In Parts One and Two, we argued that the market's indiscriminate repricing of software companies was wrong — that structural moats built on regulatory depth, proprietary data, and workflow embeddedness cannot be replicated by AI coding tools alone, regardless of how capable those tools become. The argument was primarily defensive: explaining what software incumbents will not lose.
Part Three is about what they stand to gain, and why the timing advantage belongs to them.
The Two Levers Now in Play
The strongest software companies are now pulling two levers simultaneously, which bring both future revenue and cost benefits.
The first is product velocity. AI coding tools are compressing development timelines. Features that once required months of engineering time are being shipped in weeks. For a software business with an entrenched customer base and a deep product roadmap, this change is significant to their potential to develop new revenue streams, and offers the ability to widen the competitive distance with peers. If anything, listening to executives report their financial results across the world these last few months, we’re convinced that software leaders are moving the goalposts faster than challengers can approach them.
The second lever is cost efficiency. Block and WiseTech executive teams recently announced bold headcount changes. For management teams with genuine engineering competency, reaping the cost benefits of new productivity tools — be it across engineering teams or back office operations — is a tangible possibility, if acted upon. This C-suite engineering competency however, is almost prerequisite for overcoming the classic ‘institutional imperative’ — the unseen force within organisations that drives behaviour toward inertia and peer imitation.
Jack Dorsey made the most visible statement in February, reducing Block's headcount by more than 40% while promising to continue growing product velocity. His framing was unambiguous: production code shipment per engineer had increased materially through 2025. Engineering work that once took weeks was being completed by small teams in a fraction of the time. What changed, in Dorsey's words, was that the company could see clearly that a smaller, more capable team paired with AI tools could do more — and do it better.
Most commentators are dismissing Block’s move as corrective action post COVID-era headcount bloat, but the other credible explanation is that he’s merely telling the truth, and acting accordingly.
His sentiments are mirrored by numerous credible executives: META’s Susan Li recently called out power users of AI coding tools seeing 80% output gains, while Zuckerberg noted projects that used to require big teams are now accomplished by a single, talented person; Spotify’s Soderstrom said their best developers haven’t written a single line of code since December; and Salesforce’s Benioff said AI agents are performing 30-50% of all work within Salesforce in mid 2025.
The Productivity Reinvestment Opportunity
The opportunity behind this headcount optimisation is not just a simple margin expansion story for software stocks. It’s an opportunity to unlock budget and recycle it into high ROI initiatives.
Take Block for example.
The first order effect of Dorsey’s headcount reduction is estimated gross savings of over $1 billion.
The likely second order effect, however, is that these savings will be — at least partially — reinvested into high ROI initiatives. In their case, the direct field salesforce to better penetrate Square’s payments ecosystem with larger merchants; or consumer incentives to convert Cash App users to primary banking customers. This adds fuel to sales growth.
If Dorsey’s thesis is correct, by acting fast, they aren’t just stepping up pace of releasing monetisable products, but also their resources to sell them and drive growth.
Racing Up a Steep Innovation Curve
Winners, in this new era, will be those with a long list of client pain points to solve, and hence a lot more innovation to complete and monetise — and a lack of peers with the capability, resources or unique (non-software) assets to do so.
In contrast, a software company with a flat innovation curve — one where its problems for customers are largely already solved — in the cash flow harvesting phase of its lifecycle, is just waiting for peers to catch up and compete for its lunch.
Those with the capacity to shift the goal post faster than competitors can keep up, have a unique opportunity to continually drive new growth.
Let’s stick to the Block example, and consider what this means for the Square business.
Square - From Payment Processor to Autonomous Operating System
Over the last 18 months, Square has consistently increased the pace of new innovations solving real merchant problems. AI Voice Ordering now answers all incoming phone calls for restaurant sellers, handling complex menu customisation, suggesting add-ons, sending payment links, and flowing confirmed orders directly into the kitchen — eliminating the double entry pain in busy service periods. Order Guide standardises vendor prices by unit ($/oz) and auto-converts menus into complete ingredient lists, giving SKU-level cost insights that most small operators have never had access to. A unified POS app consolidated restaurants, retail, and services into a single platform with seven customisable modes.
Block’s Investor Day offered insight on upcoming innovations. ManagerBot is an AI assistant designed to sit across every data layer of a merchant’s business. Dorsey framed it as giving the merchant a personal COO, “We want them to think of this as hiring a manager or hiring a COO to truly operate their business end-to-end.” The vision is proactive recommendations on stock management, automated workflows across payroll, scheduling and inventory, and adaptive interfaces that generate custom dashboards for each business based on what matters most to that operator.
Alongside this, Square shared its intention to build a scaled vendor network that connects sellers together, automates accounts payable and financing through standardised catalogues, and eventually enables group purchasing and negotiated rates across the base. Order Guide doesn’t just give sellers cost visibility — it creates the data layer for a procurement network that would make Square’s merchant base collectively more powerful than any individual operator. That’s a platform strategy, not a feature release.
Square AI could evolve into an autonomous business operator: reordering inventory when stock runs low whilst understanding the propensity for stock error rates; automating staff scheduling based on an anticipated Thursday-night-rush while understanding various staff no-show rates; generating and A/B testing marketing campaigns, and optimising menu pricing based on dynamic supply costs.
The key insight is what this does to switching costs. The longer Square AI is trained on a merchant’s specific data such as trading patterns, seasonal demand cycles, staff behaviour, supplier pricing history, the more valuable it becomes, and the harder it is to leave. A restaurant that’s been on Square for three years would have an AI that knows the business like an experienced COO. Migrating to a competitor means starting that learning curve from zero. This builds competitive advantage via the institutional knowledge the AI has accumulated.
Not Dead Yet
Is Software dead? Hardly. The best software businesses are entering their most productive era. The companies with the steepest innovation curves, the strongest non-replicable assets, and the willingness to act on the cost opportunity in front of them — these are the compounders that hold significant opportunity.
The article has been prepared by ECP Asset Management Pty Ltd (ECP). ECP is a funds management firm based in Sydney, Australia. For further information, visit www.ecpam.com. This material has been prepared for informational purposes only and is not intended to provide and should not be relied on for financial advice. ABN 26 158 827 582, AFSL 421704, CAR 44198.
