‘Generative artificial intelligence (AI)’ has emerged as a competitive variable within the cloud market. Generative AI is re-energizing the cloud market, which has been slowing growth.

As the expansion rate of cloud service sales this yr is predicted to drop to half of last yr, the three major cloud corporations including Amazon Web Services (AWS), Microsoft (MS), and Google are introducing generative AI technology one after one other.

Analysts expected the expansion rate of the cloud service market this yr to be 18%, half the extent of last yr. In truth, AWS’ revenue growth last quarter was 20%, the bottom ever.

As a way to reply to this slowdown in growth, the plan chosen by the three cloud corporations is to focus on corporations using generative AI. We expect generative AI to be the catalyst for brand new cloud growth.

Thomas Kurian, CEO of Google Cloud, emphasized that “AI will open up tremendous market opportunities,” and emphasized that “it continues to be in its infancy.” AI will probably be the catalyst for the cloud business, as many systems will probably be moved to the cloud.”

Customer reactions don’t appear to be different. Accounting firm KPMG said spending on cloud computing has declined in recent quarters, but spending on AI services is more likely to increase after Microsoft’s Azure OpenAI service.

One other potential customer for the cloud is a small AI startup. It takes loads of computing resources for these corporations to develop and run AI-based applications. Enterprise capitalist Andreessen Horowitz estimates that 10 to twenty percent of the revenue that startups make from generative AI apps goes to cloud service providers.

AWS plans to supply as much as $300,000 of free computing resources to pick out corporations as a part of a support program for generative AI startups. Google can also be planning to supply free credits price as much as $250,000 to latest AI startups.

The more corporations create and utilize generative AI services with specific cloud technologies, the more the cloud service provider can secure stable customer demand and increase its market share in the long term.

For that reason, cloud service providers are currently specializing in investing in generative AI startups or expanding alliances.

In January, MS invested $10 billion in OpenAI, exclusively supplied Azure cloud services to OpenAI, and OpenAI’s ChatGPT became a signboard service within the Azure environment.

In February, Google invested $300 million in Antropic, a startup that competes with OpenAI’s ChatGPT, and supplied Google Cloud services to Antropic, and in May, it launched Runway, an AI startup that permits users to create videos from text descriptions. invested about $100 million in Runway has been using AWS as a cloud service, but it surely is predicted that it’s going to use Google’s cloud service with Google’s investment as a possibility. As well as, ‘Vertex AI’, which provides its own large-scale language model (LLM) ‘PaLM 2’, and AI chatbot ‘Bard’ will probably be installed on Google Cloud.

AWS provides cloud services in partnership with Stability AI, an image-generating AI developer, in November last yr. As well as, it has entered right into a partnership with the startup Hugging Face and hosts the LLM of Hugging Face on AWS.

As well as, AWS launched ‘Bedrock’, an enterprise cloud service that gives AI models developed by AWS in addition to AI models from startup partners. It offers its own LLM referred to as ‘Titan’.

Google and AWS even have close ties with Toronto-based Coher and AI21 Labs.

Oracle, the fourth-largest cloud service provider in the US, also recently participated in a $270 million investment round by generative AI model developer Coher.

Koher is an AI startup founded in 2019 by researchers from Alphabet and is referred to as a competitor to OpenAI because it has developed a conversational AI model just like ChatGPT. Particularly, it’s attracting attention for its approach focused on targeting the company market, claiming to be an LLM specialized in corporations.

It shouldn’t be known exactly how much Oracle invested, but industry experts predict that Oracle will strengthen its cloud business through generative AI technology.

An official aware of the industry situation said, “The fact of Oracle is that it cannot survive with DB alone now.”

Oracle currently has no AI model to talk of. For this reason there may be an evaluation that a reversal will probably be made through Coher, which doesn’t give preferential treatment to specific cloud service providers. For Oracle, a latecomer as a cloud service provider, incorporating such a model is taken into account a better method to increase market share than developing its own technology.

Indeed, riding on the generative AI boom, Oracle said it performed well last quarter as demand for its AI tools increased. “In point of fact, demand for AI processing is bigger than available capability,” Oracle chairman Larry Ellison said on an investor conference call.

Meanwhile, Salesforce, an organization specializing in customer relationship management (CRM), also joined the competition within the cloud market with its AI cloud service for businesses. Using its greater than 150,000 corporate customers as a stepping stone, it plans to distinguish itself with business-to-business transactions that concentrate on corporate customers quite than individual customers.

Salesforce has launched an ‘AI Cloud’ service that gives AI models, especially text generation models, from various partners, including AWS, Antropic, Coher, and OpenAI. Salesforce AI models are also available, supporting functions corresponding to AI code generation to business process automation. It’s a type of providing cloud services in order that corporations can develop their very own chatbots based on generative AI. Alternatively, customers can bring custom-trained AI models to the AI ​​Cloud while storing data on their very own infrastructure.

Salesforce’s AI cloud is comparable to ‘Bedrock’, an enterprise cloud service launched by Amazon. As well as, by providing security services with enhanced personal information protection, cloud access has been increased for corporate customers who’re concerned about leakage of sensitive information.

Recently, HPE, which sold hardware corresponding to computer servers, entered the AI ​​cloud market with its Exascale Cray supercomputing infrastructure. HPE has launched ‘GreenLake’, a latest cloud service that exclusively supplies German startup Aleph Alpha’s LLM ‘Luminous’ from supercomputer infrastructure.

Customers can have access to pre-trained LLM ‘Luminous’ and might use their very own data to coach and fine-tune custom AI models. The service allows customers to construct various sorts of AI applications and integrate them into their business workflows.

In this fashion, if synergistic collaboration results are visualized by combining AI startup technology and cloud infrastructure to expand the bottom of the cloud ecosystem, the shape of collaboration is predicted to turn out to be more solid.

Particularly, competition for generative AI technologies and services that may intensify is predicted to have a major impact on the cloud market, which is the core business of those big techs and is directly related to their performance. Amazon CEO Andy Jassy said, “Truthfully, 6-9 months ago, generative AI wasn’t so attractive, but it surely’s gotten higher and higher because it gets much larger and faster,” but added, “It’s rare to comprehend just what number of latest cloud businesses the approaching flood of generative AI will create over the following few years. He expressed his willingness to take a position in the sector of generative AI.

Through partnerships, cloud providers can increase their market share by securing stable customer demand, and AI startups can secure the computing power needed to operate AI models. Particularly, the boom in generative AI highlights the importance of large-scale, advanced cloud computing infrastructure to run large-scale AI models.

It is thought that OpenAI needs billions of dollars to maintain ChatGPT running, and that’s the reason it has joined hands with Microsoft.

As such, the explanation why AI startups form partnerships with cloud corporations is ‘money’ in spite of everything. It requires a big data center and a graphics processing unit (GPU), and requires an unlimited amount of electricity.

For that reason, it’s difficult for generative AI startups to refuse investment offers from cloud corporations, corresponding to providing computing infrastructure. As a way to maintain this relationship, startups surrender their shares and use the cloud cheaply, but there are concerns that they might find yourself within the arms of cloud corporations.

Since running an LLM with billions of parameters requires a stable computing infrastructure, it’s difficult to transfer to other platforms once model training has begun.

Jonathan Frankl, co-founder of MosaicML, also said: “Partnerships like this leave AI startups with little selection but to leap into the arms of a cloud company that may provide essential cloud computing at a reduction and access the huge amounts of capital it needs.” identified.

It’s explained that the mixture of a generative AI startup with Big Tech, which provides cloud services, is an inevitable selection to proceed generative AI research and services that require high-performance computing capabilities.

Such deals will not be formally exclusive, but in practice make the startup’s technology or product depending on the cloud platform. In consequence, cloud conglomerates are turning into a proxy war available in the market over the initiative of generative AI of affiliated startups.

Meanwhile, there are growing concerns that this type of integration will limit opportunities for generative AI startups by cloud giants because of the undeniable fact that the core service provider can also be a competitor. In consequence, the voices of startups criticizing the cloud giants are also heard.

Jonathan Frankl, co-founder of MosaicML, accused cloud giants of “loving monopolies and forcing startups to sign massive, multi-year contracts.”

George Washington University professor William Kovacic identified that “one of these merger is a typical sort of monopoly that antitrust regulators are watching.”

Reporter Park Chan cpark@aitimes.com