Separating Signal from Noise

Concerns about AI disruption returned to the fore during September. The releases of OpenAI's latest frontier model and Meta's Muse AI agent reignited fears that AI could replicate or disintermediate many of the services offered by online platforms, and specialised software and data providers. Against this backdrop, the share prices of several of our portfolio holdings, mainly auto and property classifieds, fell over the month, giving back many of the gains that followed their strong Q2 results.

At Fairlight, we systematically stress test every holding against the opportunities and threats that AI presents, distinguishing between genuine and perceived risk. Our research continues to suggest that the impact of AI is more nuanced than the market’s knee-jerk reactions suggest. As we’ve discussed in recent reports though, we are encouraged that share price swings are increasingly occurring in both directions and believe that ultimately our patience will be rewarded. In fact, some of our holdings which were initially branded as ‘AI losers’ by the market, such as Softcat, (discussed below), have delivered strong year-to-date share price performance as the narrative reversed, with consistent earnings results showing AI to be a tailwind rather than an existential threat.

In summary, while September featured an adverse swing in the AI narrative for some portfolio companies, we have been pleased to receive additional strong updates from portfolio companies through the month.

Softcat

Softcat, one of the UK's leading IT infrastructure resellers, announced the acquisition of GDT, a Dallas-based multi-vendor IT solutions provider specialising in networking and data centre infrastructure. The US$1.05 billion investment will be funded through cash, new debt facilities and a £350 million equity placement which led to a 10% increase in the share count. The Fund participated in the raising after discussing the merits of the deal with management.

Large acquisitions deserve scrutiny, and we have written previously about the risks of buying sizeable businesses from sophisticated sellers. Several features of this deal give us comfort:

• Softcat had followed GDT for around four years and negotiated with the private equity owner on an exclusive basis rather than through a competitive auction.

• GDT's entire management team is being retained and its operating structure left intact.

• Softcat's financial model includes only revenue synergies, built around serving multinational customers on both sides of the Atlantic and leveraging GDT's networking and data centre expertise across Softcat's own customer base.

• GDT's recent growth has been broad-based across legacy customers, new account managers and new clients, and is tied to enterprise infrastructure investment rather than hyperscaler capex.

• Management expects high single-digit to low double-digit EPS accretion in the first full year of consolidation and financial leverage to return to around 1x within 18 months.

Along with the acquisition, management also announced that, excluding the impact of GDT, they expect Softcat’s current business to grow operating profits in FY27 at a high single-digit rate, after this year’s growth of 18%.

As discussed last month, the extreme volatility of Softcat’s share price over the past 12 months is a useful reminder of how quickly market sentiment can turn. Not long ago, the company was widely regarded as an “AI loser”, with investors concerned that AI would shrink the software licences and seat counts its customers buy or allow vendors to sell direct. Today, as enterprises accelerate investment in networking, and AI infrastructure, the market increasingly views Softcat as an “AI winner”.

The underlying business changed far less than the narrative did. Shifts like this can happen quickly, but their timing is difficult to judge, which is why we anchor our decisions on fundamentals and valuation rather than attempting to predict when and where consensus will move.

Halma

Halma is a long-held manufacturing business that owns a growing portfolio of more than 50 safety, environmental and healthcare technology companies. One of these, Avo Photonics, has built a relationship of more than a decade with a large hyperscaler technology customer, co-designing and manufacturing optical switches critical to data centre infrastructure. Unsurprisingly, business has been booming over the past couple of years. Avo alone contributed to almost half of Halma’s exceptional 16% organic revenue growth in the financial year ended in March 2026. While Avo is still growing fast, we are supportive of management’s decision of reallocating Avo’s supernormal profits to diversify the group further and thus see Halma’s recent acquisition activity across niche medical, safety and water monitoring businesses favourably.

In September, Halma released its first-quarter trading update for FY27. Revenue guidance was unchanged, with management continuing to expect low-double-digit organic growth, also reflecting roughly 30% growth for Avo Photonics. Importantly though, the EBIT margin outlook was raised to 23.5–24.0%, up from the previous guidance of circa 23%. This profitability upgrade was broad-based, confirming the strength of Halma’s portfolio of niche businesses rather than pointing to an overreliance on a single company highly leveraged to AI infrastructure spend. The share price reaction was muted but the investment thesis remains firmly on track.

Diploma

Distribution company Diploma hosted an investor call during the month focused on its successful acquisition strategy. Since 2019, Diploma has completed around 60 acquisitions for £1.7 billion, increasing its cadence from a couple of deals a year to between eight and twelve. The playbook, however, has not changed. Acquisitions are selected for their ability to lift the organic growth of the whole group, the hurdle is a return of around 20% within five to ten years, and acquired businesses are largely left to run themselves.

What stood out to us was management's emphasis on people. Culture and management quality are the most common reasons deals are abandoned late in the process, and deal structures are designed around retaining key individuals. Diploma frequently wins assets without being the highest bidder, competing on certainty and speed rather than price. With leverage expected at 0.7–0.8x by year-end, the company has meaningful capacity for further deals. This is the type of disciplined, decentralised capital allocation we look for in an acquisitive compounder.

The Fairlight View

September was a reminder that share prices of quality businesses can move sharply on shifts in sentiment that have little bearing on their long-term earnings power. The portfolio's 32 holdings now trade on 18.4x forward earnings, with forecast EPS growth of around 13%, consistent with our long-term objective. We believe this represents an attractive entry point into a portfolio of businesses with resilient growth, high returns on capital, and low financial leverage. Most portfolio companies will report Q3 earnings results over the coming weeks. We look forward to sharing further feedback next month.