Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minutePC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Some links on this page are affiliate links: if you buy through them we may earn a commission, at no extra cost to you.
The reported $10 billion xAI–Oracle server deal was never a signed contract. On July 9, 2024, Reuters, citing The Information, reported that the companies had ended talks over a potential multiyear arrangement to rent Oracle’s Nvidia-equipped AI capacity. Reports pointed to disagreement over deployment speed and power availability. xAI’s move toward building its own infrastructure did not mean it stopped using Oracle: current company and Oracle materials show that relationship continues.
What ended—and what did not
The negotiations concerned an expansion of an existing xAI–Oracle arrangement, not the cancellation of a completed $10 billion purchase. Reuters reported that xAI already rented Nvidia AI chips through Oracle Cloud and had a contract to train models in Oracle’s Gen2 Cloud. The proposed multiyear deal would have added substantial server capacity for a planned supercomputer. The reported $10 billion was an estimated potential contract value, not a disclosed upfront payment or final price for equipment. (Reuters report; The Information’s earlier report)
So the accurate description is that talks over a major expansion ended. The available reporting does not say xAI stopped using Oracle, that Oracle stopped serving xAI, or that a signed $10 billion contract was canceled.
Why did the talks end?
Reuters’ account, based on reporting by The Information and people familiar with the discussions, cited a mismatch over timing. Musk wanted the supercomputer built faster than Oracle considered feasible. Oracle also reportedly questioned whether xAI’s preferred site had enough electrical power. These are reported explanations from private negotiations, not a jointly confirmed account by both companies.
#1 Best Overall
- New
- Mint Condition
- Dispatch same day for order received before 12 noon
- Guaranteed packaging
- No quibbles returns
Capacity allocation was another practical factor: Reuters reported that the Oracle capacity under discussion was subsequently contracted to another customer. That helps explain why the proposed arrangement did not simply remain available while the companies worked through differences.
Why speed and power matter for AI servers
A large AI training cluster is more than a shipment of GPUs. It needs servers, high-speed networking, storage, a prepared data-center site, cooling and dependable electricity. Each component has its own procurement and deployment schedule. A building can be available while its power supply is not; GPUs can be ordered but not yet installed in a working cluster.
That makes power a central constraint, not a minor facilities detail. If a site cannot deliver enough electricity or cooling capacity, adding more processors does not make the system ready to train a model. The reporting does not disclose the specific site, its power requirements, the number of GPUs in the proposed Oracle capacity, or the detailed terms the companies discussed.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →xAI’s alternative: more control through its own infrastructure
Reuters reported that xAI chose to obtain Nvidia hardware and build the system itself, with Musk describing that approach as the fastest route to completion. The coverage referred to Nvidia H100 GPUs in the system xAI was building. This was a shift toward greater control over deployment and design—not proof that xAI owned every facility involved or abandoned cloud services altogether.
| Approach | Potential advantage | Main trade-off |
|---|---|---|
| Rent cloud capacity | Access to an operating provider platform without taking on all construction and facilities work directly | Timing, location, available capacity, power and service terms depend partly on the provider |
| Build or arrange its own infrastructure | More direct control over hardware configuration, networking and deployment decisions | Requires capital and expertise, while construction, utility connections, cooling and equipment delivery remain risks |
Buying chips is not the same as building and operating a data center. The public reporting does not establish where all the chips were installed, whether third parties supplied colocation, the final cluster size, or how much capacity was used for training rather than inference. Nor does it show that self-building was cheaper overall. It suggests that xAI prioritized control of the schedule.
Did xAI leave Oracle?
No. Current official materials show an ongoing connection, though they do not establish that the original $10 billion proposal was revived. xAI’s subprocessor list names Oracle for cloud infrastructure, services and support. Oracle’s OCI documentation says xAI’s Grok models are hosted in an Oracle Cloud Infrastructure data center in a tenancy provisioned for xAI.
Those disclosures distinguish two ideas that headlines can blur: xAI could expand its own compute facilities while continuing to use Oracle for cloud infrastructure or particular workloads. A company’s infrastructure strategy need not be an all-or-nothing choice between owning everything and renting everything.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Clear out junk files and repair common Windows errorsFree Scan →What happened to xAI’s buildout afterward?
In a January 2026 announcement, xAI said it had expanded Colossus I and II and ended 2025 with more than one million H100 GPU equivalents. That scale is xAI’s own claim, not an independently audited capacity figure in the announcement. xAI also said it raised $20 billion in Series E financing, above its stated $15 billion target, with Nvidia and Cisco Investments among the strategic investors. (xAI’s announcement)
This later progress provides context for the company’s infrastructure direction; it does not prove that the 2024 choice was obviously right at the time, nor does it show that the Oracle discussions later resumed. A GPU-equivalent count also does not, by itself, reveal usable compute: networking, utilization, uptime and workload all matter.
What the reported $10 billion figure means for readers
There is no basis in the reported facts for treating $10 billion as revenue Oracle definitely lost, money xAI saved, or a completed bill. It was the estimated value of a potential multiyear rental arrangement. No final contract value, duration, GPU-hour price or detailed service-level terms were disclosed. For business and investor readers, the distinction matters: a large proposed deal can signal demand for AI infrastructure without becoming booked revenue or a binding commitment.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
Recommended Free Tools
