Generative artificial intelligence (AI)" has become a competitive variable in the cloud market. Generative AI is reinvigorating the slowing growth of the cloud market.
It is expected that the growth rate of cloud service sales this year will fall to half of last year's level, and Amazon Web Services (AWS), Microsoft (MS), Google and other three major cloud companies are successively introducing generated AI technology.
Analysts forecast cloud services market/market growth of 18% this year, half the level of last year. In fact, AWS's sales growth rate of 20% last quarter was the lowest value in history. In response to this slowdown in growth, the company has chosen to generate AI to use the corporate playbook. Expect generative AI to be the initiator of new cloud growth.
Google CEO Thomas Kurian emphasized: "AI will open up huge market opportunities." "It's still very early." JADSON ALTOPS MS COMMERCIAL CEO ALSO PREDICTED: "A LOT OF ENTERPRISES ARE MOVING MORE AND MORE AI-RELATED SYSTEMS TO THE CLOUD, SO AI WILL BECOME A FACILITATOR FOR CLOUD BUSINESS."
The customer response didn't seem to be any different. Accounting firm KPMG said spending on cloud computing has decreased in recent quarters, but the likelihood of increased spending on AI services has increased since using MS's AJE Open AI services.
Another potential customer for the cloud is small AI startups. These businesses require significant computing resources to develop and run AI-based applications. Andresen Horowitz, a venture capital firm, estimates that 10%-20% of the revenue raised by startups to generate AI applications will go back to cloud service providers.
AWS is one of the support programs that generate AI startups, which plans to provide up to $300,000 in free computing resources to select enterprises. Google also plans to offer up to $250,000 in free credits to new AI startups.
The more a company makes and uses the generated AI services with a specific cloud technology, the more that cloud service company can ensure stable customer demand and thus increase market share in the long term.
As a result, cloud service companies are investing in generative AI startups or expanding alliances.
In January this year, MS invested $10 billion in Open AI to provide AAJE cloud services exclusively to Open AI, and Open AI's chat GPT became the signature service of the AJE environment.
In February, Google invested $300 million in ANTOPIK, the startup of competing with Open AI's chat GPT, and provided ANTOPIK with Google cloud services, and it was observed that Lunway had previously used AWS as a cloud service, but with Google's investment as an opportunity, Google Cloud Services will be used. In addition, "Capstone AI" and AI chatbot "Bard", which provides its own large-scale language model (LLM) "PaLM 2", will be installed on Google Cloud.
AWS PARTNERED WITH IMAGE GENERATION AI DEVELOPMENT COMPANY STEVILITI AI IN NOVEMBER TO PROVIDE CLOUD SERVICES. In addition, it signed a partnership with startup Hugging Face to host Hugging Face's LLM on AWS.
In addition, AWS has also launched the enterprise cloud service "Bedrock", which provides not only self-developed AI models, but also AI models for startup partners. Offered its own LLM known as "Titan".
Google and AWS have also forged close relationships with Toronto-based companies such as KOHIR and AI21 LAPS.
Oracle, the fourth-largest U.S. cloud services company, also participated in a $270 million investment round by Kohier, a recently created AI model development company.
Kohier, an AI start company founded in 2019 by Google-turned-researchers, has developed interactive AI models similar to Hat GPT and is said to be a competitor to Open AI. In particular, the approach of promoting LLM that is specifically aimed at enterprises and focusing on corporate mayors/market strategies has attracted attention.
While it's not known exactly how much Oracle has invested, the industry predicts that Oracle will strengthen its cloud business with generated AI technology.
According to a person familiar with the industry, "Oracle's reality is that DB alone cannot survive now" and "this is understood as an attempt to paste other companies' creation AI in the Oracle Cloud to increase market share."
Oracle has not proposed such an AI model at this time. That's why some analysts say that a reversal will be sought by not favoring the Kosch language of a particular cloud service provider. As a cloud service provider, Oracle's position as a rising star is that it is easier to expand market share by adopting this model than to develop its own technology.
In fact, riding the generative AI boom, Oracle also said that the demand for the use of AI tools increased, and last quarter achieved good results. Oracle Chairman Larry Elison said during the investor conference call: "In reality, the demand for AI processing exceeds the available capacity."
At the same time, customer relationship management (CRM) professional company SALAPOS has also taken the lead in launching enterprise-generated AI cloud services and joined the cloud market competition. The plan is to use the more than 150,000 corporate customers it owns as a springboard, focusing on corporate customers rather than individual customers, and transacting between businesses to differentiate them.
The sales team launched "AI Cloud" services, including AWS, ANTOPIK, KOHIL, OPAI, etc., providing a variety of partner AI models, especially text generation models. Starting from AI code generation, AI models of sales forces that support functions such as business process automation can also be used. This is a form of providing cloud services for companies to develop their own chatbots, etc. based on generated AI. Alternatively, customers can store data in their own infrastructure while importing custom-trained AI models into the AI cloud.
The sales team's AI cloud is similar to Amazon's enterprise cloud service "Bedlock". In addition, security services that strengthen the protection of personal information are provided, improving cloud accessibility for enterprise customers who are concerned about sensitive information leakage.
Recently, HPE, which sells hardware such as computer servers, entered the AI cloud market ahead of Exas Kray's supercomputing infrastructure, owned by HPE. HPE launched a new cloud service "GreenLake", LLM "LUMINOUS" exclusively supplied by supercomputer infrastructure.
Customers can access the pre-trained LLM "Luminus" and use their own data to train and refine customized AI models. The service enables customers to build multiple AI applications and integrate them into their own business workflows.
In this way, in order to expand the grassroots of the cloud ecosystem, the technical capabilities of AI-initiated enterprises are combined with cloud infrastructure, and the synergy results are visualized, and the joint vertical and horizontal form will be further consolidated.
In particular, the increasing competition in generative AI technologies and services will eventually have a significant impact on the market structure directly related to the core business and performance of these big technologies. Andy Jassy Amazon CEO said: "To be honest, 6~9 months ago, AI was not so attractive, but as AI gets faster and bigger, it gets better and better." At the same time, "few recognize that the upcoming C6 partnership will allow cloud enterprises to gain steady customer demand, increase market share, and AI start enterprises to gain the computing power needed to operate AI models."
Especially with the boom in generating AI, the importance of large-scale cutting-edge cloud computing infrastructure in order to run large-scale AI models is emphasized.
It is reported that Open AI needs billions of dollars in order to continue to implement chat GPT, so it has joined forces with MS. As such, the reason AI startups form partnerships with cloud businesses is ultimately "money." Large data centers and graphics processing units (GPUs) are required, which require a lot of power.
Therefore, it is difficult for generative AI startups to turn down investment offers from cloud ventures such as computing infrastructure. There are also concerns that in order to maintain this relationship, entrepreneurs will hand over their shares and use the cloud at a low price, but eventually they will throw themselves into the arms of cloud enterprises.
Operating LLM with billions of parameters requires a stable computing infrastructure, so once model training begins, it is difficult to move to other platforms.
ML co-founder Jonathan Frankle Mosarieve ML also pointed out: "In addition to AI START enterprises providing the necessary cloud computing at a favorable price and investing in the arms of cloud enterprises that can obtain the huge amount of capital required, this partnership shows that the combination of generative AI startups and large technologies that provide cloud services is a helpless choice for generative AI research and services that continue to require high-performance computing power."
This kind of deal is not formally exclusive, but actually subordinates the startup's technology or product to the cloud platform. As a result, cloud giants are becoming proxy battles in the marketplace for the generation of AI dominance in cooperative startups.
On the other hand, this form of incorporation is the fact that the core service providers are also competitors, so there is growing concern that cloud giants will limit the opportunities for the generated AI to launch the enterprise. Therefore, there are also voices criticizing entrepreneurs of large cloud enterprises.
Jonathan Flancle Mosaic ML co-founder accused "cloud giants like monopolies, forcing startups to sign large-scale, multi-year contracts."