Two SaaS founders meet at a conference. One has a payments platform, the other a logistics tool. Their customers overlap. They sketch a product on the back of a napkin — a bundled offering that solves a problem neither company can solve alone. They agree to split the build, share the revenue, and figure out the legals later.
Six months in, the product is live, customers are paying, and nobody knows who owns the code. One company has been acquired by a competitor of the other. The exclusivity arrangement was never written down. The “later” never happened.
This is the most expensive way to structure a joint venture: by not structuring it at all. The legal frameworks for co-building products in Australia are well established. Choosing one early — and documenting it properly — is the difference between a partnership that scales into a standalone business and a dispute that consumes both companies.
The Three Structures Worth Considering
Australian startups have effectively three structures to choose from for a product-building JV. The choice turns on duration, depth of integration, and whether the JV might eventually become its own business.
Incorporated JV. A new Pty Ltd company — call it JV Co — is established and owned by both parties, typically 50/50 but often with a casting vote or independent director to break ties. JV Co owns the foreground IP, hires staff (or takes secondees from each parent), signs customer contracts, and has its own tax file number. This is the right structure where the collaboration is long-term, the product will have its own go-to-market, the parties want ring-fenced liability, and there is any prospect of external funding or a future sale. It is the default choice for two founders co-building a real product.
Unincorporated (contractual) JV. Two existing companies agree by contract to collaborate on a defined project. Each retains its own books, staff, customer contracts, and tax position. There is no new entity. This works for discrete, time-limited projects — a single integration, a co-marketed pilot, a one-off product release — where neither party wants the overhead of forming a new company. It is cheap to set up but quickly becomes messy if customer contracts, IP, and revenue start commingling.
Partnership. A partnership arises automatically under the relevant state Partnership Act — for example, the Partnership Act 1892 (NSW) — wherever two parties carry on a business in common with a view to profit. It creates joint and several liability between the parties for each other’s debts and acts within the scope of the partnership. For a startup founder, that exposure is almost always unacceptable. Partnerships should be avoided as a structure, and JV agreements should include an express anti-partnership clause to reduce the risk that the relationship is later characterised as one.
The Deemed Partnership Trap
The anti-partnership clause is not a magic spell. Whether a partnership exists is a question of substance. If two parties jointly derive income from a venture, share profits, and intend to carry on business in common, a court or the ATO can find a partnership exists regardless of what the contract says (see ATO Taxation Ruling TR 94/8).
The safest way to avoid the trap is to structure the cash flows correctly. In a contractual JV, each party should take its share of the product or output separately — not jointly derive income that is then split. In an incorporated JV, all revenue should flow through JV Co and be distributed as dividends or service fees on arm’s-length terms.
The IP Question
For a product JV, IP allocation is the most important commercial issue and the most commonly mishandled. There are two categories that need separate treatment.
Background IP is each party’s pre-existing IP — the codebase, brand, datasets, and know-how each side brings to the table. The default position is that each party retains ownership and grants JV Co (or the other party) a licence. The scope of that licence matters: purpose, territory, exclusivity, sublicensing rights, term, and — critically — whether it survives termination. A licence that terminates with the JV can leave one party holding the product hostage.
Foreground IP is the IP created during the JV. There are three viable models: JV Co owns it outright (cleanest, with a licence-back to parents for non-JV use); one party owns it and licenses it to the other; or the parties co-own it. Co-ownership sounds equitable but creates serious problems. Joint owners of an Australian patent generally cannot grant a licence without the consent of the other co-owner under section 16 of the Patents Act 1990. Joint copyright owners face similar consent requirements for commercial exploitation. The result is a deadlock that destroys commercial value. Avoid co-ownership unless the parties have a clear consent regime and a buyout mechanism.
The agreement should also spell out what happens to the IP if the JV ends — whether the IP is sold to a third party, bought out by one parent, divided by domain, or destroyed. Defaults are vague and adversarial. Pre-agreed waterfalls are not.
The Australian Traps to Watch
Several Australia-specific issues catch out startup JVs:
- Cartel risk. Two parties that compete in any market need to be careful. The cartel provisions in the Competition and Consumer Act 2010 (Cth) (CCA) prohibit price-fixing, market sharing, and output restriction between competitors. The joint venture exception at sections 45AO and 45AP CCA is narrow, untested, and places the burden of proof on the defendant. Where there is genuine overlap, ACCC authorisation or notification is the safer path.
- Mandatory merger control from 1 January 2026. The Treasury Laws Amendment (Mergers and Acquisitions Reform) Act 2024 introduces Australia’s first mandatory, suspensory merger regime. Forming or restructuring a JV that crosses the combined-turnover and transaction-value thresholds requires ACCC clearance before completion. Most pure-startup JVs will sit below thresholds, but a JV with a large corporate parent may not.
- ESS aggregation trap. The startup ESS tax concession in Division 83A-33 of the Income Tax Assessment Act 1997 requires JV Co to be unlisted, less than 10 years old, and have aggregated turnover under $50 million. If one of the parents is a large corporate, JV Co’s turnover may be aggregated under section 328-115, killing the concession and making employee options far less attractive.
- FIRB. If one party is a foreign person, forming the JV can be a notifiable action under the Foreign Acquisitions and Takeovers Act 1975, particularly in critical technology, data, or defence-adjacent sectors where the dollar thresholds drop to nil.
- GST. Most software product JVs do not qualify for the GST joint venture regime in Division 51 of the GST Act (which is largely confined to resources and construction). Each party must account for its own GST, or JV Co (if incorporated) registers and accounts in its own right.
Governance and Deadlock
A 50/50 JV deadlocks easily. Sensible agreements layer multiple mechanisms: a reserved-matters list requiring unanimity for major decisions, tiered escalation through the management chain, mediation for technical disputes, and a deadlock-resolution mechanism of last resort.
The traditional “shoot-out” mechanisms — Russian roulette, Texas shoot-out, Dutch auction — are quick but penalise the cash-poor party. Where one founder is materially better capitalised, these mechanisms are coercive rather than fair. A put-call option at an independently determined valuation, or a triggered wind-up with a pre-agreed IP waterfall, is usually a better fit for early-stage JVs.
What Founders Should Do
Pick the structure before the build starts. The cost of forming an incorporated JV is modest. The cost of unwinding a poorly structured collaboration after the product is live is not.
Document the IP regime in detail. Background IP, foreground IP, what happens on termination. Vague IP clauses are the single most common source of disputes between JV parties.
Run the competition check. If the other party is a competitor in any sense, get advice on whether the cartel provisions are engaged and whether the JV exception genuinely covers the arrangement.
Build a deadlock mechanism for the realistic case. Two founders both believe they will always agree. They won’t. Plan for the day they don’t.
Watch the ESS aggregation rules. If the other parent is a large corporate, the startup ESS concession may be off the table for JV Co staff. That changes how the JV can compete for talent and needs to be priced into the deal.
The Bottom Line
Joint ventures are a powerful way for Australian startups to extend their product reach without the time and capital cost of building everything in-house. Done well, they can create a standalone business worth more than either parent. Done badly, they create disputes that consume management bandwidth and destroy goodwill on both sides.
The structural choice is not difficult. The documentation is well understood. What founders most often skip is the conversation about what happens when things change — when one party is acquired, when a customer wants exclusivity, when the IP becomes more valuable than the venture. Have that conversation early, and reduce it to writing. Everything else follows from there.
