"Catch Up with Chat GPT"... Busy Pairing of Mega AI and Cloud

Google Partners with Anthropic to Develop ChatGPT Rival
Combining Cloud and Massive AI Like MS and OpenAI
Naver Also Aims for Synergy with 'NewCloud'

[Asia Economy Reporter Yuri Choi] Big tech companies are busy moving to catch up with Chat GPT. Just as Microsoft (MS) and OpenAI have joined forces, cloud technology and massive artificial intelligence (AI) are working as one. Google, which issued a red alert (code red) following the emergence of Chat GPT, and Naver, which announced 'Search GPT' to compete with Chat GPT, have adopted the same strategy.

Bloomberg reported on the 4th (local time) that Google invested $400 million (about 500 billion KRW) in AI startup Anthropic. Anthropic was founded in 2021 by siblings Daniela Amodei and Dario Amodei, founding members of OpenAI. Last month, they released a test version of the AI chatbot 'Claude,' considered a rival to Chat GPT.

Along with the investment, Google Cloud signed a partnership with Anthropic. The plan is to develop and deploy Claude using Google Cloud's computing service, Tensor Processing Unit (TPU). TPU is a semiconductor chip used for AI training. It acts as the brain of AI with extremely powerful computing capabilities. The TPU v4, the 4th generation version of TPU released by Google Cloud in June last year, supports 9 exaflops of computing power, meaning it can perform 90 quadrillion operations per second.

MS and OpenAI are collaborating in the same way. MS provides cloud computing resources necessary for OpenAI to develop AI models and has obtained exclusive rights to use GPT-3, the foundation of Chat GPT. Recently, MS invested $10 billion (about 12 trillion KRW) in OpenAI and is gradually integrating Chat GPT into its own services. AI chatbots like Chat GPT have been applied to the search engine 'Bing' and collaboration tool 'Teams.' Furthermore, MS plans to utilize Chat GPT technology in cloud service 'Azure' as well as Office programs such as Word, PowerPoint, Excel, and Outlook.

Experts agree that the combination of massive AI and cloud is essential. AI needs cloud-based computing resources to become smarter by training large-scale language models and to handle many user requests based on that. To answer questions from over 100 million users monthly, Chat GPT requires processing an enormous amount of data, a role fulfilled by MS Cloud.

The active collaboration between massive AI and cloud also facilitates naturally bringing AI services into cloud products. MS has focused on developing Chat GPT as a business-to-business (B2B) solution, recognized as a tool to enhance work productivity across various fields. It is provided to corporate clients through MS Azure.

Ha Jung-woo, head of Naver Cloud AI Lab, explained, "As AI grows larger, infrastructure becomes important, and compatibility with cloud has become key. Google and MS are adopting the same strategy from this perspective."

Domestically, Naver is responding with its 'New Cloud Strategy.' Earlier this year, Naver reorganized its structure by integrating Clova, an in-house independent company (CIC) responsible for AI research and development (R&D), with Naver Cloud, a Naver subsidiary. This aims to create synergy between Naver's massive AI 'Hyper Clova' and cloud.

Naver has experience operating 'Clova Studio,' which provides Hyper Clova through its own cloud. This offered an environment where external startups could create AI services or apps. Going forward, Naver plans to develop AI solutions specialized by domain and offer them as cloud products. For example, providing AI solutions specialized for sales teams to companies would enable even new employees to respond like veteran salespeople.

Park Jong-yeol, leader of Naver Cloud's business strategy, said, "The quality of information can vary depending on what learning information experts provide to the AI model. We are considering a direction where the AI learns information from specific vertical domains and outputs the necessary information."