If 2023 was the year of ChatGPT, 2025 has unquestionably been the year of AI deals. The past several months have been a blur of multi-million and multi-billion dollar collaboration deals and purchase agreements. To help keep track, we’ve created this handy tracker.
In the short term, it provides a way to keep tabs on who is doing what and with whom. In the long term, it could shed insight into how real the dollar figures in these deals turn out to be.
So without further ado, behold our Encyclopedia of AI deals. And yes, just like an encyclopedia, everything is listed alphabetically.
This story will be updated on a rolling basis.
Amazon/AWS
In May, AWS unveiled a deal with HUMAIN to invest more than $5 billion in AI infrastructure, AWS services, and AI training and talent development in Saudi Arabia.
In November, AWS and HUMAIN announced plans to deploy up to 150,000 AI accelerators (including AWS’ Trainium AI chip and Nvidia’s GB300 infrastructure) in a data center facility known as an “AI Zone” in Riyadh.
The same month, Amazon inked a $38 billion cloud services supply deal with OpenAI.
AMD
In May, AMD signed a partnership deal with Cisco and Saudi Arabia’s HUMAIN initiative to develop AI infrastructure in the country.
In October, AMD inked a partnership deal with OpenAI to deploy 6 GW worth of AMD’s Instinct GPUs, with an initial 1 GW set to roll out in the back half of 2026. AMD CEO Lisa Su said the deal has the potential to generate “well over $100 billion in revenue” over the next few years.
As part of the arrangement, AMD gave OpenAI a warrant for up to 160 million shares of AMD common stock.
The same month, AMD revealed work with the U.S. Department of Energy to build two AI supercomputers. It said the work represented a combined $1 billion investment of private and public funding.
In November, AMD released further details about its work with Cisco and HUMAIN, outlining plans to deploy up to 1 GW of AI infrastructure by 2030, starting with an initial 100 MW in Saudi Arabia which is set to be rolled out in 2026. A dollar figure was not attached to the partnership.
Anthropic
In October, Anthropic announced it would expand its use of Google’s AI chips, planning to use up to one million TPUs in an arrangement worth “tens of billions of dollars.”
In November, Anthropic unveiled plans to invest $50 billion in American computing infrastructure, starting with the construction of data centers with Fluidstack in Texas and New York.
The same month it inked a deal with Microsoft to scale its Claude model on Azure. Anthropic agreed to purchase $30 billion worth of compute capacity and Microsoft said it would invest up to $5 billion in Anthropic as part of the arrangement.
As part of the same deal, Anthropic announced Nvidia would invest up to $10 billion in the model maker.
Broadcom
In October, Broadcom announced a multi-year collaboration with OpenAI to deploy 10 gigawatts of custom AI accelerators starting in the second half of 2026.
Cisco
In May, Cisco signed a partnership deal with Saudi Arabia’s HUMAIN initiative to develop AI infrastructure in the country. The same month, it also inked an MoU with UAE company G42 to get the ball rolling on AI infrastructure deployments there.
Further information about the collaboration was released in November, with HUMAIN, Cisco and AMD outlining plans to deploy up to 1 GW of AI infrastructure by 2030, starting with an initial 100 MW in Saudi Arabia which is set to be rolled out in 2026. A dollar figure was not attached to the partnership.
Coreweave
In March 2025, CoreWeave announced an initial infrastructure supply agreement with OpenAI with a contract value up to $11.9 billion, followed by an expanded agreement worth up to $4 billion in May 2025. In September, the pair expanded their deal by another $6.5 billion, bringing the total contract value to approximately $22.4 billion.
In September, Coreweave revealed that it had inked a $6.3 billion deal with Nvidia to supply compute capacity to the chip giant through 2032.
G42
G42 is a government-backed AI technology company based in the United Arab Emirates that is part of the country’s national AI development strategy. It is building a 1 GW Stargate cluster in the country alongside Oracle, Cisco, NVIDIA and SoftBank Group, which will be part of a larger 5 GW AI infrastructure hub in the country. It also received a $1.5 billion equity investment from Microsoft in 2024.
In May, it inked an MoU with Cisco to get the ball rolling on AI infrastructure deployments there.
In October, G42 announced it had secured permission to receive exports of advanced AI chips from the U.S., opening the door for it to press ahead with its infrastructure projects.
In November, G42 said it was working with Microsoft on a 200 MW expansion of data center capacity, with compute expected to begin coming online before the end of 2026.
Google/Google Cloud
In October, Google locked in a deal to provide Anthropic with up to one million TPUs in an arrangement worth “tens of billions of dollars.”
HUMAIN
HUMAIN is Saudi Arabia's state-backed AI company, which was formed in May 2025 to help facilitate the deployment of AI infrastructure there.
In May, HUMAIN unveiled a deal with AWS to invest more than $5 billion in AI infrastructure, AWS services, and AI training and talent development in Saudi Arabia.
At the time, it also announced plans to work with Cisco on its data center rollouts.
Further information about the collaboration was released in November, with HUMAIN, Cisco and AMD outlining plans to deploy up to 1 GW of AI infrastructure by 2030, starting with an initial 100 MW in Saudi Arabia which is set to be rolled out in 2026. A dollar figure was not attached to the partnership.
In November, HUMAIN helped lead a $900 million funding round for frontier AI model company Luma AI.
The same month it also signed a deal to develop GPU data centers with Elon Musk’s xAI, starting with a 500 MW+ facility; struck a deal with sovereign AI infrastructure company Global AI to build data centers in the U.S.; and inked an agreement with AWS to deploy up to 150,000 AI accelerators (including AWS’ Trainium AI chip and Nvidia’s GB300 infrastructure) in a data center facility known as an “AI Zone” in Riyadh. The November AWS announcement was a follow up to the $5 billion deal announced in May.
Microsoft
In November, Microsoft inked a deal to help model maker Anthropic scale its Claude model on Azure. Anthropic agreed to purchase $30 billion worth of compute capacity and Microsoft said it would invest up to $5 billion in Anthropic as part of the arrangement.
The same month, Microsoft announced it had made a $1.5 billion equity investment in G42 and had spent more than $4.6 billion in capital expenses for advanced AI and cloud datacenters in the United Arab Emirates since 2023. It added it planned to spend another $7.9 billion in the country, with $5.5 billion going to the expansion of its AI and cloud infrastructure there.
Nvidia
Nvidia made a series of strategic investment deals in 2025, including one to invest up to $10 billion in model maker Anthropic and another to invest up to $100 billion in OpenAI. It also invested $1 billion each in telecom equipment maker Nokia and AI startup Poolside.
OpenAI
(See also: Stargate)
In March 2025,OpenAI announced an initial infrastructure supply agreement with CoreWeave with a contract value up to $11.9 billion. It expanded that agreement in May by $4 billion. In September, the pair expanded their deal by another $6.5 billion, bringing the total contract value to approximately $22.4 billion.
In July, OpenAI inked a deal worth more than $300 billion over five years with Oracle for compute infrastructure. The deal, which was reported and publicly acknowledged in September, appeared to be tied to OpenAI’s work with Oracle on U.S. Stargate campus deployments.
In September, OpenAI announced a deal with Nvidia to build and deploy at least 10 gigawatts of AI data center capacity using Nvidia systems, in a move that would require “millions” of Nvidia GPUs. As part of the deal, Nvidia said it would invest up to $100 billion in OpenAI.
In October, OpenAI inked a partnership deal with chipmaker AMD to deploy 6 GW worth of AMD’s Instinct GPUs, with an initial 1 GW set to roll out in the back half of 2026. The deal was said to be worth well over $100 billion over the next few years.
The same month, it announced a multi-year collaboration with Broadcom to deploy 10 gigawatts of custom AI accelerators starting in the second half of 2026.
In November, OpenAI inked a $38 billion deal to secure cloud compute capacity from Amazon.
Oracle
In May, the Financial Times reported Oracle struck a deal to spend $40 billion in Nvidia chips for its Stargate deployments.
In July, Oracle inked a deal worth more than $300 billion over five years with OpenAI for compute infrastructure. The deal, which was reported and publicly acknowledged in September, appeared to be tied to OpenAI’s work with Oracle on U.S. Stargate campus deployments.
In October, Oracle confirmed it had signed a $20 billion deal with Meta. At the time, Oracle co-CEO Clay Magouyrk said the company had inked $65 billion in new cloud infrastructure commitments over the past 30 days.
Just days before, Oracle announced a deal to deploy 50,000 AMD GPUs starting in calendar Q3 2026 and expanding in 2027 and beyond, though it did not attach a dollar figure to the agreement.
Stargate
Stargate is a joint venture formed by OpenAI and a series of partners in January 2025 to build AI compute infrastructure for the leading model provider. See our primer on Stargate projects here.