Pro Ticker Research Brief · Memory Chips: Reconcile Our Glut Story Before Selling Tesla’s Shortage Hedge

Research created at 8:39 AM ET, November 10, 2026 · Full analysis

1. EXECUTIVE SUMMARY

• REVENUE GROWTH IS NOT EVIDENCE OF A GLUT. Gartner’s August 24 forecast put 2026 memory revenue at $837.3 billion, versus $220.1 billion in 2025—about 3.8 times higher. That measures dollars, not excess physical supply; an August glut argument based on that increase requires correction. [10]

• PRODUCT AND FORECAST PERIOD EXPLAIN PART OF THE APPARENT CONTRADICTION. TrendForce’s July outlook anticipated NAND supply constraints easing in second-half 2027, while Micron subsequently projected demand exceeding supply in 2027 and 2028. Those are different market scopes and outlooks, not interchangeable descriptions of today’s memory balance. [30][6]

• THE SHORTAGE CASE RESTS ON PRICING AND DELIVERY CONDITIONS. Deloitte reported AI-server DRAM costs roughly doubling in first-quarter 2026, with a fourfold full-year increase expected; Team Group’s November 2 warning pointed to worsening conditions from end-2026 through end-2027. These support procurement-risk analysis, not automatic outside chip revenue for Tesla. [7][4]

2. WHAT IS HAPPENING

As of 8:39 AM ET on Tuesday, November 10, the central issue is whether the memory shortage supports a defensible Tesla Terafab business case—not whether a large memory-market forecast can be presented as Tesla’s prospective revenue.

The latest industry warnings favor continued tightness. Team Group CEO Chen Ching-wen warned on November 2 that conditions could deteriorate from end-2026 through end-2027, with equilibrium potentially delayed until the latter half of 2028. Micron’s recent commentary likewise projected demand exceeding supply in calendar 2027 and 2028. Together, those statements suggest that availability, as well as price, belongs in buyers’ procurement planning. [4][6]

The August-versus-November comparison needs three separate tests:

• AUGUST 24 GARTNER FORECAST: What the figures describe Total memory revenue of $837.3 billion in 2026, versus $220.1 billion in 2025. [10] · Implication Revenue expansion alone establishes neither a glut nor a shortage.
• JULY 21 TRENDFORCE OUTLOOK: What the figures describe NAND supply growth outpacing demand in 2027, easing constraints in the second half. [30] · Implication A forward NAND easing scenario can coexist with current shortages.

• NOVEMBER SHORTAGE COVERAGE: What the figures describe Team Group’s worsening outlook through end-2027 and Micron’s demand-over-supply outlook for 2027–2028. [4][6] · Implication Broader shortage forecasts still need product-specific comparison with NAND expectations.

PRODUCT MIX AND REPORTING PERIOD PROVIDE A PARTIAL RECONCILIATION; THEY DO NOT RESCUE A REVENUE-BASED GLUT CLAIM. A headline about future NAND easing can be consistent with current AI-memory scarcity. A headline asserting a present, industry-wide glut because revenue forecasts increased confuses market value with supply balance and should be corrected on that basis.

Tesla’s opportunity requires the same discipline. TrendForce’s May 29 forecast of $618.7 billion in 2026 DRAM revenue, as reproduced in the shortage chronology, concerns DRAM rather than the entire memory market. It should not be added to Gartner’s broader memory forecast, treated as Tesla’s addressable sales, or used to establish a competitor’s wafer capacity. [11][10]

3. WHY IT IS HAPPENING

PRICING, MIX AND CAPACITY OPERATE ON DIFFERENT CLOCKS

Deloitte’s July 28 analysis attributed the crunch to AI infrastructure demand, reporting that AI-server DRAM costs roughly doubled in the first quarter and forecasting a fourfold increase over the full year. Fourfold means approximately a 300% increase, not 400%; the distinction matters when translating industry forecasts into procurement budgets. [7]

Revenue can therefore rise sharply even when shipment growth is modest or customers cannot obtain enough product. Conceptually, revenue reflects quantity, price and product mix. A higher-value AI-memory mix and rising prices can enlarge the dollar market without establishing abundant capacity.

Capacity forecasts also differ by segment and project. TrendForce anticipated NAND constraints easing in second-half 2027; CSIS described partial nearer-term relief and a larger expansion later, while Deloitte placed substantial additions in 2029–2030. These should be read as differing supply horizons, not combined into a single industry-wide completion date. [30][8][7]

That creates a timing problem for vertical integration: a facility justified by today’s scarcity may reach economic scale under a different pricing regime. The relevant comparison is not simply current outside purchase price against prospective manufacturing cost. It is purchase prices over the facility’s operating life versus fully loaded internal costs and capital requirements.

AVOIDED PURCHASES AND MERCHANT SALES ARE DIFFERENT BUSINESSES

For a Terafab valuation framework, internally consumed chips and chips sold to outside customers require separate calculations:

• INTERNAL PROCUREMENT SAVINGS: Appropriate analytical treatment Qualified internal units multiplied by the difference between outside purchase cost and fully loaded internal production cost.
• SUPPLY CONTINUITY: Appropriate analytical treatment Expected production losses avoided through more reliable access, without counting the same benefit again in procurement savings.

• EXTERNAL CHIP REVENUE: Appropriate analytical treatment Qualified units actually sold to outside customers multiplied by realized selling prices.
• INVESTMENT RETURN: Appropriate analytical treatment Cash benefits after operating costs, capital spending, working capital and ramp requirements.

These are analytical distinctions, not Tesla guidance. Internal chip consumption does not create consolidated third-party revenue. Outside sales introduce a separate commercial test: product suitability, customer qualification, delivery commitments and competitive pricing.

Likewise, wafer starts are not saleable output. A meaningful competitor-capacity comparison must align product, process, wafer size, utilization, yield, packaging constraints and time period. An annual dollar-market forecast cannot resolve those physical-capacity questions.

4. WHO IS AFFECTED

Tesla (TSLA)

For Tesla, the potential benefit is a combination of procurement resilience and manufacturing economics—not a mechanical share of a large memory-market forecast. Team Group’s shortage horizon through end-2027 strengthens the conceptual value of dependable supply, but the Terafab thesis depends on whether relevant, qualified output arrives while that scarcity still matters. [4]

The investment trade-off is asymmetric across the cycle: internal production could protect operations during a shortage, yet become expensive relative to purchased chips after supply normalizes. External chip sales would need to stand on their own economics rather than serving as an assumed offset to factory costs.

MICRON TECHNOLOGY (MU)

Those figures imply approximately 30.8% sequential revenue growth and 33.5% net-income growth, illustrating how strongly current conditions favor suppliers. [28]

Its demand-over-supply outlook for 2027–2028 extends that favorable backdrop, but investors must distinguish earnings supported by durable demand from earnings supported by unusually favorable pricing. [6]

SanDisk (SNDK)

The contrast between strong reported results and the subsequent equity decline suggests investors are discounting a less profitable future supply environment. The scheduled November 19 eight-for-one split changes share denomination, not that underlying economic question. [31][26]

Apple (AAPL) and automotive buyers

The commitment increase should not be described as a dollar-for-dollar memory expense or attributed entirely to the margin decline. [32]

Reuters reported automotive and retail warnings about sustained near-term price increases and supply disruption on June 3. That makes memory availability relevant beyond data centers: downstream companies face a choice among absorbing costs, passing them through, changing product configurations or securing additional supply. [9]

RATES AND TECHNOLOGY VALUATIONS

The reversal cautions against interpreting broad technology strength as confirmation of a particular memory or Terafab thesis. [1][2]

Higher required returns would make distant factory cash flows less valuable and capital-intensive projects harder to justify, even if near-term shortage conditions remain favorable. [3][2]

5. THE NUMBERS

• GARTNER TOTAL MEMORY REVENUE FORECAST: Latest $837.3B for 2026 · Change About +280.4% versus $220.1B in 2025 · What is driving it Dollar-market expansion; not a physical supply measure. [10]
• AI-SERVER DRAM COSTS: Latest Fourfold full-year 2026 increase expected · Change Roughly doubled in Q1 · What is driving it AI infrastructure demand and tight supply. [7]

• NASDAQ-100 TRACKING ETF, QQQ: Latest $671.26 at 8:39 AM ET · Change −1.19% today · What is driving it Broader technology valuation backdrop. [2]

6. WHAT INVESTORS SHOULD WATCH NEXT

• THIS WEEK: PRODUCT-SPECIFIC PRICING AND CUSTOMER COMMENTARY. Compare AI-server DRAM conditions with NAND’s projected second-half 2027 easing rather than extrapolating one segment’s outlook across all memory. [7][30]
• TESLA’S INTERNAL-USE VERSUS OUTSIDE-SALES ALLOCATION. Evaluate any Terafab disclosure through separate procurement-savings and merchant-revenue models, with qualified output and production timing determining the applicable market.

• CAPACITY COMPARISONS ON A COMMON BASIS. Track wafer starts, yields, utilization and packaged output separately; the decisive question is usable supply, not a headline capacity ranking.
• NOVEMBER 19: SANDISK’S EIGHT-FOR-ONE STOCK SPLIT. Adjust price comparisons consistently while keeping attention on earnings durability and NAND supply conditions. [26][30]

• 2027–2028: COMPETING SHORTAGE FORECASTS. Watch whether NAND easing broadens or whether Micron’s demand-over-supply outlook persists; that divergence will help distinguish product-specific relief from an industry-wide turn. [30][6]
• LATER CAPACITY RAMPS: Compare project execution with Deloitte’s 2029–2030 supply horizon, because delayed output could prolong scarcity while synchronized additions could increase subsequent oversupply risk. [7][8]

7. ANALYST TAKEAWAY

THE DEFENSIBLE RECONCILIATION IS CONDITIONAL BUT CLEAR: future NAND easing and current AI-memory shortages can coexist. Revenue growth does not establish a glut, however, so an August oversupply conclusion drawn from a larger dollar-market forecast requires correction rather than a product-mix explanation. [30][7][10]

In the bull scenario, shortages persist into 2027–2028 and a Terafab operation delivers relevant, qualified output at competitive cost. The signposts would be sustained customer demand, reliable production milestones and internal cost advantages that survive conservative pricing assumptions. The opportunity would first be procurement resilience and avoided costs; outside revenue would be an additional, separately demonstrated business. [4][6]

In the bear scenario, NAND easing spreads, AI investment growth moderates and new capacity arrives before a new entrant achieves competitive scale. TrendForce’s second-half 2027 easing outlook and Dan Ives’s expectation that AI spending growth could peak in 2027 identify plausible demand-and-supply pressures on that scenario. [30][52]

The biggest risk is valuing a long-lived factory using peak shortage prices while counting internal savings as external sales. The biggest opportunity is genuine supply security at an attractive through-cycle cost. A credible Tesla thesis must establish those economics before attaching a share of the global memory market to Terafab.

Prepared for informational and educational purposes only. Not financial advice.

SOURCES

This article is for informational purposes only and is not investment advice. Always do your own research or consult a licensed financial professional before making investment decisions.