The Minotaur Global Opportunities Fund was relatively flat (-0.5%) in January, outperforming the MSCI ACWI (Net, AUD) which fell 2.0%.

The memory pairs trade that contributed strongly in December continued to deliver, with Micron and SK Hynix both up strongly. Hut 8 also extended its run to be our third-largest contributor. This was offset by profit-taking in Artrya, and negative contributions from Atlassian (which we have since sold) and First Solar.

Software: From “Misplaced Fears” to “Maybe Not”

The most significant portfolio change this month was our move to a short position in software, a meaningful about-face from our more constructive stance in December.

We spoke about this in-depth in our December Quarterly, but as a recap, we were fairly vocal about the opportunity we saw in software, arguing that parts of the sector were being punished by misplaced fears around AI disruption. Since then, our view has evolved. We now think those fears are less misplaced than we initially believed.

In January, we moved from net long software to net short. We exited some software longs and initiated a diversified portfolio of 12 software shorts, focusing on companies with the least protected business models. We’re still open-minded about where the winners ultimately emerge, but in the near term we see the balance of risks skewing differently than it did a month ago.

For over a decade, SAAS was viewed as incredibly high quality: recurring revenues, low churn, pricing power, hard to switch. Terminal values were bankable. That narrative is now being challenged.

The market is now pricing in uncertainty about what this sector looks like five years from now. Will new competitors emerge faster and cheaper? Will per-seat pricing models survive as teams shrink? Forecasting the impact of these changes five years out is inherently a theoretical activity. Channel checks with IT departments today won’t give you much insight – they’re focused on immediate purchasing decisions, not how the competitive landscape might evolve. Markets discount the long-term future, not the present.

So what turns it around? Individual companies need to demonstrate they’re AI winners, not AI losers. Every CEO has a plan to benefit from AI, but that needs to show up in the numbers through revenue acceleration as they add more value to customers that they can capture. If concrete examples emerge, the valuation story shifts quickly. Until then, uncertainty reigns.

Gold, Silver, and the “Just in Case” Trade

We also initiated a position in gold, silver and PGM stocks. The logic is straightforward: central bank diversification away from US assets continues, industrial demand for silver remains structurally supported, and gold is still under-owned globally (we estimate it’s ~3% of global AUM vs. the 5-10% it could be in a world that’s re-learning what geopolitics looks like).

There was some volatility late in the month as markets digested Trump’s nomination of Kevin Warsh to replace Jerome Powell as Fed Chair. Warsh is viewed as hawkish and independent and markets had been pricing in aggressive rate cuts and dollar weakness but this move signals that might not be the case. We’re not too concerned. If anything, the chop is a reminder of why we like holding some “hard asset insurance” in a portfolio built for weird outcomes.

Genie 3 and the Pace of Research

On January 30th, DeepMind announced Genie 3 – an AI system that can generate playable 3D game worlds from a single image or text description. Unlike video generators that create passive footage, Genie 3 produces interactive environments. A player can explore them, manipulate objects, and the world responds consistently. It’s a significant step toward AI that can simulate reality rather than just describe it.

Within hours, footage started appearing on X. The market reaction was immediate and severe: Unity fell 24%, Roblox 13%, Take-Two and CD Projekt (which we own) fell 7%. These are multi-billion dollar gaming companies shedding billions in market cap because people were posting videos of AI-generated worlds.

We immediately spun up a research process using the iterative techniques we described in our December Quarterly. From a 127-word prompt asking for implications on the games industry, our AI system iteratively chose what to explore: value chain analysis, five-year scenarios with falsifiable signposts, unit economics ($/minute cost models), IP and licensing questions, and a winners/losers matrix across engines, platforms, and publishers. Over 50 iterations it built out each section, cited sources, and stress-tested its own conclusions.

We viewed the selloff as overdone. While Genie 3 is impressive, it generates simple explorable environments, not the complex systems, narratives, and polished experiences that AAA studios like CD Projekt deliver. The gap between “playable world” and “commercial game” remains vast. Established developers with strong IP and live-service expertise remain well-positioned. Read the Genie 3 report here.

This is not polished, definitive research – it’s a demo of how we can mobilise AI-assisted analysis when the frontier moves. Some sections remain undeveloped because we stopped at a certain iteration point. You can always go further, redirect it, or prune sections. But we had a structured, multi-scenario analysis within hours of the announcement.

Iteration and Adaptation

As “Aristotle” supposedly said: “It is the mark of an educated mind to be able to entertain a thought without accepting it.” We operate the same way – open to changing our minds, but only once the research earns it.