The Gap Begins to Close

Three months ago, we argued that the gap between what our companies were earning and what the market was willing to pay for them was a source of future return rather than a reason for concern. In the third quarter, some of that return arrived. One good quarter after a difficult start proves very little on its own. It is, however, the first clear sign that the process is doing what it was designed to do.

Navigating uncertain markets

In our launch letter we warned investors to expect short-term volatility, including valuation compression driven by AI disappointment. We have had plenty of it. Since launch, the fund's volatility has been around 20%, against roughly 12% for the index, and the NAV fell 16% from its January high to its low on 30 March.

That drawdown was a multiple event, not an earnings event. Consensus estimates for our largest laggards rose throughout the sell off, even as their share prices fell. The market decided to sell first and read the earnings revisions later. Investors who stayed with us through March have since been paid for their patience. We do not expect it to be the last such episode: shorter cycles and sharper swings are what the four forces imply, and they are why we built a fund that can act on them.

Software led, as we said it might

In Q2 we argued that the market was paying for demonstrated AI revenue and pricing promised AI revenue at zero, even where earnings estimates kept rising. Through the first half of the year it paid handsomely for the picks and shovels of AI infrastructure while discounting the application layer on fears of disintermediation, largely regardless of what reported numbers were doing.

In the third quarter that began to change. Our largest contributors were Zeta Global, Veeva, GitLab, Doximity, Money Forward and Dynatrace. Several of them, Veeva, GitLab and Doximity in particular, had been among our largest detractors at the end of June.

Veeva illustrates the pattern. Over the twelve months to June, consensus earnings for FY2027 rose from around $8.34 to $9.06 a share, yet the shares fell some 40% from their November peak. Nothing in the business justified that fall. In Q3 the shares rose 61%. It remains one of our larger holdings, sized for a business whose share price has caught up with part, but not all, of its earnings.

We think we are in the early innings of the application era, not at its end. Infrastructure gets built and paid for first; the applications that make it worthwhile arrive later than the capex. The gap between estimates and share prices has narrowed, but in places it remains wide.

Semiconductors gave some back

Soitec was again one of our best contributors and is once more our largest position. Its silicon photonics business has turned it from a smartphone supply chain story into an AI data centre beneficiary, and sales estimates have only just started to turn in what could be a multi-year cycle.

Elsewhere, last quarter's winners handed back part of their gains. PDF Solutions, ACM Research and ASML were among our weakest names in Q3. We used the weakness: we added to all three over the quarter.  Recent share price action indicates this was the correct decision.

Leadership within the portfolio changed almost completely in a single quarter. The rotation that hurt us in the first half, out of software and into semiconductors, ran in reverse. We did not predict its timing, and we do not expect to predict the next one. What we can do is own businesses which have their own drivers across more than one end market, so that we are not dependent on getting the rotation right.

Healthcare coming good?

Healthcare has been the hardest part of the portfolio since launch. We chose it partly as a counterweight to technology: our healthcare holdings have a beta of around 1.0 (the healthcare sector as a whole is lower still), against nearly 1.5 for our technology names[1]. In a calmer world that would have dampened the fund's volatility. Instead, the sector faced two headwinds at once. The conflict involving Iran unsettled markets and corporate confidence, and the US administration's drive to lower drug prices led large pharmaceutical companies to cut back on research and development, and on sales and marketing. That spending is revenue for several of our holdings, Doximity and Phreesia among them, and earnings disappointments followed.

We believe the worst is now behind us. In our view drug pricing is no longer at the top of Washington's agenda, and in Q3 healthcare added value for the first time since launch. DexCom, a manufacturer of blood glucose monitors, shows where the long-term opportunity lies. Having established itself in type 1 diabetes, it is now selling into type 2 and, increasingly, to people monitoring their health for wellbeing rather than disease, a far larger market. The demographic forces we set out in our recent healthcare note have not gone away. If anything, the pressure on health budgets makes the companies that improve outcomes and lower costs more valuable, not less. We see the weakness as cyclical and not structural.

Agile investing in practice

Agility does not mean trading for its own sake. It means acting decisively when price and value diverge, or when the facts change.

Pinewood Technologies completes a story we began in our first letter. We sold the position into the Apax approach in February at 475p, rebuilt it at an average of 218p after the bid was withdrawn, and sold again in August at around 440p, still below the 500p Apax had indicated it might pay. The rebuilt holding roughly doubled in four months.

Soitec shows the same discipline over a longer period. We added heavily in February at around €31, after meeting management and concluding that the fundamentals were closer to turning than the share price implied. We took profits on the way up through the spring and rebuilt the position after the summer pullback.

Not every decision works. Enovis was a mistake, and we sold at a loss of around 22%. We had discussed cash generation specifically with the CEO, and the company then made an acquisition that pushed it further out. When a management team acts against what it has told us, we look for better opportunities, particularly when cash generation is the question.

Our one new holding is Bakkafrost, the salmon farmer. I have owned salmon stocks since 2011, and the sector has changed markedly in that time: consolidation, and regulation requiring operators to farm sustainably, have improved returns considerably. The thesis is simple. Salmon prices should rise as supply and demand move in producers' favour. Demand is growing quickly, led by China, and salmon is a healthy protein that suits consumers moving to healthier lifestyles, a trend reinforced by GLP-1 drugs. On the supply side, the strong growth in Norwegian output, helped by biological advances, is now slowing. Bakkafrost has an exceptional asset base in the Faroe Islands, and after a number of false starts we believe it is on the cusp of building a strong franchise in Scotland. It sits squarely within two of our four forces: ageing populations seeking healthier diets, and an environment in which protein has to be produced sustainably.

Re-rating, part of the plan

Some of this quarter's return came from the market paying more for our companies, not just from the companies earning more. Even so, the effect on valuation has been modest. On our own forecasts the portfolio trades on around 21x next year's earnings (harmonic average, excluding three businesses that are only marginally profitable today), against 19x at the end of March, even though the fund has risen by more than a third since then. Earnings expectations have risen alongside share prices. That is the thesis we set out in Q2 beginning to play out.

It does mean the portfolio now trades at a premium to the market on next year's earnings. We think that premium is justified. Consensus expects our companies to grow earnings by around 19% a year. Our own forecasts, built bottom up over five years, have them growing at 28%. That gap is where we believe our edge lies, and it is also where our risk lies.

On a five-year view, the opportunity has not been used up. If our forecasts are right, the portfolio needs an exit multiple of just 11.2x earnings at the end of our forecast period to achieve our targeted return profile. In March that figure was 11.3x. The share prices have moved; so have the earnings, and the margin of safety is essentially unchanged. Our forecasts already assume that valuations contract over the period, so the returns we expect rest on earnings growth rather than further re-rating. If we are wrong, the premium will narrow; that is the risk. We have built this process over 20 years. AI now lets us build and update our models faster, but the assumptions remain ours, based on experience.

Outlook

The history of technology suggests that the companies leading one era rarely lead the next. Hyperscaler capital expenditure now exceeds the cash those businesses generate, and circular investments increasingly blur the economics. The makers of picks and shovels tend to benefit early in a new cycle, but they rarely sustain leadership once demand disappoints or competition erodes returns. We prefer the businesses that will use AI to improve their own returns over those spending heavily to build it.

We believe the application layer is in its early innings. The gap between rising estimates and depressed share prices has started to close, but it remains wide in places, and that is where we continue to find opportunities.

We would like to thank our investors for their patience through a difficult start, and to welcome those who have joined us this quarter. It is early days, and we do not take it for granted. Our process has not changed, and we intend to keep it that way.

Valuation figures: consensus from the Goodhart quarter pack; forecasts from Goodhart's proprietary models, which are subject to change and are not a guarantee of future returns. Source: Bloomberg, Goodhart Partners.

[1] The 2 year weekly beta of the DMHC index vs DM index is 0.546 (0.697 for the adjusted beta) up until 30th September. This is the health care sector out of our benchmark ( Bloomberg Developed Markets Health Care Large & Mid Cap Price Return Index).