Introduction
Mid-market technology acquisitions — transactions valued between $10M and $50M — represent the sweet spot of AI M&A activity in Southeast Asia. These deals are large enough to be transformative for both buyer and seller, but often too small to justify the comprehensive (and expensive) due diligence processes used in billion-dollar transactions. The result is that certain categories of legal risk are routinely overlooked, creating post-closing disputes that could have been avoided with proper pre-signing diligence.
Over the past five years, we have advised on dozens of technology acquisitions in this size range. Here are the five legal pitfalls we see most frequently — and practical guidance on how to avoid them.
Pitfall 1: Incomplete IP Assignments
This is the single most common issue we encounter, and it is remarkably persistent despite being well-documented in M&A literature. In our experience, approximately 40% of AI startups in Singapore have at least one gap in their IP assignment chain. The most common scenarios are: early code written by a co-founder before the company was incorporated (and never formally assigned); contributions from freelance developers or contractors who were engaged without written IP assignment agreements; and jointly developed IP from university or research institution collaborations where the ownership split was never formalized.
Under Singapore law, the default position is that the creator of a work owns the copyright. For code written by employees in the course of employment, Section 30(6) of the Copyright Act 2021 vests first ownership in the employer. But for independent contractors, the default position is that the contractor retains ownership. If your CTO wrote the first version of your core algorithm while freelancing before joining full-time, you need a written assignment. If you engaged A*STAR researchers under a collaboration agreement, you need to confirm that the IP ownership terms give you the rights that your acquirer expects to receive. Acquirers' counsel will check every one of these items, and gaps discovered during due diligence will either delay closing or reduce the purchase price.
Pitfall 2: Change-of-Control Clauses
Material commercial contracts frequently contain change-of-control provisions that are triggered by an acquisition. These clauses give the counterparty the right to terminate or renegotiate the contract upon a change in the ownership or control of the company. In technology acquisitions, the contracts most likely to contain change-of-control clauses are cloud infrastructure agreements (AWS, GCP, Azure), enterprise customer contracts, data licensing agreements, and strategic partnership agreements.
The practical risk is straightforward: if a key customer or data provider can terminate their contract upon your acquisition, the acquirer's commercial thesis may be undermined. We have seen deals where the target company had 60% of its revenue concentrated in three enterprise contracts, each containing a change-of-control clause. The acquirer required consent from all three counterparties as a condition to closing, adding six weeks to the timeline and requiring substantial commercial concessions. The solution is simple: review every material contract for change-of-control provisions at least three months before engaging with potential acquirers, and negotiate amendments where possible.
Pitfall 3: Employee Matters
In AI acquisitions, the team is often as valuable as the technology. Key employee retention is a critical concern for acquirers, and Singapore's employment law framework creates several issues that must be addressed in the transaction structure. First, under the Employment Act 1968 and common law, employees cannot be unilaterally transferred from one employer to another. In a share acquisition (where the acquirer buys shares in the target company), the employment relationships continue uninterrupted because the employer entity remains the same. But in an asset acquisition (where the acquirer buys assets from the target), employees must be offered new employment and must consent to the transfer.
Second, employee stock option plans (ESOPs) are almost universally present in AI startups, and their treatment in an acquisition requires careful planning. Options may accelerate upon a change of control (single trigger), upon a change of control followed by termination (double trigger), or not at all — depending on the terms of the ESOP. The tax treatment of option exercises in the context of an acquisition can also be complex, particularly for employees who are Singapore tax residents holding options over shares in a foreign-incorporated parent company.
Pitfalls 4 & 5: Data Compliance and Earnout Structures
The fourth pitfall is inadequate data protection compliance. Acquirers in 2026 are conducting data protection due diligence as a standard workstream, not an afterthought. If your company processes personal data — and virtually every AI company does — you should expect detailed questions about your PDPA compliance, your data protection officer appointment, your data protection impact assessments, and your breach notification procedures. Companies that have not appointed a DPO or that cannot produce documented consent records for their data processing activities will face valuation adjustments or additional indemnity requirements.
The fifth pitfall is poorly structured earnout provisions. Earnouts are common in AI acquisitions because there is often a significant gap between the seller's view of the technology's future value and the buyer's willingness to pay upfront. However, earnout disputes are among the most litigated issues in M&A — and the disputes are almost always traceable to ambiguous drafting. Every earnout should define the earn-out metric with specificity (revenue, ARR, user count, model performance), the measurement period, the accounting methodology, the seller's right to operate the business during the earnout period, and the dispute resolution mechanism. In Singapore, the most common approach is to use an independent accounting expert for earnout disputes, with the Singapore International Arbitration Centre (SIAC) as a backstop for disputes that cannot be resolved through the expert determination process.