A report of a failed negotiation has sent already pressured semiconductor stocks lower once again.
According to a Business Insider report after the market close on Tuesday, June 16, Microsoft (MSFT) had recently engaged in talks with Oracle (ORCL) to lease the latter's cloud infrastructure, but the discussions ultimately fell apart over security and compliance issues, involving a sum exceeding $3 billion. The news sent Oracle's stock and the broader semiconductor sector lower, further exacerbating the day's already weak market performance.
The Core Hurdle for the Deal
The central obstacle to the deal was a security certification framework called FedRAMP, a standard required by the U.S. government for cloud service providers handling government data. The report, citing sources familiar with the matter, stated that Oracle's public cloud lacks this certification and was unwilling to undertake the modifications necessary to obtain it. An Oracle executive reportedly noted that adding FedRAMP certification to the public cloud "would be a huge engineering lift."
Oracle's Response and Market Context
Oracle promptly denied the report. A spokesperson said, "The details as described in the article are not accurate. Microsoft is both a partner and a customer in our OCI platform. We have an extremely tight and productive relationship and we often discuss ways to expand our existing partnership." The spokesperson declined to specify which details were inaccurate. Microsoft declined to comment.
This marks the second incident within a week where news related to AI infrastructure has triggered market volatility. Previously, AI infrastructure company Crusoe paused its 1.8GW data center project in Wyoming at a client's request, causing related stocks to tumble intraday.
Why Microsoft Sought the Deal
Microsoft anticipates its calendar year 2026 capital expenditures to reach $190 billion, primarily for data center expansion. Even so, its own capacity remains insufficient—the company had previously turned to Amazon to supplement computing power for its GitHub code development business to address recent service disruptions.
Sources familiar told media that Microsoft is seeking agreements with other cloud service providers to prioritize its own Azure cloud resources for its customers. Microsoft stated, "We are looking for capacity everywhere."
Beyond Oracle, Microsoft has other options. The public clouds of Amazon and Google already possess FedRAMP certification. According to sources, Microsoft is still evaluating and exploring options for leasing cloud infrastructure.
This phenomenon of "large tech companies leasing capacity from other large tech companies" is becoming a new normal in the AI era. Google and SpaceX recently disclosed that Google will pay SpaceX $92 million per month from October 2026 to June 2029 for AI computing power. Just two months ago, Google's cloud business signed an agreement to sell AI compute capacity to Anthropic.
Recent Sensitivity to Infrastructure News
This is not the first time recently that AI infrastructure news has triggered a market chain reaction.
Just a week earlier, on June 10, AI infrastructure company Crusoe announced the suspension of its 1.8GW large-scale data center project in Cheyenne, Wyoming (codenamed "Project Emerald"), at the request of an undisclosed client. The news caused a more than 9% plunge in the stock of Bloom Energy, which has supply chain ties to the project, and contributed to declines in the Nasdaq and S&P 500 that day.
The fallout continued. According to Bloomberg, the direct reason for Crusoe's forced exit from the project was Google raising concerns about the project's costs and timeline under Crusoe's leadership. Energy company Black Hills subsequently stated the project would proceed without Crusoe's involvement, with Google finalizing an agreement with the remaining partners to purchase compute capacity from the site. The project is expected to be operational by early 2028, with the timeline unchanged.
Why the Market Is So Sensitive
The two incidents, separated by only a week, share a highly similar trigger logic: a single point of weakness in the AI infrastructure chain immediately prompts the market to re-evaluate the delivery pace of the entire chain.
Goldman Sachs partner Rich Privorotsky promptly highlighted the event, pointing to the market's core vulnerability: "In a market where everything is tied to AI capital expenditure, even sporadic delays, deferrals, or reprioritizations are enough to force investors to re-evaluate their assumptions about future demand."
He provided data showing that current market momentum returns are in the 90th percentile over the last five years, total exposure is in the 99th percentile, financing spreads have widened as leverage demand rises, and retail participation via leveraged ETFs remains substantial.
"Everything is increasingly tied to AI spend. It is part of the hardware long trade, it is driving part of GDP, and it is driving much of the market's performance. The circularity is becoming harder to ignore," he wrote.
This high degree of correlation means any "pause" or "abandonment" in any part of the chain is amplified and reinterpreted by the market. Microsoft's abandonment of this $3 billion deal with Oracle is, in essence, a negotiation failure due to technical compliance issues. Yet, under current market sentiment, it is being read as a larger question: Can the expansion of AI capital expenditure truly continue to be delivered at the originally anticipated pace?
Comments