Global equities see strong start to 2026

In the first half of 2026, our Global Equity Strategy returned 19.9%, outperforming the MSCI All Country World Index by 7.8% net of fees on an asset-weighted basis. But we did not earn it by sitting still. Market breadth was among the narrowest on record: in April and May, only one in four stocks in the Index outperformed it. When three out of four stocks in the market are hay, finding the needles is harder and more valuable. Every position must earn its keep.

The biggest winners have been concentrated in AI. But not all AI exposure is equal. We frame ours in four groups. The first is the "Core": direct exposure through companies like Samsung Electronics, Taiwan Semiconductor, and SK Square. The second is "Enablers": the businesses providing the power, materials, and real estate that hyperscalers need to operate. The third is AI "Powered": companies where AI accelerates an already compelling model, like Alphabet. The fourth is the "Overlooked": resilient businesses the market has mispriced as AI casualties. The first three are widely understood. The fourth is where we believe the market is most wrong, and where we are hunting.

Take insurance brokerage. Brown & Brown and Ryan Specialty are brokers: they sit between companies and insurers, matching clients to the right coverage and placing it for a fee, without carrying the risk themselves. Both are priced as AI casualties. Two fears weigh on the stocks: a softening pricing cycle and AI disintermediation. The pricing cycle is already in the price. On AI, we disagree.

AI may automate commodity cover like personal auto, but that is under 5% of Brown & Brown's book. The rest is commercial and specialty insurance, where the broker exercises judgement and answers for a denied claim. It is a people business, and Brown & Brown has spent more than eighty years serving mid-sized companies that lack the in-house expertise in insurance.

Ryan Specialty makes the case from the other end of the market: a leading specialist in excess and surplus, the part of insurance built for complex, hard-to-place, and emerging risk. That market is structurally expanding, and we view Ryan as its most capable navigator.

The pattern holds across the Overlooked. We are not buying businesses AI leaves alone. We are buying businesses that put AI to work. Scaled brokers with proprietary data are sharpened by it, not replaced.

The floor test

The same logic reaches well beyond insurance. In any industry, the dividing line is rarely whether a company adopts new technology. It is how. We have seen this before.

When electric motors arrived in factories at the turn of the last century, most owners did the obvious thing: they ripped out the central steam engine and dropped in an electric motor. Productivity improved modestly. Nothing fundamental changed.

The real gains came later. Manufacturers realised they no longer needed to organise the factory around the power source. Smaller motors could be distributed to individual machines, and the floor redesigned around workflow rather than around the engine room. That redesign was where the value was created, not the motor itself.

We think about AI through the same lens. Most early corporate AI adoption is a motor swap: existing workflows, existing structures, existing assumptions, with AI bolted on. The companies that compound the most value are those willing to redesign the floor. This is a question I ask of every management team I invest behind. QXO is doing exactly that. As it consolidates the prosaic, low-tech business of building-products distribution, it is rebuilding the operating model itself, from pricing and procurement to inventory and branch data, rather than bolting technology onto the old way of working.

At Orbis, we have wired over two decades of internal research into AI, making the firm's accumulated knowledge searchable for every analyst. It's valuable, but a motor swap. The deeper ambition is improving our decision-making. We are turning years of analyst decisions and investment reasoning into a dataset we can learn from, identifying where our research engine adds the most value and where our reasoning has led us astray. The goal is not to replace our stock pickers' judgement. It is to raise the ceiling on what they can do. We are in the early stages. We are adapting how we work, not our principles. In time, we believe it can make us better stewards of your capital.

Strong convictions, loosely held

Our founder Allan Gray had a phrase many of us still live by: "strong convictions, loosely held." It means conviction grounded in evidence over emotion, and adapting your view when the facts change.

The owners who won the electrical age were not those who believed most strongly in electricity, but those who held their old assumptions loosely enough to rebuild around a new reality. The same holds true today. Geopolitical realignment, shifting trade, and a generational technology are all in motion at once. And we do not pretend to know how they resolve. The job is not to predict. It is to build a portfolio that benefits more from being right than it suffers from being wrong.