Last price
$175.06
The old king reinvented as an AI-training cloud landlord — Autonomous Database plus OCI Gen2 make Oracle one of the least-appreciated AI-infrastructure positions in public markets.
Live quote sourced from Yahoo Finance. Prices cited in narrative below reflect the original memo date and may be stale.
Oracle is two thesis-aligned businesses bolted onto a third legacy one. (1) Autonomous Database is the clearest working autopilot in enterprise software — a service that replaces the DBA entirely and is priced as a managed outcome. (2) OCI Gen2 has become the de-facto neutral AI-training cloud: cheaper and faster interconnect than AWS/Azure for dense GPU clusters, which is why OpenAI, xAI, and Meta signed mega-deals. (3) Fusion SaaS is the incumbent-apps franchise where agents will eventually automate finance and HR workflows — not the thesis's sharpest point but the execution layer for future agents.
Oracle in 2026 reads less like a database dinosaur and more like a capex-heavy infrastructure platform with embedded AI-service franchises. The key debate is whether OCI's capacity lead translates into durable economics once hyperscalers match interconnect architecture. Bulls argue the capacity deals lock revenue for a decade and the software margin profile remains intact. Bears worry about capex dilution, Larry-Ellison succession risk, and capacity-glut scenarios if AI-training demand plateaus. The Autonomous Database line item is the under-appreciated piece; it is a services-as-software business growing 40%+ inside a company the market prices as a low-teens-growth value name.
OCI Gen2 became the neutral AI-training cloud between 2023 and 2025. The architectural wins were RDMA-at-scale networking, cheaper H100/H200/B200 clusters, and willingness to sign multi-year GPU capacity commitments. Combined with Autonomous Database — a zero-DBA managed service that sells outcomes not instances — Oracle now has two thesis-aligned franchises underneath the Fusion apps layer. The market treats Oracle's AI exposure as Azure-lite; the reality is that OCI has become a preferred bet-the-company training cloud for frontier labs that don't want to live inside Microsoft or Google.
Oracle now runs capex at hyperscaler scale without hyperscaler cash flow. Free-cash-flow conversion has compressed while debt has risen. If AI-training demand plateaus, capacity glut risk is concentrated here. Meanwhile AWS/Azure/Google will close the interconnect gap by 2027; Oracle's lead is narrower than the stock implies. Larry Ellison remains the architect; succession risk is real for a company built around one founder's deal-making. Fusion apps still trail Workday/Salesforce in competitive net-new.
| Segment | Approx. mix | AI posture | Services-as-software read |
|---|---|---|---|
| Cloud Infrastructure (OCI) | ~20% | AI-training capacity + autonomous DB | Core thesis — software sold as managed outcome |
| Cloud Applications (Fusion + NetSuite) | ~20% | Copilot + emerging agents | Thesis-adjacent — agents as upgrade motion |
| License support (on-prem DB) | ~35% | Cash engine, some AI features | Non-thesis legacy moat |
| Hardware (Exadata, servers) | ~10% | Attached to Autonomous DB | Thesis-adjacent via autopilot DB |
| Services + other | ~15% | Mostly human-led | Non-thesis |
Oracle signed generational capacity deals with OpenAI, xAI, and Meta because it offered better networking and better commercial terms than AWS/Azure. That RPO is real and underwrites OCI revenue through 2030. If the thesis that frontier labs need a neutral partner holds, OCI is the franchise.
Zero-DBA, self-patching, self-tuning database at GA since 2018 — priced as a managed outcome, with a margin profile that shames traditional DBaaS. This is a quiet thesis-native franchise generating several billion in ARR and growing 40%+.
Finance, HR, and supply-chain agents need a system of record to act against. Oracle owns more of that system of record than any vendor except SAP. When autonomous finance agents go mainstream, Oracle is the surface.
Even with capex intensification, Oracle prints >$15B FCF and returns most of it via buybacks and dividends. Capital discipline is a structural advantage versus pure-play AI infrastructure names.
Oracle's capex load is now near hyperscaler levels but without the matching cash flow. If AI-training demand plateaus or deal terms compress, free cash flow conversion craters and the multiple re-rates.
AWS Nitro and Azure's Maia/Cobalt programs are catching OCI's architectural lead. By 2027–28, Oracle's technical moat in frontier-training clouds may be neutralised.
In net-new SaaS deals Fusion trails Workday in HCM and Salesforce in CRM. Oracle's apps growth is partly driven by install-base upgrades, not share gains.
Oracle's dealmaking culture lives in Larry Ellison. Post-Ellison, the company has to operate without its top salesperson and architect — a material transition for a $600B market cap.
Oracle is thesis-positive because two of its three main businesses — OCI and Autonomous Database — are thesis-native. OCI is the infrastructure layer for AI-as-service at frontier scale, and Autonomous Database is one of the clearest working examples of software sold as a managed outcome. Fusion is thesis-adjacent as the execution surface for autonomous finance/HR agents. The legacy license business is a cash engine funding the transformation. Oracle is not a pure services-as-software name, but the thesis-native portion of the business is both large and growing fast.
A hybrid AI-infrastructure + autonomous-database play trading like a value software stock. Own through capex digestion; OCI + Autonomous DB are the durable franchises.