Emerging Frontiers of AI Supremacy
Discover the battleground for AI supremacy, where leading chatbot technologies like Google's Bard and OpenAI's ChatGPT are shaped by a strategic fusion of essential components Unveil the limited cloud market, escalating regulatory scrutiny, concerns of critics, and cloud companies' staunch defense of their achievements
Firing up an artificial intelligence chatbot, such as Google's Bard or OpenAI's ChatGPT, means interacting with the culmination of three or four essential components. These include the engineering expertise required to develop the chatbot's AI model, the extensive training data necessary for the model to learn how to respond to prompts, and the use of advanced semiconductor chips for the time-consuming training process.
Cloud platforms are quickly becoming the fourth crucial component in AI, gathering data from high-demand semiconductor chips, offering online storage and various other services. These platforms rent out processing power and storage space to AI companies, fundamentally shaping the trajectory of the AI industry. This places cloud companies at the center of a technology that is set to revolutionize the way people live, work and learn.
Only a few key companies, such as Amazon, Microsoft, and Google, control the massive half-trillion-dollar cloud market. This has raised concerns among policymakers and industry critics, who fear that the power these Big Tech firms hold in cloud markets could have significant and potentially anticompetitive implications for the future of AI.
"I am deeply concerned that a handful of Big Tech firms dominate cloud computing and storage," expressed Massachusetts Democratic Sen. Elizabeth Warren in an interview with CNN. "Without sensible regulations, these companies will further solidify their control over AI, stifle competition, and jeopardize consumer privacy and safety, innovation, and national security. It is essential for us to preserve competition within this crucial industry."
None of the three companies provided a comment for this article. Advocates for the industry argue that the top companies in the sector engage in strong competition for business and that their size allows them to offer comprehensive solutions for all cloud computing needs.
However, as with any crucial industry dominated by major players, policymakers are concerned about the possibility of inflated pricing, anticompetitive behavior, exploitative contract terms, or other practices that could impact the cost of using AI services or the types of products available in the market.
The cloud market is limited to only a few major players
AIs share of total cloud spend is expected to grow substantially as generative AI picks up steam.
The public cloud market is expected to grow significantly, with overall spending projected to increase by over 20% to $679 billion next year, according to market research firm Gartner. Analysts predict that artificial intelligence could represent between 30% and 50% of that total within five years. Cloudflare CEO Matthew Prince noted that only a handful of cloud platforms have the capacity to handle the extensive processing demands of AI companies and their clientele.
In an interview, Prince mentioned that as AI developers and businesses relying on AI shift towards cloud providers, the providers will play a larger role in determining the success of companies in the AI industry. Currently, artificial intelligence makes up a small fraction, possibly less than 10%, of the $563 billion public cloud market, according to Gartner, a market research firm.
The top three providers of AI in the cloud are Amazon, Microsoft and Google by a wide margin, followed by smaller cloud providers including IBM and Oracle.
Regulatory scrutiny at home and abroad is growing
The growing prominence of the industry has caught the attention of governments worldwide. In the United States, both the Federal Trade Commission and President Joe Biden have expressed concerns about competition in cloud markets and its potential impact on AI development.
The FTC warned in June that if a single company or a group of companies dominates the cloud industry or any other critical component of AI development, they could have excessive control over a large portion of economic activity, given that AI is becoming increasingly integrated into everyday life. President Biden also acknowledged these concerns when he issued an executive order in October specifically focusing on artificial intelligence.
The order emphasized the importance of promoting a fair and open market for AI development. It stated that this goal could be achieved by preventing unlawful collusion and addressing the risks associated with dominant firms using key assets (such as semiconductors, computing power, cloud storage, and data) to disadvantage their competitors. The order did not specifically name any providers or allege any misconduct within the industry.
In recent years, the cloud industry has come under scrutiny by the FTC and other global competition regulators, who have conducted multiple comprehensive studies on the sector. For example, the UK government launched an investigation this fall focusing on Microsoft and its software licensing practices, which are largely unrelated to AI. In March, the FTC announced that it was also investigating the industry in collaboration with France, Japan, the Netherlands, and South Korea.
Government inquiries have consistently pointed to a highly concentrated market, according to Sarah Myers West, managing director at the AI Now Institute and former FTC senior advisor on AI.
One reason for the partnership between cloud platforms and AI companies is the continuous demand for cloud resources when their products are used by consumers or businesses.
Large tech companies can use this to gather even more data from businesses and internet users. According to her, their goal is to lead people into their cloud ecosystems in order to profit from these tools.
What cloud critics say they worry about
There are a handful of specific cloud practices that raise potential competition concerns, according to some industry analysts.
AI companies often sign exclusive agreements with one or two cloud providers in exchange for significant investments, unlike other sectors where using multiple cloud providers is more common, according to West.
Many AI startups currently in the market have already established exclusive licensing agreements with cloud providers, according to the speaker. Even smaller players in the industry have connections with a single cloud infrastructure company. For instance, OpenAI's ChatGPT tool, which garnered attention and sparked the current AI excitement, relies significantly on a partnership with Microsoft. Microsoft has pledged $13 billion in cash and free cloud computing usage to support this relationship.
According to Microsoft CEO Satya Nadella in an interview with CNN contributor Kara Swisher, OpenAI's existence is indebted to Microsoft's deep partnership with the company. He also expressed that if needed, Microsoft has the ability to handle all of OpenAI's AI work due to having the necessary resources.
In a similar fashion earlier this year, Amazon made an investment of up to $4 billion in the AI startup Anthropic in exchange for a minority stake and the status of "primary" cloud provider. This came after Anthropic had previously announced a deal to use Google as its "preferred" cloud provider.
Experts caution that offering free cloud credits could potentially result in powerful cloud companies influencing AI companies, leading to a lack of competition and making it difficult for them to switch providers or customize features. "The more we commit to this path, the more challenging it becomes to untangle," noted Steven Weber, a professor at the University of California's Berkeley School of Information, during an FTC roundtable discussion this year.
Some partnership between cloud and AI firms, such as the one between Amazon and Anthropic, can offer big tech companies the chance to acquire stakes in influential AI startups. The UK competition regulator is mulling over a potential antitrust investigation into Microsoft's association with OpenAI in this regard, indicating that OpenAI's recent leadership issues may have resulted in a merger worthy of scrutiny. Microsoft has refuted claims of any ownership of OpenAI, despite gaining a non-voting position on the startup's board.
Another area drawing attention is the fees charged by certain providers for accessing data from a cloud provider.
"It just makes AI more expensive," stated Prince. If AI companies could conveniently and affordably transfer their data from one provider to another, "that would drive the cost of [AI] training down significantly ... my hunch is that you could cut the cost of training probably in half."
How cloud companies defend their record
Cloud providers and their advocates claim that while the cloud market may be concentrated due to the high cost of building cloud services at massive scales, it is still highly competitive. Amazon informed the FTC during its industry study this summer that since the inception of AWS, it has substantially reduced prices for services like computer processing, data storage, and data transfer. They added that competition among cloud and other IT providers is flourishing, ultimately benefiting the U.S. economy.
Google is stepping up its competition for AI cloud customers with the announcement of a new generation of cloud-based computing processors. According to Google, these processors can train large AI models nearly three times faster than the previous generation. As an example, Google used these chips to train its most advanced AI model, Gemini.
Google Cloud stated in a recent blog post that their commitment to being the leading hyperscale cloud provider extends to their AI ecosystem. They emphasized their dedication to helping organizations access and innovate with generative AI and large language models through partnerships, as well as applying evolving AI and ML capabilities to real industry use cases, and developing new applications utilizing these capabilities. According to Microsoft, cloud customers, including AI developers, are well-informed when entering agreements with a cloud provider.
"Service providers engage in extensive negotiations with customers, covering various aspects such as pricing, storage capacity, and contract duration," Microsoft informed the FTC. "For cloud vendors, every project is essentially a sales pitch, as customers seek value and quality. If a customer feels they are not receiving the expected level of service, they will explore other options, driving competitive dynamics within the industry."
Exclusive agreements between AI companies and cloud providers can be beneficial, according to Brandon Jung, vice president of ecosystem and business development at the AI startup Tabnine and an early employee of Google's cloud platform.
Jung mentioned that the advantage is increased efficiency for developers and deployment, as well as the potential for enhanced security by using a single vendor. However, he also believes that it may further integrate people into cloud services.
Critics argue that some of the concerns policymakers have raised are a result of longstanding government suspicion towards app stores, social media, and e-commerce giants.
Several of these happen to be the same companies that operate large-scale cloud computing enterprises, according to an individual familiar with the perspective of major cloud vendors who requested anonymity due to not being authorized to speak publicly.
"There is an overall unease among regulators globally with Big Tech and digital markets, and a belief that the level of dominance in those markets is problematic," the individual stated. "Unfortunately, they have adopted some of those same worries and concerns and applied them to their perception of cloud computing, which serves a very different customer base."
Some AI companies have chosen to avoid forming exclusive partnerships with cloud vendors.
For example, Cohere AI, which offers AI models to businesses rather than individual consumers, stated that it collaborates with all major cloud providers, even if it has resulted in potentially missing out on certain funding opportunities.
"In our most recent round of funding, we made a conscious decision not to accept any single large check," stated Coheres COO Martin Kon. "We didn't want to be beholden to any one individual who might try to dictate our actions."
However, according to West, companies like Cohere are rare exceptions to the norm.
"I think it makes clear that the cloud firms hold a tremendous amount of power in the market," she said.
