Oracle expands AI tools as valuation debate grows

Oracle (ORCL) announced new agentic AI innovations for Oracle AI Database at its AI World Tour in London, aiming to help customers build, deploy and scale secure agentic AI applications suited to production workloads across multicloud and on‑premises environments.

The Oracle AI Database is designed to architect agentic AI and data together across operational databases and analytic lakehouses, allowing AI agents to securely access real‑time enterprise data wherever it resides and combine that data with large language models trained on public data.

Customers can select AI models, agentic frameworks, open data formats and deployment platforms, with Oracle Exadata users gaining Exadata Powered AI Search for accelerated AI queries in high‑volume, multi‑step agentic workloads.

Oracle introduced Autonomous AI Vector Database, which provides vector database functionality on top of Oracle Autonomous AI Database, offering intuitive APIs, a web interface, and enterprise‑grade security, reliability and scalability. It is in limited availability via Oracle Cloud free and low‑cost developer tiers, with one‑click upgrades to the full database.

The AI Database Private Agent Factory offers a no‑code AI agent builder that runs in public clouds or on‑premises, including pre‑built agents such as a Database Knowledge Agent, Structured Data Analysis Agent and Deep Data Research Agent, allowing customers to deploy data‑driven agents without sharing data with third parties.

Oracle also highlighted Oracle Unified Memory Core, enabling low‑latency reasoning for AI agents across vector, JSON, graph, relational, text, spatial and columnar data in a single converged engine with consistent transactions and security.

New security features include Oracle Deep Data Security, which enforces end‑user‑specific data access rules in the database so each user or AI agent only sees permitted data, using persona and function‑based rules to implement least‑privilege access and guardrails against AI‑era threats such as prompt injection.

Oracle Private AI Services Container lets customers with strict security requirements run private instances of AI models without sharing data with external providers, while offloading compute‑intensive AI tasks and keeping data within their environment in public cloud, private cloud or on‑premises, including air‑gapped deployments.

Oracle Vectors on Ice provides native support for vector data stored in Apache Iceberg tables, enabling AI Vector Search on data lake data and unified search across databases and data lakes, while Oracle Autonomous AI Database MCP Server allows external AI agents and MCP clients to access database capabilities securely without custom integration code.

Separately, Oracle launched Fusion Agentic Applications, a suite of AI‑driven tools within Oracle Fusion Cloud, alongside AI‑powered upgrades to its Fusion financial suite that add agent‑style capabilities for finance users.

These products embed decision‑making AI into core enterprise functions, including finance, human resources, supply chain and customer experience, with Oracle positioning AI agents to act on unified policies, workflows and approvals across Fusion Cloud rather than relying only on dashboards.

The launches align with Oracle’s broader push into industry‑specific AI such as healthcare agents and embedded finance, and arrive as Oracle shares trade around $154.34, with returns of 77.5% over three years, 132.4% over five years, but only 0.7% over one year and a 21.1% decline year to date.

A discounted cash flow analysis from Simply Wall St estimates Oracle’s intrinsic value at about $257.71 per share, implying the stock is 40.1% undervalued versus a recent price of $154.34, using a two‑stage free cash flow to equity model that assumes free cash flow turns positive and reaches $28.5 billion in 2030.

On earnings metrics, Oracle trades at a P/E of 27.4 times, below the software industry average of 29.9 times and a peer group average of 56.8 times; Simply Wall St’s proprietary Fair Ratio suggests a higher P/E of 56.3 times, indicating potential undervaluation on this measure as well.

Narrative‑based fair value scenarios on the same platform range widely, from a cautious case around $119.97 per share with mid‑single‑digit revenue growth to a bullish AI‑focused case near $389.81 with 28% revenue growth assumptions, underscoring differing expectations around Oracle’s AI and cloud trajectory.

In a market note, BlackRock (BLK) said the traditional 60/40 stock‑bond portfolio is not working as effectively because stocks and bonds have recently tended to move together, citing a week in March 2026 when the S&P 500 (SPX) fell while 10‑year U.S. Treasury yields rose to 4.28% and government bonds offered little refuge.

BlackRock highlighted quality U.S. equities, particularly AI‑focused large‑cap technology stocks, as well as emerging‑market hard‑currency debt in commodity‑exporting countries such as Brazil, as areas it currently favors, pointing to its iShares J.P. Morgan USD Emerging Markets Bond ETF as one option.

Separately, valuation work on BlackRock’s own shares using an Excess Returns model produced an intrinsic value estimate of about $999.55 per share versus a recent price of $974.58, implying roughly 2.5% undervaluation, while a P/E of 27.30 times sits below the Capital Markets industry average but above a company‑specific Fair Ratio of 19.06 times.

Cboe Global Markets shares recently traded around $280.62 after modest short‑term declines but multi‑year gains of 32.5% over one year, 120.9% over three years and 209.2% over five years, keeping the exchange operator on many investor watchlists.

An Excess Returns analysis estimated Cboe’s intrinsic value at about $266.27 per share, putting the stock roughly 5.4% above that level, while its current P/E of 26.8 times sits below Capital Markets peers but above a Fair Ratio of 14.4 times, suggesting a premium relative to that framework.