By Motlatjo Seima, editor of Citywire South Africa.
Original Article: Read Here
Investors have piled into AI ‘winners’ while dumping platform businesses. But earnings are still growing at double-digit rates, and the multiples may have overshot.
Investors chasing artificial intelligence winners may be overlooking one of the market’s most compelling opportunities, according to Centaur Asset Management, which believes fears of AI disruption have pushed several high-quality global platform businesses to unjustifiably low valuations.
Laing Walker (pictured above), portfolio manager at Centaur, told Citywire South Africa that companies including Uber, Tencent and Trip.com continue to deliver strong earnings growth and high returns on equity, yet are being priced as though AI will render their business models obsolete.
‘The market is selling a doomsday narrative, and we’re buying the fundamentals,’ he said.
Walker believes investors have failed to distinguish between platform businesses whose competitive advantages are genuinely under threat and those whose scale, data and network effects are likely to become even more valuable in an AI-driven economy.
The distinction matters. Companies such as Uber, Tencent and Trip.com are being priced as if disruption is inevitable, despite continuing to grow Ebitda by 15% to 30% a year while generating returns on equity of around 30%.
Walker argues the market has conflated two very different types of platform business. One group, including Upwork, Etsy and Wix, primarily provides simple matching services. Those businesses are already seeing users displaced by AI and their competitive advantages are weakening.
The second group is far more deeply entrenched. Uber has spent nearly two decades refining its logistics algorithms using billions of miles of driving data. Tencent has become the operating system for 1.4 billion people in China, while Trip.com has built exclusive supplier relationships across Asia that AI cannot easily replicate.
Market could be wrong
Walker identifies three reasons why the market’s AI disruption thesis may be overstated.
First, AI struggles to replicate physical infrastructure, network effects and proprietary data.
‘How is AI going to replicate Amazon’s massive logistics and fulfilment network?’ he asked.
Second, many of these businesses are already embedding AI into their own platforms.
‘Many of these platforms have already developed their own agentic capabilities,’ Walker said. ‘They can already provide an agent inside their ecosystem that fulfils a transaction from end to end.’
Third, the large language model providers may have little incentive to replace platforms altogether.
‘These LLMs and their agents don’t want the liabilities,’ Walker said. ‘They really just want to be directing traffic.’
He added that many platform businesses are already among the largest customers of AI providers, making widespread disintermediation economically unattractive.
The bear case
Walker acknowledges the risks.
If AI agents eventually replicate Uber’s logistics capabilities or Trip.com’s supplier relationships, the investment case would weaken. The same applies if incumbent platforms fail to adapt quickly enough.
Margin pressure is also likely in the near term.
‘There is likely to be a period where AI investment runs ahead of monetisation, which could pressure margins,’ he said. ‘We’re already seeing that today.’
However, Centaur is deliberately backing businesses with multiple competitive advantages.
‘We’re focused on businesses with adaptive management teams and multiple layers of defence, including physical execution, proprietary data, network effects and distribution,’ Walker said.
Closing the gap
Walker believes earnings growth, rather than a change in investor sentiment, will ultimately close the valuation gap.
Around two-thirds of Centaur’s expected 20% internal rate of return over the next three years is expected to come from earnings growth, with only one-third dependent on multiple expansion.
‘That’s important to us because the investment case doesn’t rely on sentiment changing,’ he said.
‘If sentiment remains weak, we still have businesses growing earnings at attractive rates. If the market becomes more comfortable with AI’s impact, the rerating could be significant.’