← Back to list
Scottish Mortgage (Baillie Gifford)Podcast25 Sep 2026Source: scottishmortgage.com

Ask the Managers

Scottish Mortgage is Baillie Gifford's flagship investment trust (founded 1909, LSE ticker SMT), known for its maximalist growth style — long-term stakes in Tesla, Amazon and ASML plus bold allocations to private companies like SpaceX and ByteDance. It is the UK retail investor's flagship vehicle for global disruptive growth.

Tom Slater、Lawrence Burns · 1909 · 英国爱丁堡Aggressive growth / Public & private

In plain words

This covers fund managers Tom Slater and Lawrence Burns answering investor questions, focusing on growth beyond AI: SpaceX, healthcare, luxury goods, and quantum computing. They're bullish on SpaceX (over 15% of the fund), arguing risk is permanent capital loss, not volatility, and lower launch costs unlock new businesses. They call NVIDIA and other chipmakers a 'royalty' on AI (they profit regardless of which model wins). Ferrari and Hermès are seen as lasting scarce assets. Moderna's mRNA tech shows promise in cancer treatment.

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

At a Glance In a Q&A session, Scottish Mortgage fund managers Tom Slater and Lawrence Burns discussed long-term investment opportunities beyond AI. SpaceX remains the largest holding, accounting for over 15% of the portfolio. Its success stems from a sharp reduction in launch costs: before SpaceX, t

~11 min full read · 10 sections
Deep Analysis

This Issue at a Glance

Tom Slater (Fund Manager) and Lawrence Burns (Fund Manager) responded to investor questions during the annual digital conference Q&A session, with the core theme being: Beyond AI, SpaceX, healthcare, luxury goods, and quantum computing form diversified growth sources. The most impactful judgment in the entire session came from Tom Slater — "A large position does not equal high risk; risk is permanent capital loss, not volatility" — thereby justifying the rationality of holding over 15% in SpaceX.


1. SpaceX: Lower Launch Costs Unlock Multiple Businesses; High Valuation but Long-Term Returns Achievable

Tom Slater believes that SpaceX’s success stems from an investment philosophy of "letting winners fully contribute returns." The fund currently holds a position slightly above 15%, having gradually reduced concentration through multiple selling windows, but it will not sell early for the sake of portfolio balance. "We do not equate large positions with risk; risk is permanent capital loss," meaning volatility is inevitable, and the key is to fully realize returns on winners.

Lawrence Burns adds that SpaceX’s core logic lies in the continuous decline of launch costs:

  • Cost to orbit: Before SpaceX, $18,000–$19,000 per kilogram → Falcon 9 list price around $2,700 → Starship target of $100–$200 (currently around $900, still in testing)
  • Starlink has proven its business model: annualized revenue of $17 billion, growing nearly 70% year-over-year, with an operating margin near 40%
  • Lower costs will unlock "multiple Starlink-scale businesses": orbital data centers, defense applications (e.g., Starshield missile defense), microgravity manufacturing

Valuation assessment: Current profit and revenue multiples are indeed high, but if multiple Starlink-scale businesses can be built, the next decade could generate strong long-term returns for shareholders.


2. AI Value Capture: "Royalties" at the Chip Layer + "Amplifiers" at the Organization Layer

Lawrence Burns divides AI value capture into two layers:

First Layer—Chip Layer (Highest Certainty): Companies such as NVIDIA, TSMC, ASML, and SK Hynix hold supply chain bottlenecks. Regardless of which AI model (Anthropic, Gemini, OpenAI) or application (autonomous driving, robotics, coding agents) prevails, they benefit. "These companies are royalties on AI development," meaning they collect revenue no matter who wins downstream. The consensus among all AI companies is "more chips are needed."

Second Layer—Organization Layer (Source of Differentiation): Founder leadership (e.g., Shopify's Toby Lütke's early focus on AI), a culture of dynamic adaptation, the ability to attract top talent, and proprietary data (e.g., MercadoLibre using transaction data for advertising and credit assessment) are key. AI is not a democratizing tool but an amplifier for outstanding organizations—in regulated industries, competitors' approval decisions take two years, while agile organizations can iterate quickly, creating a sustained advantage.

Tom Slater adds demand-side signals: The token consumption of OpenAI's largest user group is eight times that of medium users, and the gap is widening; coding use cases have already demonstrated clear economic value; and the shift from conversational models to autonomous agent models will only drive token consumption higher.


3. AI Investment Returns: Positive Returns on Both User and Capital Sides

Lawrence Burns observes a shift in corporate attitudes: 18–24 months ago, companies were "experimenting" with AI; over the past 12 months, they shifted to "believing it will have a material impact"; and more recently, they have begun "restructuring entire organizations around AI models." His personal token usage has grown exponentially (as model capabilities improve faster than usage time increases), and he believes current returns are extremely favorable—"This is the worst AI we will ever have, and it will only get better."

Tom Slater provides data from the capital side: the computing facility built by xAI (a subsidiary of SpaceX), along with two major deals signed with Google and Anthropic, generate annualized revenue approximately 2–3 times the capital invested. "The returns on current hardware investments are outstanding," meaning this explains why capital continues to pour in.


4. Healthcare: Enormous Technological Potential, but Slow Adoption Requires Careful Selection

Tom Slater believes healthcare is an area with vast potential for technological application, yet historically one of the slowest industries to adopt new technology. The fund has positioned itself through a few select companies such as Moderna (mRNA technology) and Indivior (AI screening of natural compounds):

  • Moderna: Beyond the COVID-19 vaccine, mRNA technology has achieved breakthroughs in flu vaccines and cancer treatments (positive clinical trial results for melanoma), serving as a typical example of "programmable technology"
  • A realistic perspective is needed: most healthcare technology trends are slow to materialize, and only a few companies can truly create patient value and shareholder returns

5. Ferrari and Hermès: Low Correlation, Scarcity, and Enduring Alternative Growth

Lawrence Burns explains why he holds "slow-growth" companies like Ferrari and Hermès:

  • Ferrari has ranked among the top ten in returns over the past decade, driven by modest volume growth + pricing power + margin expansion + market premium for quality perception
  • Low correlation with other tech themes in the portfolio, building diversified sources of growth
  • "In a world of technological change and material abundance, things with scarcity and physical attributes may hold more lasting value", meaning these companies will still exist 100 years from now, while tech companies remain uncertain

6. Quantum Computing: Early Learning Phase, Small Position for Observation

Tom Slater explains that the fund holds quantum computing exposure through PsiQuantum. At this stage, "the investment is in the technology, not the business," with the aim of learning about the team and understanding the probability of success. "A traditional computer is a box of accountants; a quantum computer is a box of toddlers learning to walk," meaning it can solve an entirely new class of problems (such as atomic-level simulations and cellular-level modeling). Once successful, the returns could be enormous, but for now, it remains a small position.


7. AI's Impact on Employment: The Entrepreneur's Perspective is "Human Shortage," Not "Human Surplus"

Tom Slater shares a key insight from conversations with founders: while the media worries about AI causing job losses, what founders actually fear is a shortage of human labor. "There is no upper limit to economic opportunities and consumer demand; the more tools available, the more people are needed," meaning that as AI improves efficiency, it will generate even more new opportunities. The likely outcome is that large companies will have fewer employees, but more small companies will emerge (because small teams can accomplish more with AI).

Lawrence Burns adds: AI is an "amplifier for well-organized organizations," not a "fair democratizing tool." In regulated industries, agile organizations can outpace competitors by several "AI generations" (due to approval cycles lasting up to two years), making active stock selection more valuable.


Mentioned Positions

Position Guest Stance Key Data
SpaceX Bullish (largest holding, >15%) Launch cost dropped from $18,000–19,000 to $2,700 for Falcon 9, Starship target $100–200; Starlink annualized revenue $17B, growth ~70%, margin ~40%
Moderna Bullish (long-term hold) Positive clinical results for mRNA technology in melanoma treatment; demand shrinking post-COVID but new indications advancing
Indivior Bullish (AI application) Using AI to screen natural compounds for medical use
NVIDIA Bullish (AI chip layer "royalty") Supply chain bottlenecks; all AI models and applications benefit
TSMC Bullish (same as above) Same as above
ASML Bullish (same as above) Same as above
SK Hynix Bullish (same as above) Same as above
Shopify Bullish (founder leadership) Toby Lütke focused on AI early on
MercadoLibre Bullish (proprietary data) Using transaction data for advertising and credit assessment
Ferrari Bullish (alternative growth) Top 10 returns over the past decade; volume growth + pricing power + margin expansion
Hermès Bullish (same as above) Brand scarcity, expected to endure for a century
PsiQuantum Early observation (small position) Quantum computing, strong academic background, targeting commercial-grade quantum computers
xAI Bullish (capital return) Annualized revenue from computing facilities is roughly 2–3 times capital investment
Google Neutral (customer/supplier role) Signed a large computing facility deal with xAI
Anthropic Neutral (customer/supplier role) Signed a large computing facility deal with xAI
Meta Neutral (capital deployer) Capable of investing in AI infrastructure
Amazon Neutral (capital deployer) Same as above

Judgments Worth Remembering

1. Tom Slater: "A large position is not high risk; risk is permanent capital loss, not volatility" — A SpaceX holding exceeding 15% is the result of letting winners fully contribute to returns, not a risk signal.

2. Lawrence Burns: "Chip companies are royalties on AI development" — NVIDIA, TSMC, ASML, and SK Hynix hold supply chain bottlenecks; regardless of which model or application wins, they benefit.

3. Tom Slater: "What founders worry about is not a surplus of humans, but a shortage of humans" — Economic opportunities have no ceiling; as AI improves efficiency, it will spawn even more new opportunities. The likely outcome is fewer employees at large companies but a surge of small ones.

4. Lawrence Burns: "This is the worst AI we will ever have; it will only get better" — Model capabilities continue to improve, current token usage is growing exponentially, and positive returns are seen on both the user and capital sides.

5. Lawrence Burns: "AI is not a democratizing tool; it is an amplifier for excellent organizations" — In regulated industries, agile organizations can outpace competitors by multiple "AI generations," making active stock selection more valuable.

6. Lawrence Burns: "In a world of technological change and material abundance, things with scarcity and physical attributes may hold enduring value" — Ferrari and Hermès are expected to survive for a century, have low correlation with tech themes, and provide diversified growth sources for the portfolio.

7. Tom Slater: "A traditional computer is a box of accountants; a quantum computer is a box of toddlers" — Quantum computing can solve entirely new problems, such as atomic-level simulations and cellular-level modeling. Once successful, the returns are enormous, but it remains a small-position learning phase for now.

8. Lawrence Burns: "SpaceX is essentially a near-monopoly gateway to the rest of the universe" — As launch costs drop from $18,000–19,000 to $100–200, it will unlock multiple Starlink-scale businesses (orbital data centers, defense, microgravity manufacturing).