Software Is Now a Commodity. Lean Is the Moat.
For most of software's history, the code itself was a defensible asset. Writing a working accounting system took years, a real engineering team, and domain expertise that took even longer to accumulate. The people who had done it already had an enormous head start over anyone trying to catch up. The software was the moat.
That is no longer true. LLMs have collapsed the cost of writing software to near zero. Not the cost of good judgment, or of knowing what to build, or of understanding the domain. But the raw act of producing working code - integrating an API, building a data pipeline, wiring up a compliance workflow - is now something a two-person team can do in weeks rather than quarters. The software is no longer what separates you from the competition.
What replaced it
If the software itself is commodity, what isn't? A few things remain genuinely hard to replicate: proprietary data pipelines built over time, deep distribution channels, and - the one people underestimate - cost structure.
Cost structure is durable in a way that features aren't. A competitor can copy your feature in a sprint. They cannot copy your headcount situation. If they've built a 50-person team to serve their market, they need enough revenue to sustain it. That creates a price floor they cannot go below without laying people off. You don't have that problem if you never hired the 50 people.
The incumbents are stuck
Legacy SaaS companies didn't build large teams because they were wasteful. They built them because the work required it - engineering, sales, customer success, compliance, support. That was the correct decision in 2018. The problem is that those teams are now a fixed cost the product has to justify, while a new entrant can do the same job with two people and a stack of LLM API calls.
Consider accounting software for Singapore's F&B sector. Xero charges S$30-60 per month and is generic - restaurant owners spend hours reconciling POS exports into spreadsheet-style ledgers that Xero doesn't understand. Counto, a managed accounting service targeting small businesses, charges S$300-600 per month. Their price is set by what it costs to assign a bookkeeper, a chartered accountant, and a tax specialist to each client. They can't go lower without losing money on the engagement.
A two-person team can now build the same outcome - POS data synced automatically, GST calculated correctly, IRAS returns filed via API - for S$120 per month. That isn't a promotional price. It's structural. The running costs for the product (hosting, APIs, LLM inference) sit under S$500 per month at a few hundred customers. Break-even is around 80 paying accounts. Counto cannot match S$120 without dismantling their service model. Xero cannot match it without building F&B-specific logic they have no reason to prioritise for a small regional segment.
The gap isn't closed by working harder or raising a round. It's a permanent structural advantage that accrues to the team that never built the overhead in the first place.
The micro-SaaS thesis
This is what makes micro-SaaS interesting right now. The argument was always that small teams could serve niches that larger companies ignore. That was true but fragile - eventually a well-funded competitor would notice the niche and outbuild you.
The LLM-era version of the argument is stronger. Small teams can now build and maintain the same quality of software as large teams. They can iterate faster because there's no coordination overhead. And they can price below the floor of any competitor with payroll to meet. The niche doesn't have to stay ignored - you can win even if the larger player notices, because they can't match your cost structure without a restructuring that would take years.
The bottleneck has shifted from "can we build it?" to "do we understand the customer well enough to build the right thing?" That's a question two people who care about the problem can answer better than a product team in a large company planning for Q3.
What this means for the incumbents
It isn't a death sentence for large software companies. Network effects, brand, and enterprise sales relationships still matter enormously. But it does mean that the mid-market - the segment below enterprise but above pure hobbyist - is more exposed than it's been in a long time.
The companies most at risk are the ones whose main moat was "we're the only ones who could build this." If that was ever true, it is rapidly becoming less so. The companies least at risk are the ones whose moat is in data, distribution, or regulatory relationships - things that can't be replicated in a sprint regardless of how good the LLMs get.
Software commoditisation isn't new as a prediction. It's new as a practical reality. The teams that recognise it earliest and build around a lean cost structure rather than a feature list will have an advantage that compounds quietly, right up until the incumbents finally notice and find out they can't do much about it.