If you have further insights or thoughts on this article, please contact the author: Xu Wenxin, Risk Control Department, Nanjing Innovation Investment Group Email: xuwx@njicg.com
Against the backdrop of the global wave of artificial intelligence (AI) technology, the business world is undergoing profound transformation driven by algorithms and data. China's AI technology research and industrialization have entered a critical stage of tackling core challenges. The enormous value potential of the AI sector has attracted massive capital inflows.
However, AI enterprises are generally characterized by rapid technological iteration, high professional barriers, and complex application scenarios. While experiencing rapid development, they also face multiple challenges such as unclear commercialization paths, potential valuation bubbles, and uneven technology monetization capabilities. This places higher demands on the professional depth and comprehensive judgment ability required for financial due diligence. Combining the development stages and industry characteristics of AI enterprises, this article summarizes the following key considerations from the financial due diligence process of hardware, software, and service providers for reference:
01 Verify Revenue Authenticity and Evaluate Commercialization Capabilities & Sustainable Profitability
Focus on the Authenticity and Accuracy of Revenue Recognition
AI enterprises adopt diverse business models, including integrated hardware and software sales, licensing/subscription services, customized project development, technical consulting services, and system integration. Revenue recognition methods vary across different sales models. In practice, some enterprises have underdeveloped financial systems and simply recognize revenue at the time of invoicing. For project-based businesses, revenue is commonly recognized upon receipt of customer acceptance reports, without fully considering specific contract terms such as terminal acceptance criteria, division of performance obligations, judgment of transfer of control over goods, post-period returns, and quality assurance clauses.
During due diligence, walk-through tests should be conducted by verifying sales contracts, acceptance documents, invoices, and bank receipts. Combined with bidding information and customer interviews, the authenticity of revenue can be validated. Special attention should be paid to enterprises with seasonal fluctuations, particularly concentrated revenue recognition in the fourth quarter. Post-period collection of accounts receivable and return records should be closely examined, and the rationality of product positioning, target customer groups, and pricing strategies should be analyzed to comprehensively determine whether transactions conform to commercial substance.
Focus on Customer Concentration and Business Sustainability
For early-stage enterprises, attention should be paid to the risks of relying on a single product or customer. Taking manufacturers of core robot components as an example, most are in the customer validation phase with scattered orders or limited customer procurement scale. It is necessary to evaluate customer repurchase intentions, large-scale procurement demand, and the quality of benchmark customers.
For technologically leading enterprises, if revenue mainly depends on policy-supported projects or major special programs in the short term, their ability to convert market-oriented orders and continuously secure major projects should be prioritized, along with an assessment of the potential impact of policy adjustments on operating performance.
In addition, special attention should be paid to the fairness of related transaction pricing and commercial substance, the matching degree between customer qualifications and transaction scale, risks of long-overdue bad debts, and abnormal transactions with customers facing operational difficulties. For abnormal transaction behaviors, it may be necessary to expand the scope of verification and implement in-depth penetration checks to ensure business authenticity and sustainability.
Evaluate the Feasibility of Profit Projections
Based on existing orders, potential business opportunities, and downstream customer demand, cross-validation should be conducted in conjunction with the R&D progress of core products, technological iteration, and commercialization arrangements. This will help determine whether the key assumptions in the company's profit projections are reasonable, as well as the feasibility of achieving projected break-even points and positive cash flow timing.
02 Focus on Supply Chain Security and Analyze Cost Controllability & Cost Reduction Feasibility
Focus on Supply Chain Dependence and Security
Regarding supply chain security risks in the AI sector, such as for integrated hardware and software products, systematic evaluation should be conducted on the supply stability and capacity guarantee of core raw materials, the regional distribution and concentration of suppliers, and the scalability of alternative suppliers. Special attention should be paid to the continuous supply capacity of key infrastructure and raw materials such as computing power resources and high-end chips, as well as the potential impact of overseas policy environments and transaction compliance on the company's supply chain security.
Focus on the Accuracy of Cost Accounting and Feasibility of Cost Reduction
AI enterprises typically have a high proportion of labor costs, with technical personnel often deploying resources in advance before project or product delivery. Due diligence should verify the accuracy of raw material and labor cost allocation, focusing on issues such as insufficient accrual of contract performance costs, inaccurate allocation of production materials, direct recognition of purchased invoices as costs, and the classification and allocation of labor costs between production and R&D. These accounting deviations can distort gross margins and affect the accurate judgment of profitability.
Attention should be paid to the company's bargaining power in the supply chain, analyzing the cost structure and product BOM (Bill of Materials) composition. A systematic evaluation of the feasibility of cost reduction paths—such as economies of scale, product design optimization, and localization of core components—should be conducted to assess the substantial impact of various cost reduction measures on improving product profitability.
03 Judge Technological Advancement and Identify Core Competitiveness
Focus on the Advancement and Independence of Core Technologies
R&D in the AI sector is generally characterized by large capital investment, long verification cycles, and evolving technical paths. During due diligence, interviews with technical teams, downstream customers, and industry experts should be conducted to understand the company's core R&D system and evaluate the advancement and independence of its technologies. It is necessary to check whether there is significant outsourced R&D or technological dependence, verify the source and ownership of core patents, software copyrights, and other intellectual property rights, and analyze the matching degree between R&D investment and output results (such as the number of patents and product iterations).
For model providers, for example, key attention should be paid to whether core algorithms are independently controllable, the rate of independent source code, and the degree of localization adaptation. By comparing technical indicators with mainstream competitors, potential risks of technological substitution or lagging iteration can be identified.
Focus on the Compliance and Accuracy of R&D Expenditure Accounting
The definition of R&D activities, identification of R&D personnel and investment, and allocation of R&D materials and labor costs have always been key focus areas in IPO reviews.
Due diligence often reveals that early-stage enterprises, due to underdeveloped accounting systems and incomplete basic data records, commonly have issues such as overlapping roles between R&D, production, and sales personnel, failure to allocate R&D working hours by project, and failure to correspond R&D material collection to specific projects. These problems directly affect the accurate measurement of R&D expenses and product costs.
Special vigilance should be exercised against companies that artificially adjust profits, inflate intangible assets, and embellish financial statements through capitalization of R&D expenditures. The rationality of capitalization treatment should be carefully evaluated, and by simulating the reversal of capitalized expenditures to expense treatment, the substantial impact of accounting treatment on financial statements can be quantitatively analyzed to provide a reliable basis for valuation judgment.
04 Evaluate Core Team Stability and Effectiveness of Team Structure
Focus on the Integrity and Stability of the Core Team
For AI enterprises, building an R&D team is one of the most important expenditure priorities, and the stability of the core team directly determines the company's technological R&D expansion capabilities and commercialization potential. By reviewing the educational backgrounds, research experience, project or work experience of key core members, as well as their job responsibilities and roles, an analysis can be made to judge the integrity of the team structure, talent stability, and the entrepreneurial commitment and dedication of the senior management team.
Focus on Compensation Packages and Incentive Mechanisms for Core Personnel
A detailed analysis should be conducted on the compensation structure of senior management and technical personnel to assess its alignment with job responsibilities and competitiveness within the industry. For existing or proposed equity incentive plans, the impact of share-based payments on the company's profits should be fully considered. Close attention should be paid to the tenure and turnover of core personnel to identify risks of key talent loss. Additionally, combined with the company's financial position, judgment should be made on whether existing reserves are sufficient to support continuous team development. Through multi-dimensional analysis, the effectiveness of the company's talent management mechanism can be evaluated to ensure deep integration between talent and the company.
05 Information System Verification
If an enterprise's daily operations are highly dependent on information systems, it is advisable to conduct supplementary verification of general information system controls, application controls, and their effectiveness through IT auditing. Attention should be paid to the matching degree between business data and financial data, cross-validation between system data and third-party data, as well as the commercial substance and authenticity of large or abnormal transactions.
Conclusion
Most AI enterprises are in the early stage of entrepreneurship or a phase of sustained high R&D investment, with distinct technology-intensive characteristics. Their financial structure differs significantly from traditional industries, often manifesting as high R&D investment with lagging revenue returns, long technology incubation cycles amid urgent market profit expectations, significant potential value of core technologies with slow commercial implementation, and high dependence on external financing with insufficient self-sustaining cash flow capacity.
Financial due diligence for such enterprises should not be limited to verifying the authenticity of traditional financial data. Instead, based on analyzing historical financial performance, a systematic evaluation of the company's sustainable operation capabilities and valuation rationality should be conducted through multi-dimensional assessment of technological maturity, product commercialization capabilities, supply chain stability, team structure and incentive mechanisms, and capital utilization efficiency. Only through in-depth due diligence that integrates business and financial perspectives can potential risks be identified beyond the technological halo, enabling rational decision-making amid the AI investment boom. This approach will help screen high-quality targets that possess both innovative strength and commercial potential, achieving a win-win outcome for technological innovation and investment returns.
Source: Xu Wenxin, Risk Control Department
Reviewer: Xue Yao
Publisher: You Yi
