OpenAI June 2026: Samsung Deal, Cyber Tools, IPO Plans

OpenAI closed June 2026 with a cluster of significant moves: a global enterprise deployment with Samsung Electronics, an expanded cybersecurity initiative including a new model called GPT-5.5-Cyber, and continued signals about a potential public offering — all while CEO Sam Altman publicly expressed little enthusiasm for running a public company.

Samsung Electronics Deploys ChatGPT and Codex Globally

Samsung Electronics is rolling out ChatGPT Enterprise and Codex to its entire workforce in Korea and all employees worldwide in its Device eXperience (DX) division — one of OpenAI’s largest enterprise deployments to date. The agreement covers technical and non-technical functions alike, spanning software development, marketing, product development, and manufacturing.

According to the OpenAI blog post published June 21, 2026, Samsung employees will use ChatGPT for knowledge-based tasks such as information retrieval, document drafting, and data interpretation, while Codex handles code writing, review, and debugging. OpenAI noted that Codex, which began as a developer tool, is now being applied to non-technical workflows as well.

ChatGPT Enterprise provides data protection, user and access management, and security controls — capabilities that allow Samsung to deploy AI within its existing governance framework. The scale of the rollout positions it among the most significant B2B AI agreements OpenAI has announced publicly.

OpenAI Expands Daybreak Cybersecurity Initiative

OpenAI on Monday expanded its Daybreak cybersecurity program, releasing an updated Codex Security plugin and launching a new model, GPT-5.5-Cyber, framing both around a problem it argues has grown more acute than vulnerability discovery: getting patches actually deployed.

According to SecurityWeek’s report, OpenAI contends that AI has accelerated vulnerability discovery to the point where defenders are overwhelmed by the volume of findings. The updated Codex Security plugin addresses this by scanning entire codebases, tracing attack paths, generating patches, and exporting results via SARIF files and CodeQL queries into existing vulnerability management pipelines.

Codex Security: Scale and Performance

Since a research preview launched in March 2026, Codex Security has processed more than 30 million commits across over 30,000 repositories, with human reviewers confirming more than 70,000 fixes and an additional 500,000 findings resolved automatically, according to SecurityWeek.

GPT-5.5-Cyber Model Details

The full release of GPT-5.5-Cyber follows an earlier version focused on reducing unnecessary refusals. On the CyberGym benchmark — which tests whether an agent can reproduce known vulnerabilities — the model scored 85.6%, compared to 81.8% for standard GPT-5.5, SecurityWeek reported. Access remains limited to verified defenders. OpenAI also unveiled Patch the Planet, an initiative co-founded with Trail of Bits and developed with HackerOne.

Omio Cuts Development Time by 80% Using OpenAI APIs

Travel platform Omio has rebuilt core product development workflows using ChatGPT, Codex, and OpenAI’s API, reducing the effort required to build new products to 20% of previous levels, according to a case study published on the OpenAI blog June 23, 2026. A project that previously required multiple developers over three months now takes one developer one month.

Omio connects travelers with trains, buses, ferries, and flights across 47 countries through more than 3,000 transportation providers. The company’s CTO, Tomas Vocetka, discussed the company’s push toward conversational travel — where users describe a journey in natural language and receive personalized, bookable itineraries — as a core product direction.

The Omio deployment illustrates a broader pattern in OpenAI’s enterprise strategy: positioning its tools not just as productivity add-ons but as infrastructure for rearchitecting how products are built and delivered.

Sam Altman on AGI, IPOs, and 2030

OpenAI’s public profile in June 2026 extended beyond product announcements. Fortune reported that Altman described himself as “0% excited” about the prospect of leading a public company, even as OpenAI takes steps toward what could be one of the largest IPOs in tech history. Altman’s reluctance appears rooted in the operational and disclosure demands of public markets rather than opposition to raising capital.

Separately, Fortune reported that Altman has made his boldest public prediction yet: that AI will surpass human intelligence by 2030. Rival AI executives, the publication noted, have placed that timeline even earlier. Altman’s forecast aligns with OpenAI’s internal urgency around model capability development, though the company has not attached specific technical milestones to the 2030 date.

What This Means

The Samsung deployment and the Omio case study together signal that OpenAI is moving from individual-user adoption to organization-wide infrastructure contracts — a shift with significant revenue implications. Enterprise agreements at Samsung’s scale, covering an entire national workforce and a global division, represent a different commercial model than per-seat SaaS pricing.

The Daybreak expansion is notable for its framing: OpenAI is not positioning AI as a vulnerability discovery tool but as a remediation engine. That reframe matters because it addresses the legitimate criticism that AI-generated vulnerability reports create more work for already-strained security teams. The 500,000 automatically resolved findings figure, if accurate, is a substantive operational claim — though independent verification of that number has not been reported.

Altman’s IPO ambivalence, paired with OpenAI’s structural moves toward going public, reflects a tension that will define the company’s next phase. Public markets impose transparency requirements and quarterly pressure that conflict with the long-horizon, capital-intensive bets OpenAI is making on AGI development.

FAQ

What is OpenAI’s Codex Security plugin?

Codex Security is a cybersecurity tool integrated into OpenAI’s Codex platform that scans codebases, traces attack paths, generates patches, and exports findings into vulnerability management systems. Since its March 2026 research preview, it has processed more than 30 million commits across over 30,000 repositories, according to SecurityWeek.

What is GPT-5.5-Cyber and who can access it?

GPT-5.5-Cyber is OpenAI’s most capable model for authorized security work, scoring 85.6% on the CyberGym benchmark versus 81.8% for standard GPT-5.5. Access is currently restricted to verified security defenders, not general users.

When does Sam Altman predict AI will surpass human intelligence?

Altman has publicly predicted AI will surpass human intelligence by 2030, according to Fortune. Some rival AI executives have suggested the timeline could be even shorter, though no consensus technical definition of “surpassing human intelligence” has been agreed upon across the industry.