same logic applies to memory, only more so, because Tesla and SpaceX have even less institutional knowledge in DRAM than they have in logic.

The plausible memory partner shortlist is short. SK Hynix would be the obvious first call, given they currently lead in HBM share and have the deepest stacking expertise. Samsung is the second call, and is already in the Tesla orbit through the Taylor fab logic relationship. Micron is the third call, and would be the most strategically interesting because Micron is the only American HBM producer, which matters for the political backdrop the Musk companies have been navigating.

If any of those three names show up attached to Terafab in the next 6 to 12 months, the read-through to that company's stock is real and worth paying attention to. The bull case for Micron specifically is that they get added as the American HBM partner, which would unlock a huge multi-year volume commitment without surrendering pricing power, because nobody else can fill the role.

The bear case is that Terafab proceeds without a memory partner at all, brute forces its way into DRAM production over the next decade, and eventually displaces Micron as a meaningful customer for one of the largest single buyers in the AI buildout. That's the version that should worry Micron holders. It's also the lowest probability version, because building HBM from scratch is genuinely one of the hardest things in semiconductors. Even Samsung, with decades of memory experience and unlimited capital, has struggled to keep up with SK Hynix on HBM yields.

The Rebellionaire frame for supplier stocks

This is where the Rebellionaire approach diverges from how most retail investors process Musk announcements. The Teslanaire reads a Musk announcement and asks "is this bullish for TSLA?" The Rebellionaire reads a Musk announcement and asks "what does this change about every adjacent supply chain, and which of those changes is the market mispricing?"

Terafab is bullish for Tesla in the very long run if it works, mildly bearish in the medium run because of the capex drag, and roughly neutral in the near term. But Terafab is also a piece of news for Micron, for SK Hynix, for Samsung memory, for ASML, for Applied Materials, for the entire HBM packaging supply chain. Most of those second order trades are getting almost no attention right now because the market is busy debating whether Musk's cost estimate is plausible and whether the timeline is real.

If you hold Micron, today's update is a small positive. Tesla is now signaling, by adding Intel, that Terafab is going to lean on partnerships to fill capability gaps. That increases the probability that Micron eventually gets a similar call on the memory side. It also decreases the probability that Terafab successfully builds a captive HBM operation that displaces Micron as a Tesla customer.

If you don't hold Micron, this is still a useful exercise in how to read Musk announcements as a portfolio manager rather than as a fan. Every big Musk pitch reshapes a dozen supplier stories. The Rebellionaire move is to map those stories before the second order trade gets crowded, and Terafab is one of the bigger maps to draw this year.

Quick reference

Does Terafab plan to make memory chips? Yes. Per Bloomberg's coverage of Tesla's January 28, 2026 earnings call, Musk said Terafab would be "a very big fab that includes logic, memory and packaging, domestically." Memory production is in the original Terafab blueprint.

Is Micron a current Tesla supplier? Yes. At the March 21, 2026 Terafab announcement, Musk thanked "Samsung, TSMC, Micron, and others" as existing chip partners and said Tesla will continue buying all the chips they can make.

Will Terafab hurt Micron in the near term? Almost certainly not. Terafab volume production is targeted for 2027 at the earliest, HBM specifically is even further out, and Tesla's chip needs are growing faster than Terafab can plausibly ramp. The next 2 to 3 years look like more Tesla revenue for Micron, not less.

What's the bull case for Micron from Terafab? Micron gets added as the American HBM partner to Terafab, similar to how Intel was added on the logic side, unlocking a multi-year volume commitment without surrendering pricing power.

What's the bear case for Micron from Terafab? Terafab successfully builds captive HBM production over the next decade and displaces Micron as a meaningful Tesla customer. This is a low-probability scenario because HBM is one of the hardest things to build in semiconductors.

What should Micron holders watch for next? Any announcement attaching SK Hynix, Samsung, or Micron to the Terafab project. The Intel partnership on April 7 is the template, and a memory partner is the most likely next addition.