← Back to list
Baillie GiffordDeep research14 Aug 2026Source: bailliegifford.com

The market’s short memory

Baillie Gifford is an Edinburgh investment partnership founded in 1908, famous for ultra-long-horizon, high-conviction growth investing — its early stakes in Amazon, Tesla and NIO are classics. Its "actual investors" philosophy holds world-changing companies on 5-10 year views; AUM is around $120bn. The Insights column carries its managers' investment views and thematic research.

多位合伙人 · 1908 · 英国爱丁堡Long-term growth / Global

In plain words

This report says memory chip companies like Micron are making huge profits now (revenue $41.5B, 85% gross margin), but the author thinks it's just a temporary shortage. Once new factories come online, prices will crash—it's happened three times before. So the author doesn't own Micron and instead bought more Arm and KLA during July's chip stock dip. Arm designs chips for phones and data centers, benefiting from AI growth; KLA makes inspection equipment for chip manufacturing, becoming more essential as chips get complex. The author believes both have durable growth, not reliant on scarcity.

AI SummaryAI-generated · may contain errors · verify against the original

At a Glance

One-sentence summary: The author is bearish on memory cycle stocks (e.g., Micron), arguing that the current AI-driven shortage will eventually be overtaken by supply, making pricing power unsustainable; meanwhile, the author is bullish on the structural growth opportunities of Arm and KLA. [Cautious/Bearish]

  • Micron's latest quarterly revenue was $41.5 billion (up 346% year-over-year), with a gross margin of 85%, but the author believes this remains a short-term phenomenon driven by supply shortages rather than a structural change.
  • Three historical cycles (1993-96, 2017-19, 2021-23) demonstrate that even with strong demand and industry consolidation, prices collapse once supply catches up.
  • Under the current AI investment boom, new capacity is expected to translate into supply by 2027-29, and the market may price this in ahead of time.
  • The author explicitly does not hold Micron and bought Arm and KLA during the sector's weakness in July, believing both have more durable demand that does not rely on scarcity.
~7 min full read · 6 sections
Deep Analysis

History Has Repeatedly Shown: Memory Shortages Are Eventually Met by Supply, Making Pricing Power Unsustainable

Micron’s latest quarterly revenue of $41.5 billion (up 346% year-over-year) and gross margin of 85% are fundamentally still driven by supply shortages, not structural change. The author notes that while HBM (high-bandwidth memory) is more complex than traditional DRAM and sold through multi-year contracts, "it is still ultimately built on the same underlying DRAM manufacturing base." In other words, once more suppliers gain customer certification, yields improve, and packaging capacity expands, customers will have alternative options, and economic laws will inevitably lead to competition.

The article uses three historical cycles to support this view:

Cycle Demand Driver Outcome After Supply Catches Up
1993-96 PC adoption boom In 1996, Micron’s shipments rose 77%, but revenue per bit fell about 45%, and total revenue declined slightly
2017-19 Cloud computing, smartphones, gaming In 2019, DRAM prices fell about 30%, with revenue and margins dropping sharply
2021-23 Pandemic-era PC/consumer electronics/cloud demand DRAM prices fell nearly 50%, NAND fell over 50%, gross margins turned negative, and large inventory write-downs occurred

The author summarizes the lesson: "all the bullish arguments could be true: secular demand, industry consolidation and better discipline…and yet prices could still collapse when supply catches up." That is, all bullish arguments—secular demand, industry consolidation, and more disciplined capacity management—may hold, but prices can still collapse when supply catches up with demand.

The Current AI Investment Frenzy Is Repeating the Historical Script, with a Supply Shock Expected Around 2027-29

Micron, Samsung, and SK Hynix are pouring massive capital into new fabs, packaging capacity, and manufacturing improvements, with the best estimate for these investments to translate into usable supply around 2027-29. The author emphasizes that predicting the exact timing is "little better than conjecture," but the market may begin discounting this years before supply actually comes online—Micron’s stock pullback in early July could be an early signal, or merely volatility.

HBM supply may be tighter than traditional products because it is harder to manufacture and consumes more fab capacity. However, the author concludes: "we are less convinced it is exempt from DRAM economics in the long run." That is, the author is not convinced HBM can escape the economic laws of DRAM over the long term.

Investment Implications

The author explicitly states they do not hold Micron ("we don't own Micron in US Growth") and has never considered memory companies to be high-quality long-term growth investments. The article reflects an institutional bias: as a long-term growth investor, Baillie Gifford naturally favors semiconductor themes with more durable competitive advantages, such as Arm and KLA, over the cyclical swings of memory companies. Readers should note that this view is partly shaped by the firm’s investment style—a systematic avoidance of cyclical stocks—rather than a denial of AI demand itself.


The author's bullishness on Arm and KLA is based on demand durability, not memory company scarcity

The article argues that Arm and KLA have far stronger long-term earnings potential than Micron, as they do not rely on temporary shortages and extreme price hikes. The author viewed these two companies as buying opportunities during the semiconductor sector's weakness in July. The author's original statement: "Arm and KLA, two businesses where we think the duration of demand and profitability is potentially much stronger than Micron." This means: "Arm and KLA, two companies where we believe the duration of demand and profitability could be far stronger than Micron."

  • Arm: Provides processor designs for nearly all smartphones and an increasing number of data centers. As AI workloads grow, energy efficiency becomes more critical; simultaneously, increasingly complex AI systems require more processors to coordinate models, memory, tools, and other tasks. Arm also benefits from a vast software and developer ecosystem, where replacement costs are high and disruptive.
  • KLA: Manufactures inspection equipment used to detect tiny defects during semiconductor production. As chips become larger, more expensive, and more complex—especially with more chips stacked and packaged together—a single defect can destroy more value, making inspection increasingly important.

The article notes that both companies experience semiconductor cycles, but their long-term economics rely less on temporary shortages and extreme price hikes. Arm benefits from broader adoption of its architecture; KLA benefits from increasing manufacturing complexity. The author concludes: "We believe they both have durable growth opportunities and aren’t reliant on scarcity as a moat." This means: "We believe they both have durable growth opportunities and do not rely on scarcity as a moat."

Investment Implications

The article explicitly advises investors to avoid memory companies that rely on scarcity pricing (such as Micron) and instead turn to companies like Arm and KLA, which have structural growth drivers. Readers should note that this is a position-holder's perspective—Baillie Gifford bought Arm and KLA in July, and its analysis may be biased in favor of supporting its own holdings.


Position Moves

Ticker Direction Author's One-Sentence View Key Data
Micron Not disclosed (not held) Bearish: Memory cyclical stock, pricing power unsustainable, long-term earnings potential weak Latest quarterly revenue $41.5 billion (up 346% YoY), gross margin 85%; historically prices have fallen 30%-50% during cycles
Arm Added Bullish: Durable demand, AI workload growth driving energy efficiency needs, strong ecosystem moat Viewed as a buying opportunity during the July semiconductor sector weakness
KLA Added Bullish: Increasing chip manufacturing complexity driving inspection demand, not reliant on scarcity Viewed as a buying opportunity during the July semiconductor sector weakness