Oracle Reports a $664bn Backlog, and OpenAI Rules Out a 2026 Listing
Updated: 6 days ago
Title: Oracle Reports a $664bn Backlog, and OpenAI Rules Out a 2026 Listing
Date: 15 September 2026
Type: Blog
Author: SAASiQ (contact@saasiq.ai)
Word count: 1363 words
Reading time: 6 min
Published: 15-09-2026
Oracle reported first-quarter results for fiscal 2027 on 10 September, with revenue up 30 per cent to $19.3 billion and $664 billion of contracted revenue still to come. Cloud infrastructure revenue more than doubled, and capital spending in the quarter was larger than total revenue. In the same week Sam Altman said OpenAI will not go public in 2026, Harvey raised $550 million at a $15.5 billion valuation, and Ramp's spending data showed AI costs per employee falling at the heaviest-spending firms.
Oracle's quarter in figures
The results cover the three months to 31 August 2026. Total revenue was $19.3 billion, up 30 per cent in both US dollars and constant currency. Cloud revenue, which Oracle defines as infrastructure plus applications, rose 62 per cent to $11.6 billion. Within that, Oracle Cloud Infrastructure (OCI) brought in $7.4 billion, up 121 per cent, and cloud applications $4.2 billion, up 10 per cent.
Software revenue, which covers licences and support for products customers run on their own servers, fell 3 per cent to $5.5 billion. Services grew 5 per cent to $1.4 billion and hardware 15 per cent to $0.8 billion. GAAP earnings per share were $1.56, up 55 per cent, and non-GAAP earnings per share were $1.92, up 30 per cent. The quarterly dividend is $0.50 a share.
The $664 billion backlog
Remaining performance obligations, the value of signed contracts that Oracle has not yet recognised as revenue, stood at $664 billion at the end of August. That is $209 billion more than a year earlier. Oracle said it 'booked more than $30 billion of additional AI cloud contracts in Q1'.
The quarterly 10-Q filing gives the timing. Oracle expects to recognise about 13 per cent of the total, roughly $86 billion, as revenue in the next twelve months, 37 per cent in months 13 to 36 and 34 per cent in months 37 to 60, with the remainder after that. So seven-tenths of the backlog falls between one and five years out.
In the quarter Oracle delivered 850MW of additional datacentre capacity and 'more than 300,000 GPUs to its AI Cloud customers'. Revenue on these contracts is recognised as that capacity is handed over and used.
Capital spending and cash flow
Capital expenditure was $28.5 billion in the quarter, against revenue of $19.3 billion. Operating cash flow was $23.1 billion, up 184 per cent. Free cash flow, which is operating cash flow less capital expenditure, was negative: the filing puts it at minus $5,396 million, and some early coverage rounded that to minus $5 billion.
Oracle is paying for datacentre capacity now against contracts that pay out over several years, so capital spending runs ahead of the revenue it will support. Operating cash flow nearly tripled on the year, and capital expenditure exceeded it by $5.4 billion.
Applications, data and health
Cloud applications, the line that covers Fusion, NetSuite and Oracle Health, grew 10 per cent to $4.2 billion. The release does not break out Fusion ERP or Fusion HCM separately. Infrastructure now accounts for about 64 per cent of Oracle's cloud revenue and applications for about 36 per cent.
Two product items in the release reach applications customers. Oracle says its new AI Data Platform 'fully automates the creation of Enterprise Ontologies'. An ontology in this sense is a structured description of what an organisation's data means and how its records relate to one another, and Oracle says automating it makes it cheaper and quicker for any enterprise to have AI models reason over its private data and automate its business processes.
For Oracle Health, the release describes the AI Health Care Management and electronic health records system as 'an all-new 100% Agentic system', with agents 'for every medical specialty'. There are no executive quotes in the release, and no pricing for either product.
In SAASiQ's view, the practical question for Fusion customers this quarter is how the AI Data Platform will be licensed and when it reaches their own environments, since the release covers neither.
OpenAI rules out a 2026 listing
Sam Altman told Fortune, in an interview published on 11 September, that OpenAI will not go public this year. 'I would say not 2026, yeah. We got a lot of stuff to do,' he said, adding that 'given everything happening with safety, right now would be an ill-advised moment to go public.' The San Francisco Chronicle reported that OpenAI had filed confidentially for a listing in June.
SoftBank Group, which has committed more than $60 billion to OpenAI, fell about 11 per cent to ¥5,795 in Tokyo on Monday 14 September, and the Nikkei 225 fell 2 per cent, according to Investing.com. Nikkei Asia also pointed to weak sentiment across Asian technology stocks and a jump in oil prices. On the same day Nvidia's Jensen Huang told President Trump, on a live call at the All-In Summit, 'we're not going to let [an AI slowdown] happen', as TechCrunch reported it.
Separately, OpenAI stopped taking new subscribers to its $200-a-month Pro plan on 10 September, citing demand for GPT-6 Astra. The API and the Go, Plus, Business and Enterprise plans are still on sale. 'Demand for Astra is really unprecedented,' said Thibault Sottiaux of OpenAI.
Harvey raises $550 million
Harvey, the legal AI company, raised $550 million on 9 September at a $15.5 billion valuation, according to its own announcement and TechCrunch (a few outlets reported $15.6 billion). Diffusion and Lightspeed co-led the round. Sapphire Ventures and Whale Rock joined as new investors, and Sequoia, Kleiner Perkins, a16z, Coatue, GIC and Goldman Sachs Alternatives returned. Harvey was valued at $11 billion in March and $8 billion in December 2025, and has now raised more than $1.5 billion in total.
Harvey says 80 per cent of Am Law 100 firms use it, as do five Fortune 10 in-house legal teams, and it runs separate EU and Australian instances of its application. Its first in-house model, Harvey Tenet, is built on Moonshot's Kimi K3 and post-trained on legal data with Fireworks. TechCrunch described the company as showing that an industry can 'both heavily use AI and not rely on proprietary frontier AI labs'.
AI spending per employee
Ramp's AI Index for August, reported by TechCrunch on 9 September, draws on 70,000 of Ramp's business customers. Of those, 56 per cent paid for AI products in August, up 0.4 points on July. At the top 1 per cent of firms, AI spend per employee fell almost 10 per cent to $7,205. The average price paid was $0.68 per million tokens, down from $1.15 in March.
Ara Kharazian, Ramp's economist, said competition between OpenAI and Anthropic is 'driving the price down for companies' and 'driving spend down at the top 1%'. For comparison, the US Census Bureau's survey of 23 August puts AI use at 22 per cent of US businesses.
Other deals
Profound, which helps brands appear in the answers AI assistants give (a practice it calls answer engine optimisation), raised a $180 million Series D at a $1.8 billion valuation on 15 September. Sequoia and Kleiner Perkins led, less than seven months after a $96 million Series C. Profound says revenue has tripled in six months and that it has more than 1,000 enterprise customers, including Comcast, Estée Lauder and Walmart.
OpenAI has bought Glass Imaging, which makes neural camera-processing technology, for about $300 million, according to the Wall Street Journal as reported by TechCrunch on 14 September. Its founders, Ziv Attar and Tom Bishop, previously led Apple's Portrait Mode work. OpenAI did not comment.
Listen Labs, an AI research start-up with about $30 million in annualised revenue, walked away from a $125 million round at a $1.5 billion valuation led by Menlo Ventures while it held talks about a sale to Salesforce at about $2 billion, as Business Insider reported.
Oracle's guidance
Oracle included its outlook in the results release. It expects second-quarter total revenue to grow by 30 to 34 per cent and cloud revenue by 64 to 70 per cent. For the full year to 31 May 2027 it expects revenue of at least $90 billion and non-GAAP earnings per share of $8.10. The second quarter ends on 30 November.
SAASiQ - Intelligent Solutions for SaaS ©
