By Ben James, a US equity investment specialist at Baillie Gifford
There are moments in investing when the market becomes so focused on the next few quarters that it stops distinguishing clearly between businesses. We think this is one of those moments.
Recent performance in our US growth portfolios has been disappointing. A more defensive backdrop, conflict and geopolitics have all played a part in recent portfolio performance, but the larger force has been the market’s growing conviction that ‘AI kills software’.
What began as a concern about selected parts of software has broadened into a much blunter sell-off across digital, data and technology businesses. Companies with very different economics, competitive positions and routes to value creation are increasingly being treated as though they face the same risk.
We think that interpretation is too simplistic.
Against this backdrop of uncertainty and evolving narratives, the market is rewarding what feels orderly, predictable or safe – just look at the diverging returns of energy and software sectors since September.
The closure of the Strait of Hormuz raises energy prices, and energy stocks are the first-order beneficiaries in terms of revenues and profits. The ‘AI kills software’ narrative suggests software and related companies will experience an equal and opposite decline in their fundamentals.
Despite the recent share price weakness, we do not see a broad deterioration in underlying business quality.
Rather, much of what we have experienced has been valuation compression.
Since June 2025, the S&P 500’s price-to-sales multiple has re-rated upwards, while our holdings have de-rated on average.
Beneath that average, the moves have been even more striking: 20 companies in the portfolio have de-rated by more than 25% over the past nine months. Yet the median revenue growth for those businesses in the fourth quarter of 2025 was 28% and estimate revisions have been more positive than negative. That looks more like a market that has become indiscriminate.
The simple answer behind this is the market is trying to digest a genuine technological shift, but it is doing so in a too-simplistic way.
We do not dispute that AI will change software. It will. But change does not automatically mean value destruction. More often, value shifts.
One way we think about this is through systems. When one bottleneck is removed, another tends to emerge elsewhere. That can make the underlying infrastructure, control points and workflow rails of a system more valuable, not less.
Many of the companies we own are not just ‘software’ in some generic sense. They sit at critical points in real workflows: enabling transactions, enforcing trust, managing logistics, controlling access, capturing proprietary data and helping decisions move reliably from intent to outcome. In an AI-enabled world, those positions may become more valuable, not less.
This is why we think the market’s current mood is too blunt.
DoorDash, for example, has seen strong execution overshadowed by concerns about near-term investment and delayed profitability in newer verticals. Duolingo has been caught in the crossfire of fears around AI translation, even as management is making what we see as a rational trade-off to support user growth and engagement.
CoStar has been marked down as investors have grown impatient for profits, despite the continued strength of its core franchise and its willingness to reinvest from a position of strength. These are not identical businesses. They should not be analysed through a single lens. Yet that is increasingly how the market has behaved.
There is another inconsistency in the current narrative. On the one hand, investors worry that AI capital expenditure has gone too far. On the other hand, they worry that AI is so powerful it will wipe out established software franchises. Both cannot be true in the same way at the same time. If AI proves economically trivial, it is unlikely to destroy durable incumbents. If it proves transformative, then value should accrue not only to the companies building the infrastructure, but also to those deploying AI effectively across real customer workflows.
Our excitement is sharply at odds with the prevailing market narrative and we are excited because the gap between price and business reality appears unusually wide.
Similarly, consumer staples and other physical assets that the market believes won’t be disrupted by AI have rerated. That dynamic has driven a rotation in the US stockmarket.
The IT sector now trades at a lower forward PE than consumer staples. The most dynamic, fastest-growing part of the US economy, priced more cheaply than sugar water. Nvidia, growing sales by over 60% a year, trades on the same earnings multiple as Pepsi, growing in the low-single digits.
Periods like this are uncomfortable. They test your patience and fortitude. But they can also be the moments when long-term returns are seeded. When strong businesses de-rate sharply without a corresponding weakening in their fundamentals or prospects, our instinct is not to retreat mechanically. It is to lean in thoughtfully, selectively and with humility.
