[Digital Today reporter Chi-gyu Hwang (황치규)] OpenAI has changed its partnership contract with Microsoft, allowing it to offer its models on clouds other than Microsoft Azure.

Amazon Web Services (AWS), the world’s largest public cloud service, first stepped forward to provide OpenAI models through its AI platform Bedrock.

Bedrock is a platform that lets users run AI models from various companies through a single API, including Anthropic, Amazon Nova, Cohere, Alibaba Qwen and DeepSeek. OpenAI models had been absent so far. As a leader behind the generative AI boom, the lack of OpenAI’s name on Bedrock was a disappointing point for AWS.

With deeper cooperation with OpenAI, AWS can strengthen Bedrock’s model portfolio. OpenAI can also use AWS as a backing to lay the groundwork to expand its foothold in the enterprise market.

• OpenAI-Microsoft contract revised...OpenAI products can also be sold on AWS • OpenAI finally enters AWS Bedrock...enterprise market push gains momentum

Some reports say some AWS customers are lukewarm about OpenAI models joining Bedrock. Some AWS customers say existing models on Bedrock, such as Anthropic’s, are sufficient now, even if that might not have been the case 2 to 3 years ago.

• Will AWS customers switch to OpenAI models?...Some respond it joined too late

The race is also heating up over underlying technologies that support AI agent payments. Cryptocurrency exchange OKX has unveiled an open standard dedicated to AI agent payments, the Agent Payments Protocol (Agent Payments Protocol·APP). APP will compete for a place as an open standard for agent payments with open protocol x402, incubated by Coinbase, and the Machine Payments Protocol proposed by Stripe. Payments platform Stripe introduced Link, a digital wallet that supports AI agent payments, at its annual conference.

• OKX unveils agent payment protocol APP...to compete with Coinbase x402·Stripe • Stripe launches 'Link' digital wallet for AI agents

Other moves by companies in and outside South Korea around AI are also summarised.

As AI usage surges, more companies are reshaping pricing plans in moves that effectively raise prices. Following Anthropic, Microsoft appears to be strengthening usage-based billing models. Microsoft has officially launched Agent 365, a platform that centrally manages AI agents in enterprise environments. Agents that operate within Microsoft Copilot, Teams and Microsoft 365, as well as agents from external partners, are included in its management scope. Microsoft also released Legal Agent, an AI agent for legal documents, in Word.

• Microsoft also expands usage-based AI pricing...a test in the enterprise market • Microsoft officially launches Agent 365...integrated management of third-party agents, too • Microsoft introduces 'Legal Agent' dedicated to legal documents in Word

AWS expanded its cloud call centre service Amazon Connect into 4 industry-specific AI agent services. AWS also unveiled Amazon Quick, a desktop AI app that automates various tasks.

• AWS expands Amazon Connect into 4 industry-specific AI agent services • Big Tech AI super-agent race heats up...AWS joins with 'Amazon Quick'

OpenAI will first make its cybersecurity-focused model GPT-5.5 Cyber available to selected users. OpenAI set up an application process on its website that asks applicants to submit eligibility requirements and plans for use. Google is signalling it may introduce advertising into its Gemini AI app, moving to diversify its revenue model.

• OpenAI limits distribution of GPT-5.5 Cyber...to be released first to verified security experts • Is Gemini’s free ride ending?...Will Google add ads to its AI chatbot

Cisco unveiled the open-source tool Model Provenance Kit to help companies address security and compliance issues related to using external AI models.

• Cisco unveils open-source tool to verify origins of third-party AI models • Salesforce launches 'Agentforce Operations' back-office automation AI

Adobe is expanding an agentic ecosystem spanning major technology companies, agencies and system integrators. The company said Adobe AI agents, skills and developer tools can be used within AWS, Anthropic, Google Cloud, Microsoft and OpenAI. Salesforce launched an AI system for back-office automation, Agentforce Operations.

• Adobe expands customer experience agentic AI ecosystem...strengthens links with major tech companies

Parallel Web Systems, an AI agent tool startup founded by former Twitter CEO Parag Agrawal (파라그 아그라왈), raised $100 million in a Series B round led by Sequoia. Parallel provides a web search and research API dedicated to AI agents.

• Parallel Web Systems raises additional funding at triple valuation...develops web search for AI agents

LG AI Research and Korea Electronics Technology Institute (KETI) completed phase 1 research and development of the Ministry of the Interior and Safety’s 'AI Safety Report Center' based on LG’s AI model EXAONE, and are aiming for a pilot service within the year. NHN Dooray is unveiling an AI agent service lineup and speeding up its push into the enterprise collaboration market. The company unveiled an 'action-stage' agent that goes beyond a simple conversational assistant and directly uses work data accumulated in Dooray to carry out tasks. LG CNS is expanding from a ChatGPT Enterprise reseller partnership contract with OpenAI to include a ChatGPT Edu reseller partnership, as it moves to target the education AX market. Notion signed a strategic partnership with Nexon for AI-based work innovation.

• LG EXAONE 4.5 to be used for the Interior Ministry’s 'AI Safety Report Center'...pilot service within the year • NHN Dooray unveils AI agent lineup...targets finance·public sector • LG CNS expands cooperation with OpenAI...targets education AX market • Notion-Nexon partnership for enterprise AX

CJ ENM unveiled its feature-length film 'Apartment', which combines AI and live-action footage. It raised production efficiency, including by completing filming in 4 days. A lack of consistency in backgrounds and differences in resolution between live-action and AI-generated footage are cited as limitations. CJ ENM plans to advance its AI production pipeline and expand AI use across dramas, films and advertising.

• CJ ENM unveils AI feature film 'Apartment'..."After CG comes AI"

Google DeepMind CEO Demis Hassabis (데미스 허사비스) recently attended the 'Google for Korea 2026' event and said it would "open a new renaissance in which humanity prospers," laying out a vision to realise 'AI for everyone' at the forefront of artificial general intelligence (AGI). The Ministry of Science and ICT also signed a memorandum of understanding with DeepMind focused on joint research on science and technology AI, training AI talent and responsible AI use.

• Google DeepMind: "We will open a new renaissance on the AGI frontier" • 'K-moonshot' steps up through global partnerships...government joins hands with Google DeepMind

Big cloud providers all posted results that beat expectations in first-quarter tallies. Aggressive investment strategies in AI appear to be continuing to translate into cloud business growth.

• Big cloud shows off record growth amid aggressive AI investment

As the shortage of AI chip supply deepens, big cloud providers are prioritising allocations of GPU inventory to internal demand and major customers. This is increasing the burden on AI startups. There are also reports that startups are paying relatively high costs to cloud providers to access GPU servers.

• AI computing power supply crunch worsens...AI startups feel the squeeze