AI will create enormous opportunity and enormous uncertainty. The investors thinking clearly about both are quietly allocating to dirt, steel, and real estate.

Let's be clear about something, most commentators aren't: AI will have an unprecedented impact on the economy, on SaaS, and on the creation of entirely new businesses. Many existing software companies will thrive because of it. The technology is real, the opportunity is vast, and anyone dismissing it is not paying attention.

But AI is also creating a level of uncertainty that most portfolios are not built to absorb. Which SaaS businesses will compound and which will collapse? Which new entrants will own categories that don't yet exist? These are genuinely hard questions. And for investors who would rather own durable cashflows than bet on the right answers, there is a compelling alternative: physical world assets and in particular real estate, that AI cannot disrupt, replicate, or commoditise.

The SaaS landscape: disrupted, but not uniformly

AI is collapsing the cost of building software, dissolving switching costs, and commoditising features that once justified eight figure valuations. The honest reckoning is that a large portion of the SaaS universe is more exposed than its investors currently appreciate.

But the exceptions are real and worth naming precisely. Several categories of SaaS carry moats that AI will struggle to breach. Businesses built around genuine regulatory compliance frameworks where the software is essentially co-authored by legislation sit in a different category.

So do deep systems of record like, Talli in digital payments, where years of transaction history, counter party trust, and integration depth create switching costs no language model can dissolve. Network effects businesses, where the value compounds with every additional user or data point, become harder to displace as AI raises the bar for what "good enough" looks like. And vertical SaaS players embedded in highly specialised workflows benefit from domain complexity that remains genuinely difficult to replicate cheaply.

These businesses aren't selling software. They're selling institutional memory, workflow lock-in, and risk transfer. That proposition is durable and AI may actually entrench it further.

The challenge for investors is discrimination. Identifying which ten to fifteen percent of the SaaS landscape truly fits that profile, before the market does, requires a level of analytical precision that most generalist portfolios won't apply in time.

"Physical assets don't ask you to pick winners in a disrupted category. They ask you to own something that disruption cannot reach."

Three reasons real estate and physical assets win

AI cannot deploy a digger. Physical operations require presence, planning permission, local relationships, construction timelines, regulatory sign-off. The friction that frustrates developers and operators is precisely what keeps competitors out. A well located residential development or commercial property cannot be replicated by a better algorithm. It is not a bug; it is the moat.

Scarcity is structural, not manufactured. Software businesses create artificial scarcity through lock in and switching costs, both of which AI is systematically dissolving. Real estate is constrained by geography, planning law, and the finite supply of land. A residential property in a high demand catchment area, a logistics facility near a major distribution hub, a mixed use development in an undersupplied market, that kind of scarcity does not have a "ship it" button. It compounds quietly over decades.

AI helps physical operators, it doesn't replace them. This is the twist. AI turbocharges the management of physical asset businesses without threatening the underlying assets such as dynamic rental pricing, predictive maintenance, tenant demand modelling and planning application intelligence. A residential portfolio armed with intelligent yield optimisation earns more per unit. The units are still real. Atoms plus AI is a compounding combination that the market has not yet fully priced.

What this looks like in practice

Consider a residential real estate strategy targeting undersupplied urban markets. The demand is structural, driven by population growth, planning constraints, and a chronic shortage of quality housing stock. The cash flows are recurring and needs based; people need somewhere to live regardless of what the stock market is doing. And the asset itself is backed by land, which retains value through technology cycles, inflationary episodes, and market dislocations alike.

Now layer AI on top of that. Smarter tenant acquisition, optimised pricing by micro market, predictive maintenance that reduces void periods and operational costs. The physical asset provides the floor; the intelligence provides the uplift. That combination, durable cashflows enhanced by AI, not threatened by it - is what SingleStep is building.

The investors who will define the next decade are not only those who find the right SaaS winners after the AI storm passes. They are also the ones who recognised, early, that the storm creates a once in a generation opportunity to accumulate assets it cannot touch.

Physical assets are not the old economy. They are the immune system of a portfolio in an age of digital disruption. 

SingleStep builds businesses in both the physical and digital economy - focusing on real estate strategies that generate proven cashflow, and building digital fintech with durable moats leveraging AI.