Экономист Марьяна Мазукаато призывает государство пересмотреть модель распределения выгод от технологических инноваций

In 2009, the U.S. Department of Energy (DoE) issued nearly identical loan guarantees to Tesla ($465 million) and Solyndra ($500 million), a solar panel manufacturer. While Tesla succeeded, Solyndra declared bankruptcy just two years later, attracting intense public scrutiny. The US government suffered the public backlash and cost from Solyndra’s failure but received nothing from Tesla’s success. Under the original agreement, the government would only receive three million shares if Tesla failed — a counterintuitive arrangement, since public stakes in struggling companies rarely benefit taxpayers. A more effective approach would have granted the government shares only if Tesla succeeded, allowing taxpayers to benefit from the company’s growth. Tesla’s share price, which was around $9 in 2009, then ranged between $258 and $479 in 2013, meaning that a public stake could have generated substantial funds to cover Solyndra’s losses and support future investments. The cases of Tesla and Solyndra highlight how, by not recognizing its role as a public venture capitalist, the state often ends up socializing risks while privatizing rewards. Beyond taking equity stakes, the government can incorporate conditionalities into its contracts as a pre-distribution tool to ensure that societal benefits are built in from the start. A clear example is the US 2022 CHIPS and Science Act, which aimed to bolster domestic semiconductor manufacturing by providing approximately US$53 billion in incentives for research, development, manufacturing and workforce development (CHIPS stands for ‘Creating Helpful Incentives to Produce Semiconductors’). The Act was designed to expand production and shape how the benefits of public investment are distributed — diversifying manufacturing locations, strengthening supply chain security, creating jobs, driving innovation, and promoting resilience and inclusivity. Access to CHIPS funding came with conditions that influenced value distribution upfront. For example, the Act prohibited government funds from being used for stock buybacks or dividends and favoured applicants committed to avoiding buybacks more generally. Recipients with projects exceeding US$150 million were required to share profits with the government above a certain threshold and provide plans for childcare for facility and construction workers. Proposals also had to align with programme priorities such as workforce development, R&D investment and energy- and water-efficient supply chains. It is interesting to note that some, especially those who adhere to the “abundance” agenda, claim that such conditions can impede investment and innovation. The abundance critique is that there are abundant opportunities to grow being held back by excessive regulation, planning and conditions of the sort mentioned above, which can create inertia in the system. But this view can also exaggerate the problem. All contracts are designed. The challenge is to include in their design conditions of reciprocity. If there are too many conditions, of course, that can push the system overboard and stifle entrepreneurship. But the opposite extreme just leads to deregulation and markets that only increase profits, and extractive behaviours. Indeed, there are plenty of examples where regulation and conditions are precisely the factors that stimulate innovation, because they require businesses to invest and innovate in order to receive subsidies or incentives. One could argue that conditionalities could go further than they did with CHIPS. While the rules aimed to boost production and create better-quality jobs, they applied only to certain sectors, limiting their pre-distributive impact. For example, minimum wage standards applied only to labourers and mechanics, not all categories of workers. Similarly, although proposals required commitments to community engagement, recipient firms were not required to give community stakeholders a formal role in decision making. CHIPS could have strengthened its pre-distributive effect by mandating worker representation on company boards and implementing agreements protecting the right to organize. Such considerations are critical today when advances in artificial intelligence (AI) are making a few companies very rich, notwithstanding that significant public investment powered early developments in AI, particularly in speech recognition, along with neural networks. Government agencies like DARPA were not just passive funders of basic research — they actively shaped the direction and pace of innovation. In the 1970s, DARPA’s Speech Understanding Research programme funded pioneering work at Carnegie Mellon University, MIT and Stanford Research Institute, leading to systems like systems that achieved near-human levels of recognition for a 1,000-word vocabulary. Throughout the 1980s and 1990s, DARPA continued to invest in language and learning technologies, laying critical groundwork for the voice assistants now found today in phones, cars and smart speakers. These investments were not immediately commercial, but they built the high-risk, long-term infrastructure that private firms later capitalized on. Far from being peripheral, the state catalysed the AI revolution by taking early bets on technologies that would only become commercially viable decades later. Since taxpayers bore the early risks that made today’s AI breakthroughs possible, it is both reasonable and necessary to ensure that AI systems serve broad societal goals rather than narrow corporate interests. This could involve establishing public missions to guide AI development — such as advancing healthcare, climate action or democratic participation — while embedding public value conditions into funding, procurement and regulation. Publicly funded institutions could also retain a stake in the value created, whether through equity, licensing or public data ownership. In this way, AI’s trajectory can be shaped not just by profit incentives but by deliberate, democratic choices about the kind of future we want these technologies to build. We should challenge the outdated economic paradigms that have led us to believe that we are purely self-interested, that value is created by only a few individuals and that the best a government can do is fix market failures. A big part of this is helping governments develop their capacity and capabilities to solve problems by setting ambitious goals and tackling them in a collaborative way, with attention to how all policy levers can be redesigned in an outcomes and mission-oriented way.  We must become intentional about creating better economic relationships that enable us to work together in pursuit of the common good. The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune. Excerpted from  The Common Good Economy: How to Make Capitalism Work for Us All. Copyright © 2026 by Mariana Mazzucato. Available from Basic Venture, an imprint of Hachette Book Group, Inc. Exclusive: In a new sit-down interview with Fortune, OpenAI CEO Sam Altman explains safety standards are "not at a place" to push AI capabilities much further and warns AI beyond human control is "absolutely" possible. Watch or listen here.