Memory Is the New Compute: Why Micron Is the AI Stock Outperforming Nvidia—And Could Hit $1 Trillion

 

 Memory Is the New Compute: Why Micron Is the AI Stock Outperforming Nvidia—And Could Hit $1 Trillion

**Subtitle:** *UBS just slapped a $1,625 price target on Micron. The Roundhill DRAM ETF is up 127% since April. And a “frugal” Idaho chipmaker is suddenly the hottest trade on Wall Street.*

**Reading Time:** 8 Minutes | **Category:** Investing & AI

## Introduction: The “$1 Trillion or Bust” Question

Just over a year ago, Micron Technology was a $100 billion company. It was respected, profitable in fits and starts, but largely viewed as a commodity player in a brutal boom-bust industry . Today, it has crossed the $1 trillion market capitalization threshold . The stock has soared roughly 850% over the past year .

The question on every investor’s mind is no longer “Is Micron an AI play?” It is “How big can this AI memory boom get?”

This week, Roundhill Investments CEO Dave Mazza made a compelling case that the answer is “much, much bigger.” Mazza, whose firm launched the Roundhill Memory ETF (DRAM) in early April, argues that memory chips have become the single most critical bottleneck in the AI infrastructure buildout .

“Demand for memory currently exceeds what can be supplied by the manufacturers,” Mazza said . And that supply-demand imbalance is not a temporary blip. It is a structural shift.

The numbers back him up. Micron’s revenue spiked 195% in the second quarter of fiscal 2026 to about $24 billion. Earnings per share popped 682% to $12.20 . The company’s entire calendar 2026 HBM supply is already fully booked, including next-generation HBM4 chips .

UBS analyst Timothy Arcuri just raised his price target for Micron to $1,625, arguing that long-term supply agreements will “structurally transform” the company’s earnings profile . He expects Micron to generate earnings per share of $155 in 2027—up from just $8.09 in 2025 .

In this deep-dive, we will break down the “memory bottleneck” thesis, analyze why Micron is outperforming Nvidia despite being a smaller player, and explain why the Roundhill DRAM ETF’s 28% allocation to Micron might be the smartest bet in AI .

> **The Bottom Line Up Front:** Nvidia created the AI boom. Micron is benefiting from the shortage that Nvidia’s success created. In semiconductor history, the biggest gains often go not to the market leader, but to the companies that control the critical bottlenecks. Memory is that bottleneck. And Micron is the only major U.S.-based memory supplier .

## Part 1: The “Bottleneck” Thesis – Why Memory Is the New Compute

To understand Micron’s rise, you have to understand a fundamental shift in AI infrastructure economics.

### The Processor vs. The Memory Wall

For decades, the computer industry focused on processing power. The CPU and then the GPU were the stars. Memory was an accessory—cheap, interchangeable, commoditized.

AI has flipped that equation.

A modern AI server does not just compute. It moves enormous amounts of data between processors and memory. Training a large language model involves continuously transferring datasets through the computing system. Inference workloads similarly require rapid access to model parameters.

As models grow larger, memory bandwidth becomes **nearly as important as raw processing capability** .

“If the systems doing that work didn't have enough memory capacity near the processors, the result would be a bottleneck that forced GPUs to repeatedly pause while awaiting fresh data,” explains one analysis .

### The HBM Revolution

High-Bandwidth Memory (HBM) has become the critical enabler. Unlike traditional DRAM, which places chips side by side, HBM stacks DRAM dies vertically using Through-Silicon Via (TSV) technology . This approach dramatically increases bandwidth while reducing power consumption.

But HBM comes with a catch. It consumes significantly more manufacturing resources than conventional memory. Advanced wafer processing, die stacking, packaging, testing, and yield management all add complexity and constrain supply .

### The Capacity Crunch

Here is the crux of the investment thesis. Memory manufacturers slashed capital spending during the 2022-2023 downturn. Just as they pulled back, AI exploded .

The result is a severe supply shortage. Micron can currently meet only about **50% to 66% of demand from its core customers** . Tight conditions in DRAM and NAND are set to linger “through and beyond calendar 2026” .

New DRAM capacity additions are limited, and meaningful output from expansion projects is not expected until fiscal 2028 .

| Supply/Demand Metric | Current Status | Implication |

| :--- | :--- | :--- |

| **Micron Demand Fulfillment** | 50-66% | Severe shortage |

| **DRAM Bit Demand Growth (2026)** | High-teens percentage | AI-driven surge |

| **New Capacity Timeline** | Not expected until FY2028 | Extended shortage |

| **2026 HBM Supply** | Fully booked | Pricing power |

## Part 2: The Nvidia Connection – A “Nudge” That Changed Everything

Micron’s surge was not inevitable. It took a direct intervention from the king of AI.

### The Jensen Huang Meeting

Three years ago, Nvidia CEO Jensen Huang met with Micron CEO Sanjay Mehrotra and laid out a vision that memory, not just processors, would become the critical bottleneck for AI .

Huang had long bet that memory would be the constraint. He needed Micron to rethink its technology and its spending.

“I was really grateful that Micron and Nvidia really lined up all of our road map,” Huang said .

### From Commodity to Co-Design

That meeting changed Micron. For decades, memory was a commodity: customers could switch suppliers easily, driving down prices. Micron was frugal, building factories on shoestring budgets and avoiding cutting-edge bets .

Nvidia’s AI build-out forced a rethink. Memory shifted from a commodity component to specialized HBM chips tailored to specific processors. Micron’s chips are now co-designed for customers, making its offerings for Nvidia distinct from those it sells to AMD or others .

### The Long-Term Supply Agreement

In March 2026, Micron announced it had signed its first five-year supply agreement—a landmark shift for an industry long driven by short-term pricing swings .

“They are seeing long-term customer demand, with real commitment,” said Ben Bajarin of Creative Strategies. “That is the key driver getting them to spend money” .

UBS’s Arcuri sees these long-term agreements as a “structural transformation” that will fundamentally change Micron’s valuation multiple .

| Before Nvidia | After Nvidia |

| :--- | :--- |

| Commodity memory | Custom HBM solutions |

| Short-term pricing swings | Long-term supply agreements |

| Frugal, cautious spending | Aggressive $20B capex for AI |

| 5-year earnings volatility | Structural profitability |

## Part 3: The UBS Price Target – $1,625 and the “Nvidia Multiple”

The most aggressive Wall Street call on Micron comes from UBS’s Timothy Arcuri.

### The $1,625 Target

Arcuri’s new price target for Micron is $1,625—up from just $535 . That represents a potential gain of nearly 80% from current levels.

The basis for the target is a fundamental re-rating. Arcuri argues that there is “no reason why Micron should trade a whole lot differently than Nvidia in terms of P/E” .

### The Earnings Explosion

UBS now expects Micron to generate:

| Year | Expected EPS | Prior Estimate | Change |

| :--- | :--- | :--- | :--- |

| **2027** | $155 | $133 | +$22 |

| **2028** | $167 | $122 | +$45 |

| **2029** | $117 | $77 | +$40 |

*Source: UBS *

For context, Micron earned just $8.09 per share in 2025 . The projected growth is staggering—and it is driven entirely by AI memory demand.

### The Consensus View

UBS is not alone. The average analyst price target for Micron is $780.66, with a high of $1,750 and a low of $249 . The average recommendation is a “Buy” .

Of the analysts covering Micron, **87% have a buy rating** on the stock .

| Analyst | Price Target | Rationale |

| :--- | :--- | :--- |

| **UBS** | $1,625 | Structural re-rating, Nvidia-like P/E |

| **Average (50 analysts)** | $780.66 | 87% buy rating |

| **High Estimate** | $1,750 | Bull case |

| **Low Estimate** | $249 | Bear case |

## Part 4: The Roundhill ETF – A Pure-Play Bet on the Memory Bottleneck

Roundhill Investments launched its Memory ETF (DRAM) in early April 2026. The timing was prescient. The fund has already delivered a return of more than 127% .

### The Portfolio

The DRAM ETF holds just 15 stocks, with a massive concentration in the three memory giants:

| Holding | Portfolio Weight |

| :--- | :--- |

| **Micron Technology** | **28.2%** |

| SK Hynix | 27.1% |

| Samsung Electronics | 17.9% |

| Kioxia Holdings | 6.5% |

| Sandisk | 5.1% |

*Source: Roundhill Investments *

The top five holdings alone account for **84.8% of the entire portfolio** .

### The Thesis

Roundhill CEO Dave Mazza has been explicit about the investment case. Memory chips are a “key constraint” on AI development as models grow more complex .

Unlike broad semiconductor funds, which dilute memory-specific upside, DRAM offers concentrated exposure to the companies supplying this critical layer of the AI stack .

### The Cautionary Note

Not everyone is buying the hype. Some analysts warn that the current supply-demand imbalance is unsustainable. Memory companies are frantically building more manufacturing capacity, which will boost supply and help stabilize prices .

“Once production and demand are more in sync, Micron could actually see its earnings decline, and it is unlikely that it will be able to maintain its record profit margins” .

The Roundhill ETF has already gained 127% since its April launch. Chasing that gain now carries significant risk .

## Part 5: The Trillion-Dollar Question – Can Micron Keep Growing?

Micron has crossed the $1 trillion mark. The question is whether it can stay there—and go higher.

### The Bull Case

- **Supply constraints are structural:** New DRAM capacity will not come online until 2028. Shortages persist through 2026 and 2027 .

- **Pricing power is unprecedented:** Micron’s 2026 HBM supply is fully booked at premium pricing. The company can only meet 50-66% of demand .

- **Long-term agreements transform earnings:** Five-year supply contracts replace short-term pricing volatility with predictable revenue streams .

- **U.S.-based supplier advantage:** As customers diversify away from Korea and governments push for domestic supply chains, Micron’s position as the only major U.S. memory supplier adds an edge .

### The Bear Case

- **Capacity will eventually catch up:** Memory companies are investing heavily. When new supply arrives, pricing will normalize .

- **AI companies are cost-constrained:** Uber burned through its entire 2026 AI budget in four months. If AI companies cut spending, memory demand will fall .

- **The cycle will turn:** Memory has always been cyclical. AI may make the market structurally larger, but not immune to slowdowns .

### The HBM4 Catalyst

Micron’s HBM4 chips are on track to ramp with high yields in the second quarter of 2026 . They will be a key component in Nvidia’s upcoming Vera Rubin platform .

Barron’s reports that Micron is expected to join Nvidia’s HBM4 supply chain starting in the second quarter, shifting the market from a two-horse race (Samsung and SK Hynix) into a three-way competition .

Nvidia is likely to source from all three suppliers for supply chain stability. Micron should maintain a market share of approximately **20-25% overall** .

| Catalyst | Timeline | Impact |

| :--- | :--- | :--- |

| **HBM4 Mass Production** | Q2 2026 | Nvidia supply inclusion |

| **Nvidia Vera Rubin Launch** | Late 2026 | Major HBM4 customer |

| **Long-term Supply Agreements** | Signed March 2026 | Structural earnings shift |

| **New DRAM Capacity Online** | FY2028 | Supply normalization risk |

## Frequently Asked Questions (FAQ)

**Q: Is Micron a trillion-dollar company now?**

A: Yes. Micron crossed the $1 trillion market capitalization threshold on May 26, 2026, joining an elite group of companies .

**Q: What is the Roundhill DRAM ETF?**

A: It is an exchange-traded fund launched in April 2026 that focuses specifically on memory chip companies. Its largest holding is Micron at 28.2%, followed by SK Hynix and Samsung . The fund has returned over 127% since its launch .

**Q: Why is Micron outperforming Nvidia?**

A: Nvidia created the AI demand, but memory has become the critical bottleneck. In semiconductor history, the biggest gains often go to the companies controlling the shortages, not just the market leaders. Micron is benefiting from severe supply constraints and pricing power .

**Q: What is HBM and why does it matter?**

A: High-Bandwidth Memory (HBM) is a specialized memory chip that stacks DRAM dies vertically to dramatically increase data transfer speeds. It is essential for AI workloads because it prevents GPUs from waiting for data .

**Q: Is Micron’s stock expensive?**

A: On a forward P/E basis, Micron actually looks cheap. Wall Street expects earnings per share of $155 in 2027. At current prices, that gives Micron a forward P/E of roughly 7—far below the Nasdaq-100 average of 35 . However, these earnings may not be sustainable once supply normalizes .

**Q: What is UBS’s price target for Micron?**

A: UBS raised its price target to $1,625, up from $535. The analyst argues that Micron should trade at a similar P/E multiple to Nvidia due to its structural transformation .

## Conclusion: The “Frugal” Giant Comes of Age

We started this article with a question: Can Micron become the next trillion-dollar AI giant?

The answer is that it already has. But the question now is whether it can stay there—and go higher.

The bull case is compelling. Supply is constrained. Demand is insatiable. Long-term agreements have transformed the earnings profile. And Micron’s HBM4 chips are about to become a key component in Nvidia’s next-generation platforms.

But the bear case is also real. Memory has always been cyclical. New capacity will eventually come online. And AI companies are already feeling the cost pressure of the infrastructure buildout.

**For the Investor:**

Micron is no longer a “hidden” AI play. The stock has surged 850% over the past year. The Roundhill DRAM ETF is up 127% since April. The easy money has been made. But the structural shortage may persist for years, providing continued upside.

**For the Trader:**

The volatility is extreme. Options premiums are elevated. Consider defined-risk strategies.

**For the Long-Term Believer:**

The memory bottleneck is real. Micron’s long-term supply agreements are a game-changer. And the shift from commodity to co-designed custom chips is a structural transformation. The company that was once too frugal to invest may be the one best positioned to capitalize on the AI boom.

**The Bottom Line:**

Nvidia created the AI revolution. Micron is benefiting from the shortage that revolution created. The memory bottleneck is the critical constraint on AI growth. And the only major U.S. supplier of that memory is Micron.

That is the trillion-dollar thesis. And it is just getting started.

---

**#Micron #MU #AI #HBM #Semiconductors #Roundhill #Investing #MemoryChips**

---

*Disclaimer: This article is for informational purposes only. It does not constitute financial advice. Stock markets are volatile; always consult a licensed professional before making investment decisions.*