Allocation
Priori maintains a deliberate overweight to technology and semiconductors. We add to it selectively, as company fundamentals and market pricing confirm our theses rather than on price movement alone. The position reflects a long-term view on the quality, durability and earnings power of the businesses we own, and it is reviewed continuously against valuation and new evidence.
Why the strongest technology businesses compound
We favour technology because the best companies in the sector can sustain rates of growth, incremental profitability and return on invested capital that are difficult to find elsewhere. Proprietary intellectual property, software ecosystems, switching costs, network effects and scale tend to reinforce one another over time. Once a company builds a product that is materially better than the alternatives, that advantage can be hard to dislodge, because customers design their own products, infrastructure and workflows around it.
Semiconductors raise the barriers further
These dynamics are sharpest in semiconductors. Designing a competitive processor, accelerator, memory architecture or networking product demands heavy research and development, scarce specialist talent, advanced manufacturing and a deep supporting software stack. The industry has become complex enough that only a handful of companies can compete effectively in many parts of the value chain. Long development cycles, customer qualification and embedded design wins all raise the barrier to entry.
Where complexity becomes an edge
That complexity is also where the opportunity sits for investors willing to understand the technology properly. The economic value of a new architecture may not appear in reported results for years. A product has to be designed, built, qualified and integrated into a customer's own products before the revenue becomes material. Through that period a company can look as though it is underperforming financially, even as it makes the investments that strengthen its competitive position and future earnings.
Once a semiconductor product is designed into a customer's architecture, it tends to stay there across several product generations. That is a different relationship between investment and earnings than most businesses have, where the two move closer together in time. Reading current financial statements alone is rarely enough. It has to be paired with a view on product roadmaps, technical differentiation, customer adoption, industry structure and the markets a company might go on to create.
The lag between innovation and commercialisation is often where careful research uncovers value that current earnings do not yet reflect.
The opportunity extends beyond the chip designers
The same framework shapes how we approach the wider AI infrastructure buildout. We do not see the opportunity as limited to the leading chip designers. Building this capacity also requires accelerated compute, memory, networking, power, cooling, data-centre infrastructure and specialised cloud platforms. That has opened opportunities across established semiconductor companies and a newer group of AI infrastructure providers, including neoclouds.
We stay selective here. Neoclouds can offer significant growth, but they also carry real risks around capital intensity, customer concentration, funding and the pace at which infrastructure has to be built. The aim is not simply to hold exposure to AI. It is to own businesses with credible capacity, a differentiated offering, sound unit economics and a realistic path to durable earnings and cash flow.
A bias, not a blank cheque
A structural bias does not mean holding these companies regardless of price. We reduce positions when market expectations move well ahead of what we think a business can deliver, when concentration becomes excessive, or when the original thesis no longer holds. Valuation and financial discipline remain central. A compelling technology story still has to turn into revenue, margins, cash flow and returns on capital.
Our approach to the sector rests on a few principles:
- Complex, poorly understood industries can reward investors willing to build genuine specialist knowledge.
- Scale, intellectual property, ecosystems and customer integration can produce durable advantages and lasting earnings power.
- Long-term value often depends on product roadmaps and future market structure, not only on current financial results.
- The best technology companies can create or expand markets, generating growth that looks unrealistic through a purely historical lens.
- Valuation discipline is non-negotiable. A narrative has to translate into cash flow and returns on capital.
For these reasons, technology, semiconductors and AI infrastructure are likely to remain a long-term tilt in the portfolio. It is a bias we hold with conviction and review with discipline. A forthcoming Intelligence Report will set out the changes we made at the start of Q3, together with our latest thinking on semiconductors, neoclouds and the broader AI infrastructure opportunity.