Apple Faces Supply Chain Pressures Amid AI-Driven Semiconductor Demand

The semiconductor pressure on AI has been transmitted to consumer electronics, and even Apple, with the world's strongest supply chain management capabilities, cannot remain unaffected.

Written by Mike, Frank, MSX MaiTong

A nearly flawless earnings report resulted in a 10% stock price plunge.

On July 30, Apple reported revenue of $109.42 billion for its third quarter of fiscal year 2026, a 16% year-over-year increase and the strongest June quarter on record; diluted earnings per share reached $2.02, up 29% year-over-year and significantly exceeding the market expectation of $1.89, with an overall gross margin of 50.1%, reaching an all-time high.

Almost all key metrics—from revenue and profits to core product sales—exceeded Wall Street’s expectations. Given that this was Tim Cook’s final earnings call before stepping down as CEO, it should have been a sufficiently impressive farewell.

But the capital markets responded in exactly the opposite way.

After the earnings report, Apple's stock dropped approximately 5.5% in after-hours trading, then fell nearly 10% during the following trading session, erasing nearly $500 billion in market value. The issue was not with the just-ended quarter, but with Apple’s outlook for the future—the management team forecasted a 9% to 11% year-over-year revenue growth for the September quarter, below Wall Street’s consensus expectation of around 12%; gross margin guidance also declined to 47% to 48%.

Cook further emphasized that the company is facing significant supply constraints, with little flexibility remaining in the supply chain—highlighting how the global competition for semiconductor resources driven by AI is now extending from data centers to consumer electronics, and even Apple, with its world-leading supply chain management, cannot remain unaffected.

Understanding this line is the starting point for comprehending Apple’s next-stage issues regarding costs, products, and valuation.

Over the past two years, the market has been accustomed to understanding AI infrastructure through GPUs, optical modules, networking equipment, and power systems, but has relatively overlooked another rapid development: AI data centers are not only consuming massive computing power but are also rapidly absorbing the world’s highest-quality memory production capacity.

Large AI accelerators require substantial amounts of HBM; training and inference servers also need server DRAM, enterprise-grade SSDs, and larger-scale data storage. Due to higher stacking layers and more complex manufacturing processes, HBM consumes significantly more wafers and manufacturing resources per unit than standard memory.

This does not mean that HBM directly shares the same final packaging line as LPDDR used in smartphones, but upstream memory manufacturers will reallocate wafer capacity, capital expenditures, and engineering resources based on profitability and customer certainty.

As a result, as more resources flow toward HBM, server DRAM, and enterprise storage, the supply of traditional DRAM and LPDDR for smartphones, PCs, and other consumer electronics is naturally constrained. According to industry estimates from TrendForce, AI-related memory could absorb nearly 20% of global DRAM capacity by 2026, measured in equivalent wafer consumption.

Therefore, the notion that "AI took away iPhone's memory" doesn't mean NVIDIA directly took a specific batch of LPDDR that Apple had already ordered; rather, AI customers have reshaped the entire memory industry's resource allocation through higher profitability, longer procurement cycles, and stronger prepayment capabilities.

In this reallocation of resources, the outcomes faced by different segments of the industrial chain are vastly different.

1. Memory manufacturers: Moving from cyclical price increases to resource scarcity

The most direct beneficiaries are DRAM manufacturers such as SK Hynix, Samsung Electronics, and Micron.

AI customers have offered higher product prices, longer order visibility, and stronger prepayment capabilities for HBM; meanwhile, memory manufacturers have shifted more production capacity and investment resources toward AI products, further tightening the supply of traditional DRAM and LPDDR.

Micron's revenue for its most recent fiscal quarter reached approximately $41.46 billion, a nearly 350% year-over-year increase, more than quadrupling the same period last year; this growth was driven by multiple product lines, including HBM, DRAM, and NAND.

As a result, memory manufacturers are benefiting from two sources of growth: one end is the rapid increase in high-value products such as HBM and server DRAM; the other is the tightening supply of memory for consumer electronics, driving price increases for traditional products.

Micron's revenue for its most recent fiscal quarter reached approximately $41.46 billion, a nearly 350% year-over-year increase, more than quadrupling the same period last year, driven by growth across multiple product lines including HBM, DRAM, and NAND.

This is precisely what makes the current storage cycle so unique. Past storage bull markets typically relied on a rebound in demand from traditional endpoints like smartphones and PCs, but this rally is being driven simultaneously by expanding AI demand and shrinking traditional capacity.

Data centers are continuously increasing the memory configuration per server, while consumer electronics manufacturers must pay higher prices for limited remaining production capacity. Simultaneous demand expansion and supply contraction have made storage manufacturers' profit elasticity significantly higher than in a typical cyclical recovery.

Of course, distinctions still exist among different storage companies. Micron, SK Hynix, and Samsung are direct participants in DRAM, LPDDR, and HBM logic; whereas NAND manufacturers like SanDisk benefit more from enterprise SSDs, data storage demand, and rising NAND prices.

They all belong to the mainline storage segment, but they do not share the same business logic and cannot be simply regarded as direct beneficiaries of the Apple LPDDR shortage.

2. Apple and downstream suppliers: Who ultimately bears the cost pressure?

For Apple, a memory shortage means three possible responses: raising prices, reducing profit margins, or adjusting the shipment mix across products and configurations.

Theoretically, as component costs continue to rise, Apple is more likely to prioritize high-margin, high-priced products by allocating limited supply toward the Pro series and higher-capacity models.

However, it should be emphasized that there is currently insufficient official information to confirm that Apple has significantly reduced production of any base model by one-third due to memory shortages. While some supply chain reports mention adjustments to orders for the standard iPhone 17, these may also reflect normal product cycle transitions, changes in demand, or inventory clearance ahead of a new model launch.

Therefore, a more prudent way to observe is to determine which layer of the value chain a supplier occupies:

• Assemblers, PCB manufacturers, connectors, and general component suppliers that rely on overall shipment volumes and low-margin orders are more sensitive to Apple’s production cuts;
• Suppliers that control scarce technologies such as image sensors, high-end displays, advanced packaging, and core chips have relatively stronger bargaining power;

However, greater bargaining power does not mean complete immunity.

If overall Apple shipments decline, nearly every link in the supply chain will be affected, with differences only in the extent of order reductions and whether the increased value per unit can offset the drop in shipment volume.

Guo Mingze predicts that due to LPDDR supply constraints, Apple’s actual early ordering volume for the A20 chip from the second half of 2026 to the first quarter of 2027 may be 10% to 20% lower than originally planned. However, some of this discrepancy may also stem from Apple’s prior overbooking to secure production capacity.

Apple is also evaluating additional memory suppliers, bringing CXMT into consideration. However, this path is still influenced by factors such as product validation, supply scale, and U.S. regulatory policies, making it difficult to fully replace Samsung, SK Hynix, and Micron in the short term.

3. TSMC: The true toll booth spanning both demands

TSMC is the most uniquely positioned company in this resource reassessment.

Apple relies on TSMC to produce its A-series and M-series chips, while NVIDIA, AMD, and cloud providers' custom AI chips also depend on TSMC’s advanced manufacturing processes and packaging capabilities. Therefore, regardless of whether capital flows into consumer electronics or AI data centers, TSMC stands to benefit.

But Apple and NVIDIA are not simply competing for exactly the same production lines.

Apple's primary constraints are focused on advanced wafer processes such as N3 and N2; AI accelerators, in addition to requiring advanced wafer processes, heavily depend on advanced packaging capacity like CoWoS and supporting HBM.

The bottlenecks of both overlap but are not entirely identical.

What truly matters is that strong demand for AI and consumer electronics has elevated TSMC’s advanced processes, packaging capabilities, and customer scheduling value. In the second quarter of 2026, TSMC’s revenue reached $40.2 billion, a 33.7% year-over-year increase, with a gross margin of 67.7%; the 2-nanometer process already contributes approximately 3% of wafer revenue.

TSMC benefits when Apple’s product cycle is strong, and it also benefits as demand for AI chips continues to grow.

When both occur simultaneously, TSMC gains not only improved capacity utilization but also an opportunity to reprice the scarcity of advanced manufacturing.

Understanding how industry resources are reallocated makes it easier to see why Apple’s stronger earnings report is increasingly causing market concern.

After all, on the surface, Apple's profitability in the third quarter was almost anomalously strong.

The overall gross margin reached 50.1%, an increase from the previous quarter, with approximately two percentage points attributable to tariff refunds from the U.S. government. These tariff refunds contributed about $0.11 in earnings per share. Excluding this one-time factor, Apple’s actual gross margin was approximately 48.1%.

This figure is still not bad, but far from the impressive 50.1% it appears to be.

In other words, the high gross margin in this earnings report does not fully reflect a structural improvement in Apple’s product mix or pricing power; part of the profit is merely a temporary reversal of previously incurred tariff costs, while upward pressure from rising prices of memory, chips, and other key components continues to accumulate.

Apple has previously shifted part of its costs to consumers by raising prices on certain Mac and iPad products; however, price adjustments for the iPhone, as its highest-volume and most competitively pressured core product, are clearly more sensitive.

This also explains why stronger iPhone sales this quarter have made investors more concerned about the next quarter.

Amid ongoing shortages of memory and advanced chips, some consumers may purchase existing products before prices rise, meaning the strong demand in the third quarter could simultaneously reflect normal upgrade cycles, product cycle benefits, and preemptive buying driven by expectations of price increases.

If future product prices increase, Apple needs to verify whether consumers are still willing to maintain the same upgrade pace; if prices remain unchanged, Apple must absorb the higher component costs on its own or protect profits by adjusting product structure, configurations, and shipment volumes.

Regardless of the path chosen, Apple will face a situation uncommon in the past: the rate of increase in supply chain costs is approaching the limits of what Apple’s traditional pricing, inventory, and product mix tools can absorb.

More notably, the "supply constraints" in the earnings report do not come solely from memory.

Cook stated that one of the main bottlenecks in the just-concluded third quarter was insufficient capacity for the advanced processes required to produce Apple Silicon, particularly affecting the supply of products like the Mac. Meanwhile, Apple expects memory costs to continue rising in the next quarter.

Therefore, Apple is currently facing two distinct yet simultaneous shortages: on one side, rising prices for memory such as DRAM and LPDDR are directly increasing device material costs; on the other side, constrained capacity in advanced manufacturing processes is limiting how many A-series and M-series chips Apple can produce.

Cook described the current memory market as a "hundred-year flood" unlike anything he has seen in his career, and Apple's response has been to build inventory ahead of time. As of the end of June 2026, Apple's inventory reached $11.09 billion, nearly doubling from $5.72 billion at the end of fiscal year 2025. Component inventory rose from approximately $2.12 billion to $7.65 billion, indicating that Apple is working to secure critical components as early as possible.

However, inventory can only smooth costs; it cannot create new capacity.

For a company already valued at a historical high, this state—where products can be sold but may not be producible, and production may be possible but not at previous profit levels—is sufficient to prompt the market to reprice the company.

It is worth noting that just before the earnings plunge, Apple had just completed a highly symbolic market capitalization surpass.

On July 28, Apple's market capitalization briefly surpassed $5 trillion, becoming the second company globally to reach this milestone after NVIDIA. As of that day, Apple had gained approximately 25% year-to-date and briefly reclaimed its position as the world's most valuable company.

Apple's rise does not mean the market believes it already possesses the strongest large model.

On the contrary, Apple has long been considered behind OpenAI, Google, and Anthropic in terms of foundational model capabilities, cloud computing resources, and the speed of AI product releases.

The market repricing of Apple reflects a shift in the AI narrative.

In the first phase of AI investment, the market rewards model capabilities and technological breakthroughs; in the second phase, it rewards infrastructure providers such as GPUs, networks, memory, and data centers; as capital expenditures continue to grow, the question in the third phase becomes: how much revenue, profit, and free cash flow will this computing power ultimately generate?

At this stage, Apple’s advantage is not having the model with the most parameters, but rather owning one of the world’s largest entry points to high-value consumer devices.

By early 2026, the number of active Apple devices installed exceeded 2.5 billion; at the same time, the number of paid subscriptions for Apple’s services surpassed 1.5 billion, enabling Apple to integrate AI capabilities into phones, computers, headphones, watches, and operating systems, and monetize them through hardware upgrades, iCloud+ subscriptions, and its service ecosystem.

It doesn’t need to prove, like cloud computing giants, that hundreds of billions of dollars in data center investments can generate sufficiently high returns on capital—it only needs to demonstrate that AI can increase device upgrade rates, service subscription rates, and user retention.

This “device distribution + subscription service” model is precisely why Apple has been revalued amid growing controversy over its AI capital expenditures. For the first nine months of fiscal year 2026, Apple generated $116.996 billion in operating cash flow, while capital expenditures on fixed assets amounted to approximately $6.799 billion. Apple sustains its vast ecosystem of devices and services with less than 6% of its operating cash flow allocated to fixed asset investments.

Compared to tech giants that heavily invest in building AI data centers, this business model is significantly more asset-light.

However, "light asset" does not mean there is no cost. Apple is currently adopting a hybrid AI approach: its own models run on local devices and Private Cloud Compute, while some more complex Siri capabilities are powered by Google’s Gemini technology.

This approach reduces early infrastructure investment but means Apple must rely on external partners for core model capabilities, inference costs, and product timelines.

If Siri AI usage grows rapidly in the future, Apple may still face higher cloud inference costs and be forced to increase its investment in computing power. Apple may have skipped the most aggressive capital expenditure race in the first phase, but it has not permanently escaped the costs of AI infrastructure.

A more realistic issue is that Apple's AI monetization has yet to be proven by financial data.

Service revenue for the quarter increased 12.1% year-over-year to $30.74 billion, but fell short of the market expectation of $31.22 billion. App Store gaming revenue was also impacted by regulatory changes and the opening of external payment channels. Apple Intelligence in China is still awaiting regulatory approval; in the EU, the new Siri AI will not be fully available on iPhone, iPad, and Apple Watch initially.

This means Apple has a large device ecosystem but is currently unable to release its AI capabilities simultaneously across all key markets.

The post-earnings analyst分歧 stems precisely from this.

Bullish investors believe Apple possesses user access, a hardware ecosystem, and cash flow capabilities that other AI companies struggle to replicate, and that the next iPhone cycle will mark the beginning of AI monetization. Bearish investors argue that Apple’s current valuation has already priced in AI-driven device upgrades and subscription growth, yet these revenues have not yet materialized.

After the earnings report, at least four institutions lowered their price targets for Apple, while three raised theirs, bringing the median market price target to approximately $330.

What Apple needs to answer next is no longer just “How many units of the next iPhone can we sell?”, but two deeper questions:

• First, can Apple maintain its sales volume, pricing, and profit margins amid ongoing increases in memory and advanced process costs?
• Second, can Apple Intelligence and Siri AI truly translate product features into device upgrade demand, subscription revenue, and higher user lifetime value?

Over the past two years, the dominant narrative in the AI industry chain has been NVIDIA selling more GPUs, cloud providers building more data centers, and memory manufacturers achieving higher prices.

Apple's earnings report reveals another side of this prosperity.

As AI infrastructure absorbs increasing amounts of memory, advanced manufacturing processes, and engineering resources, consumer electronics companies face higher costs, suppliers must reallocate orders, and consumers may encounter higher prices.

Apple stands at both the end of pressure and the end of opportunity.

Its hardware business is under pressure from AI’s strain on global semiconductor resources, while its services and ecosystem business may become one of the most important distribution channels as AI transitions from infrastructure investment to commercialization.

In the future, will the most valuable company be the one with the most computing power, or the one best able to turn computing power into user payments?

Apple is trying to become the latter.

But at a valuation of over $5 trillion, the market will no longer pay for a story that hasn't yet been realized.