To The Moon

In 2026, $Micron Technology(MU)$ went from a sleepy memory maker to one of the most profitable companies on the planet, with the stock up about 273% this year.

Yet the market still prices it like a cyclical stock at the top of its cycle. After the Q3 report in June, MU jumped above $1,200, then fell roughly 40% over the following month as investors bet that memory prices were about to peak.

Yet, Micron just posted its 6th revenue record in a row, and the numbers still look unreal.

• Revenue hit $54.2B, +379% Y/Y, above the high end of guidance.

• Gross margin climbed to 87.0%.

• Data center SSD revenue reached nearly $10B, more than 10x the year-ago quarter.

• Signed long-term contracts jumped from 16 to 26, with roughly $150B in RPO.

• Q1 FY2027 guidance of $61.5B came in $4.5B above analyst estimates.

In this article, I will look at Micron's Q4 FY2026 results, its long-term contracts, the capex plan, current memory pricing, and the key points from the earnings call.

Let's dig in!

1. Financial Results

Overall, Micron delivered another incredible quarter, beating all the estimates and growing much faster than analysts expected!

• Revenue: $54.2B +379% Y/Y, vs $51.1B estimate. BEAT

• Gross Margin: 87.0% vs 86.0% guidance. BEAT

• Operating Income: $44.6B +1,029% Y/Y

• ADJ EPS: $33.42 +1,003% Y/Y, vs $31.61. BEAT

• GAAP Net Income: $37.7B +1,078% Y/Y

• OCF: $44.0B +667% Y/Y

• FCF: $33.2B vs $803M a year ago

• Q1 FY2027 Revenue Guidance: $61.5B vs $57.0B. BEAT

• Q1 FY2027 EPS Guidance: $38.15 vs $35.14. BEAT

1.1. Revenues

Micron's Q4 FY2026 revenue grew +379% Y/Y to a record $54.2B!

That is +31% Q/Q and well above the $50B ± $1B guidance. It also beat the $51.1B analyst consensus by 6.2%.

Core Data Center segment overtook Cloud Memory as Micron's largest unit, growing by an insane 1,042% Y/Y to $18B!

Meanwhile, phone and PC unit sales are expected to fall by over 10% this year, because of higher memory prices. Despite that, Micron’s mobile unit grew by 249%, because prices are exploding.

1.2. Margins

Gross profit grew by 831% to $47B, with the margin expanding to an incredible 87.0%!

Think about that for a second. For every $100 of memory Micron sold, it kept $87 after manufacturing costs.

Cost of goods sold rose just +14% Y/Y to $7.0B, while revenue almost 5x'd. Nearly all of this growth came from price increases.

The memory market has run out of supply, and buyers are paying whatever it takes to get bits.

Operating expenses jumped $1.1B Q/Q to $2.6B, and even after the jump, opex is just 4.7% of revenue.

Operating income grew +1,029% Y/Y to $44.6B, an 82.3% operating margin!

Net income rose +1,007% Y/Y to $38.4B, a 68% margin!

EPS of $33.42 beat the $31.61 analyst estimate by 5.7% and grew +33% Q/Q.

1.3. Cash flow and balance sheet

OCF grew +667% Y/Y to $44B, a full 81% of revenue!

Capex was $11.1B, leaving FCF of $33.9B. A year ago, FCF was nonexistent. For the full year, FCF came in at $62.3B versus $3.7B in FY2025, growing 20x.

Last quarter, some investors, like the short seller Michael Burry, were worried that receivables were growing faster than sales. This quarter, receivables rose +17% Q/Q to $36.2B, slower than the +31% revenue growth. So that worry was just silly, and looked at a single quarter.

Not only are customers not delaying payments to Micron, but they are handing Micron $12.3B in cash prepayments tied to the long-term contracts.

Micron ended Q4 FY2026 with $43.4B in cash against just $5.2B of debt. However, that significantly understates Micron’s actual cash position, because this quarter the company spent $26B on long-term investments, increasing its long-term investment position to $30B. These are purchases of bonds and treasuries with durations over 1 year. Essentially, the company has so much cash that they are locking it in for a loinger time at 4-5% interest rate, as a result interest income is likely to rise significantly in the next few quarters.

So Micron’s actual cash-like position is about $73.5B, more than doubling Q/Q!

Inventory rose to $10.4B, with management blaming build-ahead for older nodes reaching end of life, plus incentive pay that got absorbed into inventory cost. They expect inventories to fall in the coming quarters.

You are probably wondering what Micron plans to do with all this cash?

Well, they will give it back to shareholders!

Micron bought back no stock in Q4 and just $650M in all of FY2026. That will change soon. Because the company received government subsidies from the CHIPS ACT, it was banned from increasing buybacks.

However, from December 9, 2026, the 2nd anniversary of its CHIPS Act agreements, Micron plans to step up capital returns, and over time it wants to return 100% of excess cash to shareholders.

That means Micron plans to spend tens, if not hundreds, of billions of dollars on shareholder returns in the next 2-3 years, while trading at 6x FWD P/E.

1.4. Guidance

Micron guided Q1 FY2027 revenue to $61.5B ± $1.5B, $4.5B above analyst estimates!

Meanwhile, EPS guidance of $38.15 ± $1.00 compares favourably to the $35.14 estimate.

Gross margin is guided slightly lower at 86.25% because of higher incentive pay that went into Q4 inventory, which now flows through cost of goods sold as that inventory is sold.

Management called Q1 the gross margin floor for FY2027 and expects revenue to grow every quarter of the year, with a slower pace of price increases.

Opex is set to rise about $2.5B in FY2027, mostly R&D.

While in regards to capex, Micron plans about $11.5B in Q1 and $25B in the first half, with the second half higher still. That points to more than $50B of FY2027 capex, almost double the $27.4B spent in FY2026.

This capex plan is what pushed the stock lower after hours, however, with Q4 OCF alone at $44B, I think Micron can fund it comfortably.

2. HBM

HBM revenue grew faster than Micron's total revenue in Q4 FY2026!

Micron gives no exact HBM figure, so we have to read it through the Cloud Memory unit, where most HBM sits.

Here is the problem with HBM. It is a stack of DRAM dies packed next to the GPU, and Micron has said it eats roughly 3x the wafer capacity per bit of standard DDR5. During 2026, regular DRAM prices rose so fast that HBM, priced on older annual contracts, became the lower-margin product. Every extra HBM dollar pulled CMBU margins down a little.

Micron has locked in the vast majority of its CY2027 HBM bit supply with significant price increases!

Management said this narrows the gross margin gap with conventional DRAM. To me, this is the single most important HBM line in the release. Once 2027 prices kick in, HBM mix growth should stop being a drag on margins.

On product, the HBM4 ramp is going well. Last quarter, Micron said HBM4 revenue had already passed $1B and that it was ramping 2x faster than HBM3E.

Micron is working with Nvidia on NVHBM, the industry's 1st custom HBM4E!

NVHBM will be used on Nvidia's next GPU generation and its NVLink Fusion platforms. Custom HBM is a real competitive advantage. Once a memory maker designs a custom base die with a GPU vendor, swapping in a rival supplier gets much harder.

Capacity is still the main bottleneck. Last quarter, Micron CEO Mehrotra said Micron could serve only 1/2 to 2/3 of customer HBM demand.

Some help is coming. Cleanroom work at the Singapore HBM packaging plant is ahead of plan, with first output expected in early calendar year 2027.

Micron expects industry HBM bit shipments to grow faster than conventional DRAM through 2028. It is still the smallest of the 3 HBM suppliers, behind SK Hynix and Samsung. With demand far above supply, the company is in a good position to take share.

3. DRAM

DRAM revenue grew +343% Y/Y to a record $39.8B!

That is +27% Q/Q, with DRAM making up 73% of total revenue, down from 79% a year ago as NAND catches up.

Meanwhile, bit shipments rose by a mid-single-digit % Q/Q, while prices rose by a high-teens %.

Micron is selling every bit it can make, and basically the entire DRAM growth this quarter came from price.

On supply, Micron expects industry DRAM bits to grow in the mid-20s % in CY2026, with Micron in line

For CY2027 and CY2028, industry bit growth slows to the low-20s % and the market stays supply constrained.

Micron expects server units to grow in the high teens % in CY2026 and CY2027, but because of high memory prices, memory content per server is growing a bit slower than it expected.

Customers are building servers with less memory than they want because they simply can't get enough, or they don’t want to pay a higher price.

"Even with additional industry DRAM cleanroom space plans, with robust demand trends including new upside requests from customers, we do not have line of sight to when supply and demand will return to balance."

4. NAND

NAND revenue grew +526% Y/Y to a record $14.1B!

NAND grew +42% Q/Q, much faster than DRAM's +27%. Its share of Micron's revenue climbed from 20% a year ago to 26%.

Bit shipments rose about +10% Q/Q while prices rose about +30%.

Data center SSD revenue reached nearly $10B, more than 10x the year-ago quarter!

Last quarter, data center SSD revenue was a little over $5B, so it nearly doubled Q/Q. It now makes up more than 2/3 of all NAND revenue. Micron is on track for its 5th straight year of record data center SSD market share gains.

The driver is AI inference. When an AI model works through a long chat or an agent task, it stores what it has already read in a KV cache. Longer contexts make that cache too big for GPU memory, so it spills onto fast SSDs. On top of that, high-capacity SSDs are replacing hard drives in data centers. NAND sales outside the data center were roughly $4B to $4.5B.

For CY2026, Micron expects industry NAND bits to grow in the low-20s %, slightly above its prior view. Micron's own NAND supply will grow less than the industry. For CY2027 and CY2028, industry bit growth rises to the mid-20s % and the market stays supply constrained.

5. Long-Term Contracts

The main reason why Micron’s stock trades at such a low multiple is that investors expect the current gravy train to stop sooner rather than later. Micron knows that, so it has been locking in current prices with long-term agreements.

Micron now has 26 Strategic Customer Agreements, up from 16 just 1 quarter ago!

An SCA is a multi-year take-or-pay contract. The customer commits to buy a set volume of memory. If it doesn't take the volume, it still pays.

In just 1 quarter, 10 new contracts added about $50B of RPO and $10B of customer commitments.

Pricing works in 2 layers. About 3/4 of expected SCA revenue has a defined pricing framework, and most of that uses price bands with a floor and a ceiling. The other 1/4 is repriced from time to time at market prices. Any new SCA that includes pricing is now negotiated at higher prices.

Customers also want supply beyond 2030. Micron has signed SCAs that run into 2031, and 2 existing agreements got 1-year extensions to 2031.

RPO jumped to roughly $150B!

RPO only counts SCAs with a set pricing framework. It multiplies the committed volume by the minimum price, so it is a floor, not a ceiling. Management expects actual revenue to come in well above it. Most importantly, CFO Mark Murphy repeated that even at floor prices, margins would be meaningfully above any prior cycle peak.

Customers handed Micron $12.3B in cash deposits in Q4 alone!

That is more than the $11.1B Micron spent on capex in the quarter. Customer deposits on the balance sheet stood at $12.7B at quarter-end. The cash is unrestricted, so Micron can use it to build fabs. It gets paid back toward the 2nd half of each contract term, as long as the customer meets its minimum purchases.

On the call, management said about 75% of its CY2027 output is already committed, and most customer talks have moved on to 2028. Finance website TheStreet reported a long-run goal of bringing SCAs to about 50% of revenue.

"Customers want SCA-assured supply beyond 2030, and we have now signed SCAs that extend into 2031."

6. Capex Plan

Micron's FY2027 capex is on track to pass $50B, almost 2x FY2026!

In FY2026, Micron spent $30.7B on property, plant, and equipment. Net of $3.3B in government incentives, net capex was $27.4B.

Looking at the next 3 years' capex estimates, they look too low!

So the $48.3B FY2027 estimate is already below $50B management guidance, so analysts will likely increase it in the next few days.

Is this a red flag? I don't think so. FY2026 net capex was 21% of revenue. Even at $55B, FY2027 capex would be roughly 20% of revenue if sales grow from the $61.5B Q1 guide. In past cycles, Micron spent in the mid-30s % of revenue.

Most of the increase goes to construction, building cleanrooms that should be ready from late CY2028. Micron is also pulling forward some equipment purchases to squeeze more output from cleanroom space it already has.

Micron also plans to expand and modernize its existing fab in Manassas, Virginia, as part of its ~$200B US investment plan.

So what capacity actually comes online, and when?

CY2027: the first wave is small. Tongluo is the fastest source of new DRAM because Micron bought an existing cleanroom instead of building one. The first phase starts producing wafers in mid-CY2027, but management said new fabs only become meaningful a few quarters after first output. Meanwhile, the Singapore HBM plant adds packaging capacity, which lets Micron turn more DRAM wafers into finished HBM stacks.

CY2028: the big wave starts. Tongluo ramps to volume and Hiroshima in Japan start DRAM output late in the year, and the Singapore NAND fab follows in the 2nd half. By the end of CY2028, I count roughly 1.5M sq ft of new DRAM cleanroom across the 2 Idaho fabs and Tongluo, before adding Hiroshima and Tongluo's 2nd building.

If Micron keeps pace with the industry's expected low-20s % DRAM bit growth in CY2027 and CY2028, these new fabs plus node upgrades could let it make roughly 45% more DRAM bits by 2029 than in 2026.

CY2030 and beyond: New York. The 1st New York fab is the long-term supply source, with up to 3 more fabs to follow.

Here is the problem for all the memory bears predicting a quick price drop.

Almost nothing on this list adds meaningful DRAM bits before mid-CY2027, and most of it lands in late 2028. That timing matches management's view that 2027 and 2028 will be tighter than 2026.

So the memory shortage is not only not easing in 2027 and 2028, it will likely get worse.

The AI-driven demand for memory capacity will likely grow faster than the 45% in the next 2-3 years. That means that the demand and supply imbalance won’t disappear.

7. Key Points from the Earnings Call

A buyback wave is coming!

"So near term, based on our Q1 guide and including the CapEx number that you heard, you will see free cash flow significantly higher than the $33 billion we reported in fiscal Q4." (Mark Murphy, CFO)

"Finally, I will just add that our current authorization, that you can see from our previous filings, for share repurchase stands at $2.2 billion. You can assume that we will seek additional authorization in the near term." (Mark Murphy, CFO)

Micron expects to hit its target cash level by the end of Q1 FY2027. After that, excess cash goes back to shareholders, mostly through buybacks. A $2.2B authorization is tiny next to $68.3B of net cash and $33B+ of quarterly FCF.

According to analyst estimates, Micron could generate more than $100B of FCF from Q2 to Q4 FY2027.

At a market cap of about $1.2T, returning even most of that would buy back around 8% of the company in a year. I expect a very large authorization soon after December 9.

Low capital intensity is the new normal!

"Capital intensity is low on historic levels, and that is reflective of the strategic asset that memory and storage has become, and the industry is structurally reset." (Mark Murphy, CFO)

Micron used to talk about capex in the mid-30s % of revenue, yet FY2027 is heading for about 20%. Murphy pushed back on the idea that this will snap back to historical averages.

This is the core of the bull case. Memory stocks always traded at low multiples because every boom brought a capex binge and then a bust. If capex stays near 20% of revenue while customers prepay through long term agreements, the cycle gets much milder.

New fabs take a long time to matter!

"Clean rooms take a long while to build. Even after they are built, even after first wafer output, production ramps up only gradually in the clean rooms." (Sanjay Mehrotra, CEO)

"Node transitions of the future give less productivity gain per wafer as well." (Sanjay Mehrotra, CEO)

Mehrotra listed several supply headwinds. Each HBM generation uses more wafers per bit, while new nodes add fewer bits per wafer than older node shrinks did. People expecting fast memory supply ramp don’t realize that cleanrooms take years to fill with equipment.

This is why the 2028 fab wave in the Capex Plan chapter may hit less hard than the headline square footage suggests. Node shrinks used to deliver most of the industry's bit growth for free. Now the industry has to build space, and space is slow.

Customer "de-specs" are a sign of shortage!

"So when customers reduce their content growth versus prior expectations, as I referred to in certain platforms, that really is to enable them to ship more units." (Sanjay Mehrotra, CEO)

"These optimizations also have diminishing return for any further optimizations." (Sanjay Mehrotra, CEO)

During the conference call, an analyst asked about reports that a large customer (Nvidia) is cutting HBM per GPU. Bears read this as demand weakness. Mehrotra's answer is that customers are rationing memory per server so they can build more servers.

I agree with him. It lines up with TrendForce's report that GPU makers are moving to 8-Hi HBM stacks because of supply, while end demand stays strong.

And a de-spec only works once, as you can cut a server from 12 stacks to 8, but you can't keep cutting forever while AI models keep getting bigger.

8. Conclusion

The main story is how much the business model has changed!

Micron now has 26 SCAs with about $150B of RPO and customers have committed $32B. More than 75% of its 2027 output is already committed. Customers are prepaying for supply, and capex is running near 20% of revenue against the mid-30s % of past cycles.

There are real risks. Price increases have slowed from +90% Q/Q at the start of 2026 to 10% to 15% now. PC and phone makers are hitting their limit. Capex will top $50B in FY2027, and in 2028 and 2029 a wave of new fabs from all 3 suppliers will start to ramp.

SCA price ceilings also cap part of Micron's upside while prices keep climbing.

I believe the market is still pricing Micron as if 2026 is the peak. The evidence from this report points that the peak being nowhere near, and likely won’t be reached until 2029 or later. New cleanrooms ramp slowly, and node shrinks add fewer bits than they used to. Customers are even cutting memory per server just to build more servers.

At roughly 7x FWD P/E, Micron trades like the cycle ends tomorrow!

I think the memory supercycle has at least 2 more years to run, and the SCAs make the eventual landing much softer than in any past cycle. If you own Micron, this quarter gave you a lot of reason to hold on.

Markets are always moving - and sometimes, the best move is knowing what works for you.

With Treasury yields, oil prices and rate expectations keeping markets on edge this week, investors are once again thinking carefully about where to position next. There’s no one-size-fits-all choice in investing — and the same goes for Tiger Merch. This month’s hot picks are in, featuring the Tiger Toiletry Bag, Universal Travel Adapter, Tiger Umbrella and more favourites chosen by fellow Tigers.

Explore the Monthly Hot Picks in Tiger Coin Mall, now 12% OFF for a limited time.

https://laohu8.com/J/redeemGift?goodID=100561&type=delivery