Nine out of Ten Industrial Partnerships with Technology Firms Missed Their Targets. AI Raises the Risk Further

On August 14, BCG published an analysis of technology partnerships that begins with an uncomfortable figure. In a 2024 survey, nine out of ten collaborations between industrial and technology companies were not meeting their objectives. The research included 209 companies from the automotive, healthcare, consumer-goods, and other industries, as well as another 93 technology companies. Even so, 94 percent of respondents expected the number of such partnerships to increase. The reason is understandable. A carmaker, pharmaceutical company, or bank may have detailed knowledge of customers and its industry, but it cannot economically reproduce the development of the most advanced models and infrastructure. A technology company has the opposite advantage: computing capacity and software, but often lacks knowledge of the regulation, operations, and commercial conditions of a specific sector. A partnership is intended to combine these capabilities. The results so far, however, show that complementarity alone is not enough. Autonomous AI complicates the dispute further. A system no longer merely prepares recommendations; it may act, negotiate, and execute transactions on behalf of the company. If a jointly developed agent makes an incorrect purchase or uses data without authorization, it must be clear whether responsibility lies with the technology provider, the industrial partner, or a joint venture. Ownership of data, resulting models, software code, and knowledge gained from operations is equally important. BCG distinguishes four basic types of collaboration according to their objective. A company may seek a partner for its own transformation, jointly develop a technology product, establish a new venture, or use a technology partner to enter another market. Each option requires different economics, ownership arrangements, and decision-making mechanisms. Problems arise when partners use the same contractual model for projects with entirely different strategic purposes. Recurring problems include misaligned incentives, unclear rights to data and intellectual property, and different speeds of decision-making. A technology company may change its product every few weeks, while a regulated industrial company may need months to approve a change. If the contract does not account for this asymmetry, one partner sees the other as too slow while the other sees the first as uncontrollable. For leadership, it is therefore not enough to announce a strategic partnership and create a joint steering committee. Before signing, each side should document the economic outcome it expects, who will finance development, who will capture value if the project succeeds, and how the partnership can be ended. The deeper an autonomous system is integrated into a product or operation, the more expensive a later separation may become. Dependence on a technology partner must therefore be valued from the beginning of the relationship. Terms to explain Strategic partnership: Long-term cooperation between two organizations that share resources, risks, or development activities in pursuit of a common business objective. Intellectual property: Rights to technology, software code, models, designs, and other outputs of development. Decision-making asymmetry: A situation in which two collaborating organizations operate with significantly different speeds of approval and change.

Nine out of Ten Industrial Partnerships with Technology Firms Missed Their Targets. AI Raises the Risk Further

The full article is available to users registered at Hard Skills