On behalf of the Information Technology and Innovation Foundation (ITIF), thank you for the opportunity to provide comments to the Competition and Markets Authority (CMA) on the Microsoft-OpenAI partnership.[1] ITIF is an independent, nonprofit, nonpartisan research and educational institute that has been recognized repeatedly as the world’s leading think tank for science and technology policy. ITIF’s mission is to formulate and promote policy solutions that accelerate innovation and boost productivity to spur growth, opportunity, and progress.

These comments make three points. First, the Microsoft-OpenAI partnership does not meet the criteria set for a “relevant merger situation” as set forth in the Enterprise Act 2002, and thus the CMA should not intervene. Second, even if the CMA believes that the partnership reaches the threshold to create a relevant merger situation, the partnership does not have a negative impact on competition. Third, any potentially negative impacts on competition would be justified by the pro-competitive impact of this partnership.

The CMA Should Not Intervene In the Microsoft-OpenAI Partnership

The first question CMA is considering is “whether the partnership, including recent developments, has resulted in the creation of a relevant merger situation.”[2] The Enterprise Act 2002 plainly states that “a relevant merger situation has been created if…two or more enterprises have ceased to be distinct enterprises…”[3] In the Microsoft-OpenAI case, this criterion is unmet as these two firms remain distinct. Both organizations continue to operate independently, maintaining distinct governance and operational structures. Most importantly, Microsoft does not have corporate control over OpenAI. The partnership functions as a strategic collaboration, not as an integration of business operations, which stands at the heart of the concept of a merger.

While all the details of the partnership between the companies are not public, consider the facts that the press has publicly reported publicly. Both before and after the recent ouster and then rehiring of OpenAI CEO Sam Altman, as well as the subsequent changes in OpenAI’s governance, Microsoft has not had a voting seat on OpenAI’s board.[4] Indeed, the fact that OpenAI’s board fired its CEO without any prior discussion with Microsoft highlights the independence of the two organizations.[5] After the recent board reorganization, Microsoft has gained a non-voting observer seat on the board, presumably so that it would not be blindsided by any future board-initiated changes, but it still does not have any direct corporate control over OpenAI. As Microsoft President Brad Smith wrote, “Since 2019, we’ve forged a partnership with OpenAI that has fostered more AI innovation and competition, while preserving independence for both companies. The only thing that has changed is that Microsoft will now have a non-voting observer on OpenAI’s Board…”[6] In addition, while Microsoft has reportedly made a significant multi-billion dollar investment in OpenAI, a company official issued a statement saying, “Microsoft does not own any portion of OpenAI and is simply entitled to share of profit distributions.”[7] Given these reported facts, the CMA should not consider this partnership a relevant merger situation.

The Microsoft-OpenAI Partnership Does Not Harm Competition

The partnership between Microsoft and OpenAI has not had a negative impact on competition in the UK (or elsewhere) in terms of reduced innovation, consumer welfare, or choice. Importantly, OpenAI and Microsoft are not direct competitors. On the contrary, one of the reasons that Microsoft reportedly chose to partner with OpenAI was to keep up with other tech rivals who at the time had more advanced AI. As the New Yorker describes the situation in a profile of Microsoft CTO Kevin Scott, “Microsoft, over the previous two decades, had tried to compete by spending hundreds of millions of dollars on internal A.I. projects, with few achievements.”[8] Rivals like Google and Meta have been at the forefront of producing advanced AI models (e.g., large language models that OpenAI used to produce ChatGPT) over the past few years. Microsoft’s partnership with OpenAI did not reduce competition between the two firms because Microsoft did not have a comparable rival product.[9]

Likewise, the partnership between Microsoft and OpenAI has not, and will not, prevent rivals from using OpenAI’s AI models or Microsoft’s cloud services. The partnership does not give Microsoft exclusive access to OpenAI’s application programming interface (API) and OpenAI does not have exclusive access to Microsoft’s Azure cloud service. Developers use APIs to build new products and services. OpenAI began making its API publicly available to interested parties in June 2020 and, as of July 2023, anyone could access its most advanced model to date, GPT-4.[10] Nor has this partnership limited competition by disincentivizing OpenAI from improving the quality of its models. Microsoft and OpenAI launched their partnership in 2019, and since then, OpenAI has released multiple new models and updated existing ones.[11]

Moreover, even if there was some overlap, many other firms compete with OpenAI and Microsoft. The Microsoft-OpenAI partnership does not limit the ability of rivals to innovate and offer competing services. Established tech companies such as Meta, Alphabet, and Amazon as well as startups like Anthropic, Stability AI, and Mistral AI, all compete with OpenAI with robust foundation models—AI models like GPT-3 and GPT-4 which developers can use in many different applications.[12]

In fact, Microsoft Azure provides access to competing models, including native support for products such as Meta’s Llama 2, a direct competitor to OpenAI.[13] Similarly, many companies compete with Microsoft’s Azure AI platform, including major cloud providers such as Amazon Web Services (AWS), Google Cloud, and IBM Watson. For example, developers can use AWS SageMaker and IBM Watson Labs to build, train, and deploy AI models either directly on these platforms or through development platforms like HuggingFace.[14]

The Microsoft-OpenAI Partnership Has Had Pro-competitive Effects

Far from reducing competition, the partnership between Microsoft and OpenAI has proved to be pro-competitive and is likely to bring additional consumer benefits going forward.

Microsoft provided the resources for OpenAI to become the company it is today. OpenAI needed outside support when it was in the process of developing its foundation models because of the substantial computing resources necessary to train and deploy AI systems. Microsoft’s investment in the company, both through its financial contribution and technical and legal support, created a pathway for OpenAI’s current product roadmap and success. As reported in the New York Times, “Microsoft developed and provided the vast computing power that runs OpenAI, and negotiated a slate of legal and commercial deals to protect it if something went wrong there.”[15] Without Microsoft’s partnership, OpenAI’s innovative AI models likely would not exist as they do today.

In addition, the partnership between the two companies has allowed Microsoft to quickly bring AI tools to the masses by integrating the technology into its existing suite of products, including Bing, Office, and GitHub.[16] These integrations allow users to combine OpenAI’s large language models with their own data in key business applications such as Microsoft Word, Excel, Outlook, and PowerPoint, have conversations with an AI chatbot in Bing and Edge, or use Copilot to write code more efficiently.[17] Microsoft also provides another pathway for developers to use OpenAI’s AI models. Developers can choose between accessing OpenAI’s AI models through the company’s own API or using Microsoft Azure cloud services to access OpenAI’s AI models.[18] The pricing differs between these options because using Microsoft’s Azure OpenAI Service offers additional integrations with Microsoft cloud services and features, such as Microsoft’s security and access control tools, enterprise training and support, and enterprise-level service level agreements.[19] This rapid product innovation has greatly benefited consumers and businesses and would not have occurred without this partnership.

Finally, the partnership between the two companies has also resulted in innovations that have improved the quality of the AI systems. [20] In particular, Microsoft’s expertise in AI safety and OpenAI’s expertise in AI alignment has allowed them to deploy safety guardrails that make it feasible to deploy OpenAI’s AI models in corporate settings.[21] These innovations in AI safety, which both improve the quality of the AI models and enable downstream users to more widely deploy the technology, would likely not have occurred without this partnership.

Conclusion

The Microsoft-OpenAI partnership came at a crucial time when the future of foundation models was uncertain. In hindsight, it has provided the catalyst for the current wave of innovation and competition in generative AI and did not compromise the ability of other market players to compete. And there is no evidence that this partnership presents a threat to competition in the future.

Given that this strategic partnership does not create a relevant merger situation, does not harm competition, and has a pro-competitive impact on the market, the CMA should not intervene. Intervention would have a deleterious impact on digital innovation not only in the AI sector but the information technology sector more broadly by discouraging partnerships. Many tech companies both provide a technology platform (i.e., a technology used as a foundation for building and running other applications, processes, or technologies) and use other technology platforms. Partnerships between tech companies are common because of the rich integrations and interdependencies between different technologies. For example, a payment provider may use a list of technology platforms, including a cloud provider, a network security provider, and an ad network to deliver its services, while those services may in turn use the payment provider to process payments on their platforms. As discussed previously, these partnerships can be advantageous for all parties.

Moreover, the CMA should not intervene in this partnership given that it involves two U.S.-based companies. The CMA should be cautious when considering extraterritorial enforcement of its authority, especially in hotly contested emerging markets like AI, because intervening could create the impression that the UK is attempting to stifle foreign competition by sabotaging successful technological partnerships. Instead, in the absence of evidence of harm and any clear impact on the UK market, the CMA should defer to U.S. regulators to consider whether to investigate any antitrust concerns arising from this partnership.

Thank you for taking this analysis into account in your review.

Endnotes