Steve Eisman Says Half of Oracle’s Backlog Is Tied to OpenAI, as ORCL Shares Flag Pressure

Oracle’s backlog has become a focal point for investors trying to separate durable enterprise demand from the volatility that can follow platform shifts in artificial intelligence. In a Yahoo Finance report, Big Short investor Steve Eisman said that “half of Oracle’s backlog” is tied to OpenAI, framing the company’s pipeline as increasingly linked to AI infrastructure buildouts.

The report also highlighted the scale of Oracle’s backlog, citing a figure of $638 billion. In that same summary, it said Oracle stock has lost 54% over the past year, underscoring how investor expectations for growth and margins have been under strain despite the large backlog number.

Eisman’s central claim is a directional one: he attributed a material share of the backlog to OpenAI-related demand. While the backlog itself is presented as a large pipeline of future work, backlog is not the same thing as revenue. Companies typically book backlog when customers commit to future delivery, but the timing of recognition can vary based on implementation schedules, contract terms, and project execution.

Oracle’s backlog has historically been watched as an indicator of enterprise spending coming down the pipeline, particularly for long-duration technology and services contracts. In practical terms, a high backlog can provide visibility, but it does not guarantee near-term results. Investors still need to understand how quickly backlog converts into revenue and whether associated costs align with expectations for profitability.

The Yahoo Finance summary also noted a split between market performance and Wall Street sentiment. It said 80% of analysts who cover Oracle still rate the stock a buy, even as the shares have dropped substantially over the last year. That kind of divergence often points to disagreements over timing, competitive dynamics, and the risk that headline AI partnerships do not translate smoothly into financial outcomes.

From a market-structure perspective, Eisman’s comment also highlights the degree to which AI-linked infrastructure spending has become concentrated among a handful of large customers and ecosystems. When a company’s backlog is perceived as heavily dependent on one partner or use case, investors may scrutinize contract duration, renewal risk, and whether the supplier’s role expands beyond delivery into longer-term managed services.

Oracle did not publicly detail, in the Yahoo Finance piece referenced here, a breakdown of backlog by customer or partner that can be independently verified from the excerpt alone. As a result, the “half tied to OpenAI” figure should be treated as Eisman’s assessment of the composition of the backlog, not as a company-confirmed statistic in the report summary.

Looking ahead, the key question for investors is not simply how large the backlog is, but how consistently it converts into revenue over time and how contract economics evolve as AI deployments move from buildout to operations. Watch for any company disclosures that clarify backlog composition, timing of delivery schedules, and how AI-related projects affect margins, along with updates on Oracle’s guidance and investor commentary on demand durability.