← Back
The number was on slide nine, and Tessa read it before her CTO reached it.
$7.8 million.
She had built strata management software for ten years. Levies, trust accounts, meetings, by-law records: if a strata manager did it, her platform ran it. The company had been profitable for seven years. It had never raised a dollar of outside capital.
For ten years, that had been a point of pride. On slide nine, it became a constraint.
In a good year the business produced about $2.4 million of free cash flow. The rebuild would take a little more than three years of it.
Josh, her CTO, had spent six weeks on the plan. It was careful and it was honest. Rebuild the core of the product so that AI did the first draft of most of the work: summarising a by-law dispute, drafting the minutes, preparing arrears notices, flagging trust account discrepancies for a manager to approve. Twenty months. Two products running side by side for most of that time. Every customer migrated.
The reason for the plan was on slide four. A company that had not existed three years earlier, funded by venture capital from Sydney and San Francisco, had demonstrated its product to Tessa’s second-largest customer in March.
The customer had not left. The customer had asked Tessa, politely, what her roadmap looked like.
Tessa’s first thought was that the company could not afford the plan. That was not quite true. It could afford it once, by putting three years of earnings into a single bet with no margin if it ran late.
Her second thought was worse. It could not afford not to.
Somewhere in the last eighteen months, a software question had become a capital question.
Strata management was not disappearing. That was never the problem.
Buildings would still hold annual general meetings. Levies would still be struck, collected and chased. Trust money would still have to reconcile to the cent, under rules that differ from one state to the next. What was changing was who, or what, did the work, and what the software beneath it had to look like.
If you run vertical software in Australia, the same is probably true of your market. The work is not going away. The question is what the next version will cost, and who can afford to build it.
Tessa pushed back on the number. AI makes software cheaper to build. Her own engineers were shipping faster than they had two years earlier. A small team can increasingly produce what once needed a much larger one. So why $7.8 million?
Because the code was only part of it. AI has reduced one part of the cost of rebuilding software. It has not removed the cost of moving an installed business from one architecture to another.
Ten years of customer data to migrate without losing a cent of trust money. Hundreds of users to retrain, most of them content with the old screens. Security and privacy obligations for software that now drafts and recommends on the customer’s behalf. Computing costs that land in gross margin every month the product is used. And for twenty months, two products to run instead of one.
AI made the code cheaper. It did not make the crossing cheaper.
While Tessa ran two products, the company on slide four would be running one.
The fear was shared. The options were not.
In February 2026 software shares sold off sharply as investors asked what would happen when AI began doing work that application software had traditionally organised. Salesforce was caught in it. Later that month Workday fell to its lowest level in more than five years, after a soft forecast deepened the same concern.
By September, much of the market loss had been recovered. The question had not.
What separates the giants from Tessa is not exposure. It is what they can do about it.
Salesforce had already moved. In May 2025 it agreed to buy Informatica, a data management company, at an equity value of about US$8 billion. Its chief executive explained why in one sentence: “You have to get your data right to get your AI right.” In November, Workday completed its purchase of Sana, an AI company, to make Workday “the new front door for work.”
Both purchases came from balance sheets large enough to absorb them without putting the company at risk. Whatever their engineers might have built on their own, their owners had an option Tessa did not: to buy capability and time with money they already had.
The giants were not spared the fear. They were spared the funding question.
Tessa’s board papers assumed one owner. Put slide nine in front of a different one and the arithmetic changes.
Imagine a larger property software business, already serving several thousand strata, building and property management firms in more than one country. It has an AI platform in production. Its engineers have already moved a dozen acquired products onto it. The security reviews, privacy frameworks and computing contracts are in place. Its balance sheet is many times the size of Tessa’s company.
For Tessa, the crossing is a single bet: three years of cash, twenty months with little room for delay, and most of her own wealth riding on one project.
For the larger owner, it looks different. Integration is never trivial, and moving customers, data and workflows onto a common platform can become a large programme in its own right. But much of the capability has already been paid for, and the rest is spread across many products. No single project can threaten the owner’s survival.
Same customers. Same underlying work. Two very different costs of carrying them into the next product cycle.
The asset has not changed. What may have changed is who is best placed to own it.
Nothing in Tessa’s numbers would show it. Her customers had not left. Her results were good. The mismatch can open up before any deterioration appears in the standalone business, and where another owner values what she has built highly enough, the difference can itself create value.
The only question is who captures it.
A company can be undercapitalised on its own and strategically valuable inside someone else, at the same moment.
So what would that owner actually be paying for?
Partly the product. Issue 12 made the point that a buyer takes on a codebase it must own and maintain, and architecture still shapes what it will pay. But in a crossing like this one, the buyer is buying a good deal more than the code.
Two hundred and ninety strata management firms across New South Wales, Victoria and Queensland, all trusting Tessa’s platform with other people’s money. Workflows embedded so deeply that changing them means retraining every manager who uses them. Integrations with banks and payment systems. Three states, three sets of rules, and the regulatory knowledge built into every screen to handle them.
And ten years of operating data about how strata schemes are really run, together with the contractual right to use it. In an AI product, that right matters as much as the data itself. Data a company cannot lawfully put to work is not something a buyer can build on.
An entrant may be able to build a capable product surprisingly quickly. It cannot manufacture ten years of customer trust or regulatory history at the same speed. That was part of the logic Salesforce spelled out when it bought Informatica: get the data foundation right before asking AI to work on top of it.
None of this means the value is fixed. As the company on slide four gets better, Tessa’s customers look less certain, and buyers pay less for revenue they are less sure of keeping. But a strategic buyer that needs distribution may now pay more for an installed base than it would have two years ago.
The window can move either way.
Tessa did not know which way her own window was moving. She had never looked.
Tessa could find the money in three places.
The first is her own cash. No dilution, no new owners. Also the slowest option and the most concentrated risk. Three years of free cash flow go into one bet, the distributions Tessa had come to rely on stop, and the rebuild moves at the pace the business can afford. Tessa would be funding the rebuild from customer receipts. The company on slide four was funding it from investors who had already agreed to fund losses while it crossed.
The second is outside capital. Faster, but priced at a difficult moment, because anyone funding application software today asks the question the market asked in February. The uncertainty is priced somewhere. Debt prices it through interest, security and covenants. Equity prices it through valuation, dilution and sometimes preference.
The third is someone else’s balance sheet. An owner for whom the crossing is cheaper acquires some or all of the company and funds the next version itself. The founder gives up some or all of the ownership and control of what comes next, and is paid for what already exists.
None of the three is free, and each can be the right answer. Plenty of founders will fund the rebuild themselves and be right to.
Waiting is not a fourth option. It is the first, taken slowly, while the company on slide four takes the second.
But the third option matters even to a founder who never takes it, because it puts a number on what the other two require her to forgo. If the company is worth materially more to an owner who can carry the crossing cheaply, then spending $7.8 million to build it independently is also a decision to forgo that value for now, in the expectation of creating more. That may be the right call. It deserves to be made with the number in view.
Otherwise the founder is not choosing to keep the company. She is keeping it because the alternative was never priced.
Tessa did not decide anything at that board meeting, and she was right not to.
What she did was change the question. Josh had asked the board to approve a $7.8 million rebuild. The real question sat underneath it: who should pay for the next version of this company, on what terms, and how soon.
Two parts of that could be answered from inside the company. Management could model how fast the business could fund the crossing itself, and what that would mean for cash, risk and pace. It could test what outside capital would cost, and what would be left of the share register afterwards.
The third part could not. Who would value the customers, workflows, data and trust Tessa had built, what the crossing would cost them, and what that difference was worth: none of it was visible from slide nine. It could only be found by looking outward.
That night Tessa opened a blank spreadsheet and typed three headings. Our cash. Investors. A buyer. She could fill in the first two by the weekend. The third column stayed empty.
Slide nine had carried one number. The one Tessa needed next was not on it.
She had spent ten years building a company that strata managers could not run their businesses without. Its next version was going to be paid for by someone. That part was already settled.
What was not settled was whether Tessa would choose who, or whether the company on slide four would choose for her.
Five questions, worth answering before the rebuild plan reaches the board, not after.
Salesforce, Inc., “Salesforce Signs Definitive Agreement to Acquire Informatica”, 27 May 2025, and “Salesforce Completes Acquisition of Informatica”, 18 November 2025, including the quotation attributed to Marc Benioff. Workday, Inc., “Workday Completes Acquisition of Sana”, 4 November 2025. Reuters, “Global software stocks hit by Anthropic wake-up call on AI disruption”, 4 February 2026, and Reuters reporting of 25 February 2026 on Workday’s forecast and share price. Subsequent recovery based on published share price history for Salesforce and Workday to September 2026.
Founders and transactions described in Founder Thesis are composites drawn from recurring patterns in software M&A. The larger acquirer in Section 04 is illustrative. Named companies and announced transactions are matters of public record.
Before a rebuild plan reaches the board, it helps to know which parts of the business a buyer would find hardest to recreate, and which it would discount. Nexxit helps founders see the company the way a well-capitalised acquirer would. The Assessment is the entry point to Nexxit, Cube Capital’s acquirability programme.
Begin the free assessment at nexxit.ai
Sometimes the right answer is to fund the rebuild yourself, or to raise for it. Cube Capital works only on the sell side, so its role begins where a founder wants to understand what the business is worth to an owner who could carry the next version more cheaply. That value is worth understanding whichever way the decision goes.
↓
The Rebuild You Cannot Fund
Sell-side technology M&A advisory. Sydney, Australia.
© 2026 Cube Capital Pty Ltd. All rights reserved.
Cube Capital provides M&A advisory services to wholesale clients only. This website does not constitute financial product advice. Advisory services are provided in accordance with applicable Australian financial services law.