Peter Singlehurst on Growth Investing, Mistakes and Private-Market Liquidity

Peter Singlehurst, head of private companies at Baillie Gifford, joins Harry Stebbings to discuss growth investing, mistakes, liquidity and the value of long-term ownership. His views reflect decisions made for clients and interests in several companies discussed.

Singlehurst began looking at private companies in 2014 while working on a large public-equity strategy. As companies such as Airbnb and Spotify stayed private longer, he volunteered to examine opportunities that resembled the growth businesses the firm already studied.

In hindsight, he says he was too committed to using permanent-capital vehicles alone. Those structures supported long horizons, but some clients could not use them; he now believes the firm can serve clients through both permanent and more traditional fund structures.

He distinguishes a poor outcome from an investment mistake. An uncertain event can go against a well-reasoned investment, while a mistake involves information or risk that an investor should have assessed differently.

He gives a biotech investment as an example of a known regulatory risk that materialized after an FDA rejection, causing a loss. Northvolt, by contrast, is a case where he believes Baillie Gifford overestimated the team’s execution while being attracted to the idea of European energy sovereignty.

The firm made additional Northvolt investments after its initial check, but later declined further financing. Singlehurst cites concerns about execution and financing structure; his retrospective assessment should not be mistaken for a complete account of all causes or losses.

Today the team focuses on growth-stage businesses where the product has already attracted customers. Its principal questions concern the quality and scalability of the business model rather than the earliest product-development risk.

Singlehurst says the median company at initial investment has around $200 million in revenue, approximately 70% year-on-year growth and a negative 14% EBITDA margin. These are portfolio medians, not minimum requirements or results for every holding.

He cites Wise, which he says was smaller at entry and later became a multibillion-dollar revenue business. The investment case depended not only on market size but also on a model that could scale and eventually earn a high return on equity; Baillie Gifford has an investment interest.

Return on equity matters more in his growth-stage framework than it may in the earliest venture rounds. He worries that providing companies with too much capital can weaken the economics of business formation even when growth initially appears impressive.

He sees a difference between crowded, expensive AI financings and less fashionable areas where businesses may be improving profitability. As a global generalist, he says the team can look beyond the current hot sector for quality at a price it considers reasonable.

Baillie Gifford owns shares in companies such as Databricks and Tenstorrent, but Singlehurst says it has not invested in the major large-language-model companies. The team is still trying to understand what will provide enduring advantage at the model layer.

Discipline does not mean refusing to pay a high multiple under any circumstances. Singlehurst says a truly exceptional company may justify a high price, but investors must avoid telling themselves every company is exceptional.

His analysis combines growth, enduring competitive advantage, culture and financial returns. He looks for alignment between an organization’s culture and its mission, and compares the price paid with a long-term estimate of intrinsic value.

He argues that a durable advantage may lie in processes and culture rather than a single product feature. He cites portfolio company Bending Spoons’ acquisition and integration approach as an example, while acknowledging that even such advantages can erode.

Singlehurst estimates that most of the firm’s largest investments remain founder-led, partly because of selection and the stage at which it invests. He also identifies Vinted as an example of a successful business led by a non-founder, so the observation is not proof of a universal rule.

The team models a fivefold upside case consistently across companies and then evaluates the assumptions and probability required to reach it. A modeled 5x is a comparison tool, not an expected or guaranteed realized return.

He says a 30% to 50% estimated probability of a fivefold outcome could be attractive, contrasting it with his rough historical base-rate estimate. These are subjective probabilities and do not establish the future return of any investment.

Some Baillie Gifford vehicles can recycle capital by trimming public holdings and investing in new private companies. Other funds have more traditional fixed lives; Singlehurst says opportunity cost and the mandate of each vehicle influence the decision.

He says a fund closed in 2021 deployed relatively little in 2022 and 2023 because the team considered valuations too high, then invested more in 2024. Stebbings challenges whether pricing had improved for the most sought-after outliers.

Singlehurst agrees that capital can herd into a small number of fashionable companies. He argues that a broad geographic and sector universe helps the team seek opportunities elsewhere, while still examining hot areas when the price and quality make sense.

Investing outside the biggest markets can involve additional macroeconomic and liquidity risk. His approach is not to assume those risks disappear, but to seek a price that compensates clients for taking them.

The firm considers future capital needs and dilution. Singlehurst argues that growth and current profitability should not be treated as a simple either-or choice: the long-run return on invested capital and equity are important to the analysis.

He names ByteDance as an investment he believes is misunderstood. Baillie Gifford holds the company, and he argues that its Chinese advertising and commerce businesses are central to the case, independent of the uncertain future of TikTok in the United States.

Singlehurst says the firm’s base case assumes a US TikTok ban yet models a path to at least five times its money. This is his prospective scenario for an existing investment, not a realized return or a guarantee; he acknowledges that TikTok’s outcome could still affect the case.

He expects ByteDance could eventually list and notes the possibility of company share buybacks. Timing, venue and liquidity remain uncertain, and his comments do not constitute an announced listing plan.

On why companies stay private, Singlehurst emphasizes management focus and the obligations of public reporting, not merely the availability of capital. He also recognizes reasons to list, including liquidity, an acquisition currency and transparency for regulators.

He expects company-facilitated secondary rounds to become more common, while doubting that a single exchange for private shares will easily handle varied share classes and transfer rights. He also raises private dividends as a possible future source of liquidity, not an established outcome.

The firm invested in Databricks at an earlier valuation and participated pro rata in the later round discussed at $60 billion. Singlehurst distinguishes maintaining an existing ownership position with a small check from doubling down on the company.

As companies remain private longer, investors accustomed to public markets have entered growth rounds alongside venture firms. Singlehurst hopes these owners develop the expertise to evaluate companies at scale and contribute to more disciplined pricing.

Baillie Gifford invested in Anduril, and Singlehurst compares its combination of hardware, software and large markets to patterns he saw in Tesla and SpaceX. This is the assessment of an investor in the company, not proof that Anduril will achieve those companies’ outcomes.

He says the analytical core is similar across private and public markets: identify companies that could grow many times larger and earn strong returns on equity. Sourcing, information access and the obligations of a good owner differ by market and stage.

Singlehurst worries that deglobalization may reduce opportunities and weaken international ties. He continues to see opportunity in China while acknowledging additional risks and says investors need adequate compensation for those risks.

The private team’s process starts with a broad funnel. He says it met about 1,000 companies, examined 600 financings, completed 65 initial analyses and 30 deeper reviews, and made 11 new investments in the year discussed.

Written investment cases culminate in ten questions and a team discussion. A four-person investment committee decides for the dedicated funds; Singlehurst says initial check sizes ranged from about $10 million to $150 million across different capital pools.

For reinvestments, the team revisits the original case and the fivefold upside model. A meaningful increase in exposure requires stronger upside conviction, while a small pro-rata check can also reflect the responsibilities of being an existing investor.

He says the firm invested in Stripe around a $30 billion valuation but did not participate in a subsequent down round, which he now thinks it should have joined. He recalls questions at the time about growth and the broader software offering; his recollection of round timing and valuation is approximate.

On the London market, Singlehurst sees both a supply challenge and a demand challenge for growth companies. Stebbings disagrees with his assessment of supply; they discuss the possibility of a broader European public market.

Selling or trimming a holding is an opportunity-cost decision. In funds that allow recycling, the team compares the potential return from an existing liquid position with the potential from a new investment, rather than assuming every holding should be sold at a fixed point.

He expects growth-stage investing to remain relatively concentrated among a limited group of institutions, although participants can enter and leave over a cycle. He also admits a missed opportunity in Coinbase: an elaborate model gave him a false sense of precision.

In a quick-fire round he selects Bending Spoons as a hypothetical long-term holding, praising the business model of acquiring and improving digital products. Baillie Gifford is an investor, and this is his personal investment view rather than a recommendation.

Asked to choose among major AI model companies at specified valuations, Singlehurst says he would buy none at that moment because he cannot yet identify durable competitive advantage at the model layer. That uncertainty is different from a claim that the products lack value.

He says he changed his mind about how growth investors can help companies. Beyond operating advice, he now sees value in helping build an independent board, prepare for an IPO and navigate the responsibilities of being public.

Singlehurst describes Baillie Gifford as an intergenerational partnership, with carried interest going to the firm and employees remunerated through bonuses. He links its culture to responsibility for clients and stewardship across generations.

He closes by saying he is optimistic about the supply of growth-stage companies and experienced people. He believes capital availability across the market is broadly more balanced than during the peak, while acknowledging pockets of exuberance; this is his assessment at the time, not a prediction of returns.