Oracle’s Ellison cancels plan to sell up to 50 mln shares

Oracle Executive Chair and CTO Larry Ellison cancelled a Rule 10b5-1 plan that could have permitted sales of up to 50 million Oracle shares, with no shares sold and no other current plans to sell stock. The move follows Oracle's stronger-than-expected quarterly results, a fiscal 2027 adjusted EPS outlook increase of $0.05 to $8.10, and more than $30 billion in new AI cloud contracts that lifted backlog to $664 billion. Oracle maintained planned spending of $90 billion-$95 billion despite investor concerns that have pushed shares down more than 20% year-to-date in 2026; shares closed Friday at $150.28.

Analysis

The cancelled 10b5-1 plan removes a near-term technical overhang but is not, by itself, a fundamental signal: plan cancellations can reflect tax, collateral, liquidity or governance considerations. The investable question remains whether contracted AI demand converts into high-margin recurring cloud revenue fast enough to offset the front-loaded depreciation, power and financing burden. ORCL’s equity is likely to trade on evidence of capex-to-revenue conversion rather than backlog accumulation alone over the next 1-3 quarters.

The key second-order risk is that hyperscale AI infrastructure economics favor customers with cheaper capital and vertically integrated chips/networking. ORCL’s spending program raises sensitivity to construction delays, GPU availability, power interconnection timing and customer concentration; any slippage would leave fixed costs rising before revenue recognition and could pressure both free cash flow and the earnings multiple. Conversely, successful capacity delivery would validate ORCL as the principal non-hyperscaler beneficiary of enterprise demand seeking an alternative to MSFT, AMZN and GOOGL.

Consensus may be treating the share decline as a simple valuation reset, while the more important bifurcation is balance-sheet credibility. A sustained rerating requires quarterly proof that incremental cloud revenue, remaining performance obligation conversion and operating cash flow are accelerating together; guidance increases without corresponding cash conversion will likely be discounted. Near-term downside can also be amplified if long-duration technology multiples compress on higher real yields, irrespective of company execution.

Key Decisions for Investors

• Maintain a neutral-to-tactical long bias in ORCL only after the next earnings release confirms accelerating cloud revenue and operating cash-flow conversion; target a 3-6 month holding period. Do not treat the insider-plan cancellation as an entry catalyst.
• For existing ORCL exposure, buy 3-6 month downside protection or reduce gross if the company increases its capital-spending envelope, signals incremental external financing, or reports materially weaker free-cash-flow conversion. Those events would challenge the core AI-infrastructure return-on-capital thesis.
• Pair trade watch: long ORCL / short IGV or a basket of higher-multiple software names only if ORCL demonstrates capacity delivery and cloud growth acceleration. The thesis is relative multiple catch-up; it is invalidated if spending rises while cloud revenue and cash generation fail to inflect within two reported quarters.
• Avoid a broad long-AI-infrastructure extrapolation from ORCL. A more defensive expression is to retain exposure to power and data-center infrastructure beneficiaries while monitoring whether ORCL’s buildout drives incremental demand for grid capacity, cooling and networking rather than merely reallocating spend from MSFT, AMZN and GOOGL.