Dear Partners,

Fairlight Alpha Fund’s returns in the first quarter 2026 were 4.6% net of fees. This compares to a return of -4.4% for the S&P 500 Total Return index. Overall, since inception the fund has returned 950.7% net of fees and 39.3% on an annualized basis. These results are presented for informational purposes only and do not represent a solicitation or offer.

Performance vs. the S&P 500

We track the net asset value of Fairlight Alpha Fund ((the “Fund”)) over time as calculated by our fund administrator based on our portfolio positions and prices over time on a dollar basis. The fund financials are also audited each year by an independent auditor with the assistance of the Fairlight management company and the fund administrator.

The tables below show the S&P 500 total return and partnership net returns, after fees, for each year 2019-2026 (YTD), followed by the compounded returns and annualized gains over the same period.

Annual Returns

Year	Fairlight Alpha Fund	S&P 500 ((Total Return))	Difference
2019 1 .....	38.1%	17.9%	20.1%
2020 .....	26.3%	18.4%	7.9%
2021 .....	146.5%	28.7%	117.8%
2022 .....	15.8%	-18.1%	33.9%
2023 .....	6.9%	26.3%	-19.4%
2024 .....	13.6%	25.8%	-12.2%
2025 .....	66.2%	17.2%	49.1%
2026 (YTD) .....	4.6%	-4.4%	8.9%

Compounded Returns

Year	Fairlight Alpha Fund	S&P 500 ((Total Return))	Difference
2019 1 .....	38.1%	17.9%	20.1%
2020 .....	74.4%	39.6%	34.7%
2021 .....	329.9%	79.7%	250.2%
2022 .....	397.9%	47.2%	350.7%
2023 .....	432.2%	85.9%	346.4%
2024 .....	504.6%	133.8%	370.8%
2025 .....	904.9%	173.9%	731.0%
2026 (YTD) .....	950.7%	162.0%	788.7%
Annualized Gain .....	39.3%	14.6%	24.8%

1 The Fairlight Alpha Fund was launched on 01-Mar-2019 and so the 2019 performance and returns are presented for a 10-month period.

1. (1) The S&P 500 returns represent the total return index, i.e., with dividend reinvestment included. Increases in value arise from stock price appreciation as well as dividend reinvestment. This provides a fairer comparison to the fund.
2. The Fairlight Alpha Fund performance is shown net of partnership fees and expenses.

Markets

The year started with expectations of Fed rate cuts, though January concluded with falling prices and heightened volatility in precious metals. The conflict in the Middle East soon triggered a dramatic reversal in inflation expectations. Volatility, as measured by the VIX, rose quickly above 20%, peaking at 30%, before falling back to the high teens. This quarter has been one of increased uncertainty and unknowable outcomes, but in spite of this the equity market has been relatively resilient (particularly in the safe haven of the U.S.), with only a modest fall in the main U.S. indexes to quarter end (which as of the time of writing is now into positive territory).

Gold began the year as strongly as it had finished the last, reaching intra-day highs of $5,595/oz at the end of January. Subsequently, it fell from these highs and became more volatile, possibly because of the crowded nature of gold trades up to that point. In recent weeks it seems to have stabilized, albeit in a broad range between $4,500/oz and $4,800/oz. Our junior gold miner positions had similar volatility, although we think they remain very cheap on a free cash flow to valuation basis.

On February 28 th the U.S. and Israel launched a series of coordinated strikes against Iran. This was followed in early March by counter-strikes from Iran at U.S. bases across the Middle East and Israel in retaliation. By March 2, the Islamic Revolutionary Guard Corps officially declared that the Strait of Hormuz was closed, leading to a collapse in maritime traffic through the strait, in part due to the removal of coverage by insurance companies. In turn, as most will know of course, this led to a spike in oil prices, as well as a divergence between crude oil priced in different regions, with Brent rising faster than WTI prices, and Dubai Brent rising fastest of all (rising above $160/oz at one point).

As of the time of writing, June WTI futures stand around $104 per barrel and Brent futures at $109 with the prices remaining elevated, although the spread between them having closed. The conflict has led to a series of bifurcations in the market with some stocks selling off on this volatility, but oil companies rising due to these exceptional prices and the cash flows that are now being generated. It's not our expertise to predict the future of these kinds of events, but there is clearly increased volatility and level of risk. It seems to us that some of the damage to the oil market could be permanent with damage to ships, infrastructure, and the risk of shut-ins and permanent damage to oil wells in Iran and across the rest of the Middle East. We will discuss some of our thoughts on oil companies (one of which we owned prior to this crisis) and what might happen later.

Before this conflict, inflation appeared to be trending towards the Federal Reserve's 2% target and their dot plots showed that rates could be reduced perhaps one or two times through the course of 2026. Following the Iran strikes, there was a marked rise in inflation expectations as well as the course of interest rates this year. The Fed's median expectation for rates in the U.S. is now one cut later this year, with markets even more hawkish projecting a less than 40% likelihood of one cut.

The 10-year Treasury bond has risen to 4.42% from 4.21% earlier in the year. Perhaps the most shocking aspect of all of this, is that these moves have been so small. Given that the largest oil shock in history has just occurred, these seem to be small reactions. Perhaps renewables, flexibility in energy sources, a switch to nuclear fission power, as well as increasing adoption of electric vehicles are softening the effects, or perhaps the real effects are yet to be felt.

Turning back to oil and gas, it is worth paying attention to the demand aspect of the equation. Around 20% of oil produce (and likely an increasing percentage) is tied to the petrochemical industrial industries, with crude oil and by-products being used to produce plastics, synthetic fibers (like polyester and nylon), synthetic rubber (for tires), detergents, and pharmaceuticals. And gas by-products being used to produce fertilizers (ammonia and urea), solvents, antifreeze, and specialized resins (that are used to bind aircraft parts, iPhone components and EV construction).

Perhaps this moment will accelerate the trends to reduce our dependency on oil and gas and that demand may fall faster than we expect. There may be increased usage of biomaterials and bioplastics, increasing use of electrification rather than gas burning in industrial processes, usage of AI and digital twins to monitor and increase efficiency in chemical plants and petrochemical processes. And perhaps an increased focus on recycling (which would also be beneficial in reducing dependency on ever increasing mining volumes for the modern world).

But one thing is certain, in that the above geopolitical events have occurred and had significant impacts on price levels. The duration and magnitude of these effects is still very uncertain at this time, but in our view, oil prices are likely to remain higher for longer than many markets and stock prices are anticipating.

However, our focus remains in the low correlation world of micro caps and obscured investments where idiosyncratic moves dominate these wider, macro themes. We aim to find as many of these across as many regions and sectors as possible. Outside of the major markets there have been many moves in small-cap names not coinciding with any major market news events. We will discuss some of what we have been seeing next.

Portfolio Performance

Aside from the outperformance of the oil and gas sector (we were holding Valeura Energy (VLERF) through last year and prior to the Middle East conflict), one of our best performers so far this year has been Cematrix Corporation (CTXOF). This is a new position for us built up in Q1 2026, in an exciting business that has a long-term inflection in its business strategy, pivoting more toward the manufacture of specialist cellular concrete products.

The business has a great business model, that has strengthened from their purchase of a specialist low-density foam concrete, or cellular concrete manufacturer, with expertise in tunnel projects, as well as a wide variety of other infrastructure work such as bridge approach embankments, highway retaining walls, and slope stabilization. This business has a low correlation with some of the geopolitical events facing us this year, primarily dependent on the requirement for replacement or remediation of ageing infrastructure in many parts of the U.S.

Annual revenue growth has been 25% since 2017 the company showed significant earnings growth last year, and even after the recent share price run up, our estimate of the valuation versus this year's EBITDA is less than 6 for Cematrix. Given the company's large order book we think that this growth will continue, the company itself has stated that they expect record results in Q2 2026.

A position from last year, Beng Kuang Marine Limited (BNKMF), has continued along its inflection path with an excellent move to purchase their partly owned subsidiary ASOM (Asian Offshore Services Management). They owned 49% of this subsidiary up until now, with it driving a core component of Beng Kuang's strategy to become more of an asset-light business. Previously, the company had primarily operated out of its two shipyards in Batam (Indonesia) and Kuantan (Malaysia). This was CapEx and OpEx heavy and required a lot of fixed assets to operate.

The decision was made by management in 2021 to shift away from this business to focus more on consulting and on-site repairs to FPSO and other types of energy production vessels. This reduced the need for the full Batam shipyard and has allowed them to reduce fixed assets and increase margins for services provided to customers. To illustrate this, gross margins in 2022 stood at 21.2% and in 2025 reached 37%, with this trend likely to continue for several more years, particularly after the ASOM transaction completes. This is not done yet, but there will be an EGM in the near future which should see the deal get finalized.

After this is complete, it is likely that 2026 results for Beng Kuang will be even stronger. In particular, note that in a recent press release the company stated that S$28 million of work at the ASOM subsidiary had already been secured this year, as compared to a total revenue of S$98 million last year, for the full year across the whole company. It's worth noting that much of this ASOM revenue is recurring in nature, being tied to long-term FPSO maintenance contracts involving inspection, maintenance and life extension. Their other businesses are also growing, and so it seems there are healthy tailwinds for the company over the next couple of years.

The laggards in the portfolio were the precious metal junior mining companies. These were initially hit by the volatility at the end of January and have subsequently suffered from the outlook of increased energy costs following the Iran-U.S. conflict. We think the fall in these names is somewhat overblown, in particular the ones we own such as Monument Mining (MMNGF) which have been increasing production through 2024 and 2025. Energy increases will have an impact on the AISC for these names, but as long as there isn't a complete loss of access to diesel or energy supplies they're still trading at cheap levels. Falls in the gold price have an outsized impact on them, and so it becomes a question of waiting for the "area under the curve" of the gold price versus AISC in dollars per oz to become apparent in the free cashflow dollars.

In addition we have just finished our primary analysis on a new idea, and that we are still building a position up in. This one could be very interesting with extremely high revenue growth, a net cash position, and inflecting to a growing positive EBITDA position. Projecting forwards to next year, our estimate is that it trades at EBITDA (2027) based on current prices. We will write this one up in the near future in our Substack and will likely discuss further in our future letters.

Simple Rules

Benoit Mandelbrot said, with reference to fractal patterns, that:

"Bottomless wonders spring from simple rules, which are repeated without end."

It is staggering that just using the recurrence equation produces the Mandelbrot set: a fractal image of never ending complexity and finer and finer details. In a sense AI and LLMs are creating complexity out of simplicity in a similar way. The fundamental math equations used in these models are simple, with weightings applied as they "learn".

The complexity arises, in part, due to the vast number of these inter-related equations, weightings and the learnings process that utilizes vast quantities of data, with hundreds of TB used for the latest models. Spanning the common crawl (website data), books corpuses (a large amount of human writing) and GitHub (programming code), Wikipedia, and scientific papers, amongst an increasing number of other sources. The parameters in the model (or weights) now number in the billions (currently approximately from 2-400 billion 2 ). It's interesting to compare this to a human mind, which as a parallel has over 100 trillion synapses ((similar to weights)) connecting neurons.

However, perhaps we humans have some time left, given also that our brain uses 20 Watts of power ((like a small light bulb)), compared to an LLM that might use upwards of MW or GW. Although the LLM is of course becoming every more complex, and its speed and parallelism is much higher than a human brain, the comparison of Joules per thought, to Joules per inference indicates that AI is still less efficient. I would guess that this efficiency gap won't last long.

The reason for this, perhaps seemingly unrelated, digression is that we have been experimenting extensively with AI as an additional process for idea generation and analysis assistance over the last 12-18 months. The question being whether AI can helpfully supplement our existing processes, running in parallel to produce additional idea generation.

Despite recent advancements, artificial intelligence still lacks a certain 'human element' that remains difficult to replicate. While AI is a powerful partner for brainstorming, generating insights (which still require rigorous fact-checking), it frequently overlooks nuanced data or defaults to a safe, 'average' consensus. Crucially, it lacks the human intuition needed to synthesize disparate facts or the physical presence to conduct CEO interviews and participate in earnings calls, or an AGM! It is, though, an incredibly useful tool and we will continue to work to try and extract as much value as is possible from these new models.

Outlook and Partnership

Our aim is to intensively search for ideas using all the means at our disposal: from A-Z manual searches to enhanced AI agent search and analysis, we look to accelerate idea generation as much as possible. In spite of recent volatility, which looks set to continue, we outperformed the market. There always seem to be a wide range of interesting ideas if you turn over many rocks, and so we will keep searching rapidly around the globe.

For those who want to read up on our latest ideas in real time, we post the more interesting ones on our Substack, which is free for our investors. Feel free to reach out to us there with any questions, ideas you have found interesting, or feedback.

Yours,

A. Martin

---

References

1. The Fairlight Alpha Fund was launched on 01-Mar-2019 and so the 2019 performance and returns are presented for a 10-month period.
2. The S&P 500 returns represent the total return index, i.e., with dividend reinvestment included. Increases in value arise from stock price appreciation as well as dividend reinvestment. This provides a fairer comparison to the fund.
3. The Fairlight Alpha Fund performance is shown net of partnership fees and expenses.
4. Interestingly, more weights and complexity does not always produce the best model. So, there are a vast range of model sizes, but a push to efficiency to beat AI benchmarks with ever more efficient models.

---

---

Original Post

Editor's Note: The summary bullets for this article were chosen by Seeking Alpha editors.