Two storylines dominated global markets in the first half of 2026: an on-again, off-again US-Iran conflict that has kept oil markets and volatility on edge, and the relentless build-out of global AI infrastructure. On the surface, geopolitical risk and technology capex seem worlds apart. But in practice, both have reshaped where genuine investment opportunity is found. Reading this shift correctly, as opposed to chasing the headlines, is exactly what the Old Mutual Global Managed Alpha Fund's dynamic factor model is designed to do.

Iran-again-off-again

As uncertainty around Iran ebbs and flows, investors have grown warier of crowded, richly-valued positions and have started to reward shares offering genuine relative value and resilience to volatility. At the same time, as the AI story has matured, and rather than piling further into the same mega-cap names, capital has started flowing toward the broader ecosystem of companies enabling the AI expansion – many of them under-the-radar and found in emerging markets.

Going into 2026 against this backdrop, GMA's model tilted the portfolio toward shares with strong exposure to price momentum, higher volatility and better relative value. Chart 1 shows MSCI's Value Index against our benchmark, the All Country World Index (ACWI), since GMA's December 2017 inception – for most of this period, ACWI outperformed value stocks. Chart 2 zooms into the last six months, showing value's strong recovery. Our model had detected this shift early and positioned the fund accordingly – a good example of the process working as intended, ahead of the broader market narrative.

Moving downstream

While global AI capex has not slowed – with hyperscalers on track to spend close to three-quarters of a trillion dollars on AI infrastructure this year, sharply higher than 2025 – what has changed is where investors are willing to pay for that growth. After years of concentration in a handful of platform companies, the AI trade expanded in H1 2026. As valuations at the top of the value chain stretched, capital rotated toward the physical infrastructure making AI development possible – power and grid capacity, high-bandwidth memory, and the optical components connecting server racks.

This is where the data bottlenecks sit, and, alongside power constraints, are where pricing power looks more durable than in the software layer that dominates the headlines. This is why many of the best-placed beneficiaries are not the usual mega-cap names, but industrial and emerging-market businesses that had, until recently, gone largely unnoticed – "picks and shovels" companies supplying the AI economy rather than selling it.

This has translated into good performance for the period. As per the table, the fund outperformed ACWI by 1.82% to 30 June 2026. Since inception (8 years and 6 months), the excess return has been an annualized 1.62%. This has produced top-quartile performance across all periods.

Three company overweights | Not as sexy, but good investments

GMA uses a bottom-up process, selecting stocks with the best exposure to six factor buckets: Growth, Momentum, Volatility, Company Size and Value. The model is sector agnostic, but strict risk controls cap sector over/underweights at 3% and stock-level over/underweights at 1%. Notably, we started 2026 slightly underweight most of the Magnificent Seven, with the model instead preferring companies further down the AI value chain. Below are three of our top 10 alpha contributors for the six months to June 2026.

GE Vernova (USA)

Spun out of General Electric in 2024, GE Vernova ($290bn market cap) sits in the "unexciting" industrial sector, producing power generation and grid equipment – a direct beneficiary of the aggressive data centre roll-out. Chart 4 shows its factor exposure sharpening through H2 2025, with strong momentum and volatility scores driving an overweight position.

Samsung (South Korea)

Samsung ($1.1tn market cap) has aggressively pivoted toward next-generation high-bandwidth memory for AI servers – now a national imperative as Korea competes with Taiwan's TSMC in what both governments call a memory "super-cycle." Chart 5 shows factor exposures misaligned with model predictions until September 2025, when Samsung's score improved sharply on stronger value, momentum, and volatility alignment.

Zhongji Innolight (China)

This $180bn Chinese business manufactures optical components connecting AI servers – unglamorous hardware, but Zhongji Innolight is one of the world's largest suppliers of these modules. Chart 6 shows the model turning positive on the stock in June 2025: as the market recognised its AI infrastructure exposure, the share price rose, boosting both its momentum score and its market cap – reinforcing the model's size-factor tilt in a virtuous circle.

Staying ahead of the narrative

None of these three positions were obvious calls at the time they were built. GE Vernova, Samsung and Zhongji Innolight became meaningful contributors precisely because the model identified their improving factor exposure before the broader market re-rated them on the back of AI infrastructure demand. This is the value of a disciplined, data-led process in a period defined by noisy, fast-moving information. It does not require correctly predicting the next geopolitical turn or AI capex announcement, but rather reading the market's evolving preferences as they emerge, and positioning ahead of the crowd. 

During the first half of this year, the Old Mutual Global Managed Fund reached a significant milestone, with assets under management going above $1bn. It will soon celebrate its ninth year since inception. Despite the volatility, GMA remained true to its investment process, ignoring the noise and producing solid core global equity excess returns across both developed and emerging markets.

As AI infrastructure development matures further and investors continue to broaden their search for value beyond the mega-caps, we believe this approach leaves the Old Mutual Global Managed Alpha Fund well placed to continue finding alpha in places others have yet to look.