AMD Instinct: From MI300X to Helios and the Gigawatt Deals
AMD introduced the Instinct MI300X in June 2023 and launched it in December, then moved its accelerator roadmap to annual generations spanning MI325X, MI350, MI400 and later MI500 and MI600 series. MI300X reached Azure and Oracle offerings and was reported in production use by Microsoft and broad Llama inference deployment at Meta; AMD said Instinct revenue exceeded $5 billion in 2024. AMD has now launched the MI400 family and its Helios rack-scale system, and said MI455X and Helios shipments had begun by October 2026. The next test is execution: OpenAI and Meta each agreed to deploy up to 6 gigawatts of AMD GPUs and Anthropic up to 2 gigawatts, but their largest Helios deployments are scheduled to come online from late 2026 and into 2027.
Key moments
- Jun 13, 2023AMD unveils the Instinct MI300X with 192 GB of HBM3, sampling in Q3
- Dec 6, 2023MI300X and MI300A launched; Microsoft Azure, Oracle and the major OEMs line up
- Jun 2, 2024AMD moves Instinct to an annual cadence: MI325X in Q4 2024, MI350 in 2025, MI400 in 2026
- Jun 12, 2025MI350 series rolling out at Oracle; Helios rack on MI400 previewed for 2026
- Oct 6, 2025OpenAI agrees to deploy 6 gigawatts of AMD GPUs; AMD issues a warrant for up to 160 million shares
- Nov 11, 2025Financial Analyst Day: data center revenue to grow more than 60 percent a year; Helios from Q3 2026
- Feb 24, 2026Meta signs for up to 6 gigawatts of AMD GPUs with a custom MI450-based part and a matching warrant
- Jul 22, 2026Anthropic to deploy up to 2 gigawatts of MI450 GPUs in Helios, starting in the first half of 2027
- Jul 23, 2026MI400 series launched at Advancing AI 2026; Helios customers named
- Aug 4, 2026Data center revenue doubles to US$6.7 billion as Helios begins to ramp
- Sep 30, 2026HPE announces $1.2 billion Vultr order for AMD Helios systems
- Oct 7, 2026Lisa Su says MI455X and Helios shipments have already begun
In numbers
Data and sources (4 points)
| Date | AMD data center or Instinct revenue | Entry |
|---|---|---|
| 2024 | $5B | AMD reports more than $5 billion in 2024 Instinct revenue |
| 2025 | $16.6B | AMD reports $16.6 billion in 2025 Data Center revenue |
| May 5, 2026 | $5.8B | AMD reports $5.8 billion in first-quarter Data Center revenue |
| Aug 4, 2026 | $6.7B | Data center revenue doubles to US$6.7 billion as Helios begins to ramp |
Timeline
Earlier years show only their entries' dates and titles. Click a year heading to show the details.
202621 entries
- ShipmentKey moment
Lisa Su says MI455X and Helios shipments have already begun
Speaking in Korea on cooperation with Samsung and SK hynix, Lisa Su said AMD had already begun shipping its next-generation MI455X accelerators and Helios rack-scale systems. It is the first statement that the generation behind the OpenAI, Meta and Anthropic gigawatt agreements has started to leave the factory, rather than being tested or scheduled.
- Inven Global · Oct 7, 2026 ·
We have already begun shipments of our next-generation MI455X and AMD Helios solutions.
- Inven Global · Oct 7, 2026 ·
- Report
AMD says customers are testing Helios systems at their own sites
AMD said customers were testing AMD Helios systems at their own sites. It said the testing included workload validation and preparation for deployments at scale.
- AMD Newsroom · Oct 5, 2026 ·
Customers are testing AMD Helios at their own sites – validating workloads and preparing deployments at scale – bringing the platform closer to powering production AI infrastructure.
Editor's note: The cited AMD Newsroom page explicitly says customers are testing AMD Helios at their own sites, validating workloads and preparing deployments at scale. The readable page text does not provide an event publication date to support 2026-10-05.
- AMD Newsroom · Oct 5, 2026 ·
- AnnouncementKey moment
HPE announces $1.2 billion Vultr order for AMD Helios systems
Hewlett Packard Enterprise announced a $1.2 billion order from Vultr to deploy AMD Helios AI Rack by HPE systems at Vultr's cloud data centres in the United States.
- Hewlett Packard Enterprise via Business Wire · Sep 30, 2026 ·
HPE (NYSE: HPE) today announced a $1.2 billion order from Vultr, the world’s largest privately-held cloud infrastructure company, to deploy AMD Helios™ AI Rack by HPE systems at Vultr’s cloud data center locations in the United States.
- Reuters · Sep 30, 2026 ·
HPE said it will provide AMD Helios AI racks, equipped with AI chips and HPE's networking switches and software, to Vultr's US data centers.
Editor's note: Editor: dropped a Reuters sentence about what the racks would include, which HPE's own announcement does not carry.
- Hewlett Packard Enterprise via Business Wire · Sep 30, 2026 ·
- Report
Hewlett Packard Enterprise says it will provide Helios trays
Hewlett Packard Enterprise said it would provide trays to the market. The report described the trays as a white-label supply route for server competitors working with AMD on Helios.
- Hewlett Packard Enterprise · Sep 10, 2026 ·
“We're also going to be providing to the market trays.”
Editor's note: The transcript is readable and states that HPE will provide trays to the market on a white-label basis to server competitors working with AMD on Helios. The event transcript is dated September 10, 2026.
- Hewlett Packard Enterprise · Sep 10, 2026 ·
- Financial
AMD expects Data Center sales to double in 2027
Advanced Micro Devices said it expected its Data Center business to double in 2027 to about $70 billion in sales. The company said it had undertaken supply-chain work to support doubling Data Center revenue the following year. Su added that all three anchor customers for MI450 and Helios had given forecasts above AMD's initial expectations, and that AMD was working with suppliers to support them.
- Investing.com · Sep 8, 2026 ·
“The company expects its data center business to double in 2027, reaching about $70 billion in sales.”
- EarningsCalls.dev · Sep 8, 2026 ·
“So right now, we actually feel really good about all the supply chain work we have done to ensure we can support the data center revenue to double next year.”
Editor's note: The page states that AMD expects its data center business to double in 2027 to about $70 billion in sales. It also says AMD had been working with its supply chain to support doubling data center revenue next year, and identifies the conference date as September 8, 2026.
- Investing.com · Sep 8, 2026 ·
- Shipment
AMD says MI450 production shipments begin in Q3 2026
Advanced Micro Devices said Instinct MI450 production shipments were set to begin in the third quarter of 2026. It said revenue would step up in the fourth quarter of 2026 and again in the first quarter of 2027.
- EarningsCalls.dev · Sep 8, 2026 ·
“Q3 will be the -- at the very beginning of our production shipment we expect revenue in Q3, but Q4, we'll see a very significant step up and then another step up in Q1 2027 and ramp through 2027.”
- Investing.com · Sep 8, 2026 ·
“The AI GPU business continues to ramp, with MI450 set to begin production shipments in Q3 2026.”
Editor's note: The transcript states that the MI450/Helios ramp would begin with production shipments and revenue in Q3 2026, followed by a very significant Q4 increase and another increase in Q1 2027. The page is dated 2026-09-08, matching the node date.
- EarningsCalls.dev · Sep 8, 2026 ·
- FinancialKey moment
Data center revenue doubles to US$6.7 billion as Helios begins to ramp
Second-quarter 2026 results showed revenue of US$11.5 billion and data center segment revenue of US$6.7 billion, up 107 percent from a year earlier on demand for EPYC CPUs and Instinct GPUs. Lisa Su said Instinct deployments were scaling and Helios was beginning to ramp.
- AMD Investor Relations · Aug 4, 2026 ·
Data Center segment revenue was $6.7 billion, up 107% year-over-year, driven by strong demand for AMD EPYC™ processors and AMD Instinct™ GPUs.
- AMD Investor Relations · Aug 4, 2026 ·
- Announcement
AMD schedules Instinct MI600-series GPUs for 2028
AMD said Instinct MI600-series GPUs were coming in 2028. The statement extended AMD's disclosed accelerator roadmap beyond the MI500 series planned for 2027.
- AMD · Jul 23, 2026 ·
AMD Instinct MI600 Series GPUs are coming in 2028.
Editor's note: The page states that AMD Instinct MI600 Series GPUs are coming in 2028 and identifies MI500 Series GPUs as coming in 2027. It is dated July 23, 2026, matching the node date.
- AMD · Jul 23, 2026 ·
- Announcement
AMD says Meta is validating workloads on Helios racks
AMD said Meta had begun testing and validating workloads on AMD Helios racks. AMD also said Meta was validating sixth-generation EPYC CPU platforms in its labs.
- AMD · Jul 23, 2026 ·
Meta is now validating 6th Gen EPYC CPU platforms in its labs and has begun testing and validating workloads on AMD Helios racks as they prepare to deploy at scale.
Editor's note: The page states that Meta is validating 6th Gen EPYC CPU platforms in its labs and has begun testing and validating workloads on AMD Helios racks. It is dated July 23, 2026.
- AMD · Jul 23, 2026 ·
- Announcement
AMD says OpenAI expects Helios to begin operating in Q4 2026
AMD said OpenAI expected to bring Helios online beginning in the fourth quarter of 2026. AMD said deployments were expected to accelerate throughout 2027.
- AMD · Jul 23, 2026 ·
OpenAI expects to bring Helios online beginning in the fourth quarter of 2026, with deployments accelerating throughout 2027.
Editor's note: The AMD press release is dated July 23, 2026 and states that OpenAI expects to bring Helios online beginning in the fourth quarter of 2026, with deployments accelerating throughout 2027.
- AMD · Jul 23, 2026 ·
- Launch
AMD launches the Instinct MI350P GPU
AMD launched the Instinct MI350P GPU at Advancing AI 2026. AMD said the GPU brought AI acceleration to existing infrastructure.
- AMD · Jul 23, 2026 ·
AMD also launched the Instinct MI350P GPU, bringing seamless AI acceleration to existing infrastructure with leadership token economics.
- AMD · Aug 4, 2026 ·
Launched the AMD Instinct MI350P GPU, designed to bring seamless AI acceleration to existing infrastructure with leadership token economics.
Editor's note: The press release states that AMD launched the Instinct MI350P GPU at Advancing AI 2026 and includes the quoted sentence. It is dated July 23, 2026, matching the node date.
- AMD · Jul 23, 2026 ·
- LaunchKey moment
MI400 series launched at Advancing AI 2026; Helios customers named
AMD launched the Instinct MI400 series, with the MI455X in Helios racks for frontier AI and the MI430X for sovereign AI and HPC, and named OpenAI, Anthropic, Meta, Microsoft, Oracle, HUMAIN, Tensorwave, Vultr and Cirrascale among the organisations choosing Helios.
- AMD Investor Relations · Jul 23, 2026 ·
Leading AI labs and cloud providers are choosing AMD Helios for its open, full-stack performance.
- AMD Investor Relations · Jul 23, 2026 ·
- Announcement
AMD says Anthropic was already using Instinct MI355X GPUs
AMD said Anthropic was using AMD Instinct MI355X GPUs before its planned MI450-series deployment. AMD described the planned deployment as building on that existing use.
- AMD · Jul 22, 2026 ·
This deployment builds on Anthropic’s existing use of AMD Instinct MI355X GPUs.
Editor's note: The page states that the MI450-series deployment “builds on Anthropic’s existing use of AMD Instinct MI355X GPUs.” It is dated July 22, 2026.
- AMD · Jul 22, 2026 ·
- Financial
AMD commits up to $5 billion for a future Anthropic investment
AMD committed to make a strategic equity investment of up to $5 billion in Anthropic in the future. The commitment was disclosed in AMD's Anthropic partnership announcement.
- AMD · Jul 22, 2026 ·
AMD has committed to make a strategic equity investment of up to $5 billion in Anthropic in the future.
Editor's note: The AMD press release states that AMD committed to make a strategic equity investment of up to $5 billion in Anthropic in the future. It is dated July 22, 2026, matching the node date.
- AMD · Jul 22, 2026 ·
- AnnouncementKey moment
Anthropic to deploy up to 2 gigawatts of MI450 GPUs in Helios, starting in the first half of 2027
AMD and Anthropic announced a partnership to deploy up to 2 gigawatts of Instinct MI450 series GPUs in AMD Helios rack-scale systems, with the first gigawatt beginning in the first half of 2027.
- AMD Investor Relations · Jul 22, 2026 ·
with deployment of the first gigawatt beginning in the first half of 2027
- AMD Investor Relations · Jul 22, 2026 ·
- Announcement
Microsoft commits to ramp Helios at scale on Azure
AMD said Microsoft would ramp AMD Helios at scale on Azure. The deployment was intended to power frontier-model inference for Microsoft, its AI customers and Azure AI services.
- AMD · Jul 20, 2026 ·
Microsoft to ramp AMD Helios™ at scale on Azure to power frontier model inference for Microsoft, its AI customers and Azure AI services.
Editor's note: The AMD press release states that Microsoft will ramp AMD Helios at scale on Azure for frontier-model inference for Microsoft, its AI customers, and Azure AI services. It is dated and released July 20, 2026.
- AMD · Jul 20, 2026 ·
- Financial
AMD reports $5.8 billion in first-quarter Data Center revenue
AMD reported $5.8 billion in first-quarter 2026 Data Center segment revenue, up 57% year over year. AMD attributed the result to EPYC processor demand and the continued ramp of Instinct GPU shipments.
- AMD · May 5, 2026 ·
Data Center segment revenue was $5.8 billion, up 57% year-over-year, driven by strong demand for AMD EPYC™ processors and the continued ramp of AMD Instinct™ GPU shipments.
Editor's note: The page states that AMD reported first-quarter 2026 Data Center segment revenue of $5.8 billion, up 57% year over year, driven by EPYC processor demand and the ramp of Instinct GPU shipments. It is dated May 05, 2026.
- AMD · May 5, 2026 ·
- AnnouncementKey moment
Meta signs for up to 6 gigawatts of AMD GPUs with a custom MI450-based part and a matching warrant
AMD and Meta agreed a multi-year, multi-generation partnership to deploy up to 6 gigawatts of Instinct GPUs. Shipments for the first gigawatt were expected to begin in the second half of 2026 using a custom GPU based on the MI450 architecture, and AMD issued Meta a performance-based warrant for up to 160 million shares.
- AMD Investor Relations · Feb 24, 2026 ·
AMD has issued Meta a performance-based warrant for up to 160 million shares of AMD common stock, structured to vest as specific milestones associated with Instinct GPU shipments are achieved.
- AMD Form 8-K · Feb 24, 2026 ·
On February 23, 2026, in connection with a strategic arrangement between Advanced Micro Devices, Inc. (“AMD”) and Meta Platforms, Inc. (“Meta”) governing the purchase of AMD Instinct™ GPU products by Meta, AMD issued to Meta a performance-based warrant (the “Warrant”) to purchase up to an aggregate of 160 million shares of common stock of AMD (the “Warrant Shares”) at an exercise price of $0.01 per share.
is contradicted by Meta signs a multibillion-dollar deal to rent Google Cloud TPUs for its next models on Custom AI Silicon at the Hyperscalers: TPU, Trainium, Maia, MTIA and the OpenAI–Broadcom Chip.
- AMD Investor Relations · Feb 24, 2026 ·
- Financial
AMD reports $16.6 billion in 2025 Data Center revenue
AMD reported full-year 2025 Data Center segment revenue of $16.6 billion. It said the result reflected growth in EPYC CPUs and Instinct GPUs.
- AMD · Feb 3, 2026 ·
For the full year 2025, Data Center segment revenue was a record $16.6 billion, up 32% year-over-year, reflecting growth across both EPYC CPUs and AMD Instinct GPUs.
Editor's note: The page states that full-year 2025 Data Center segment revenue was a record $16.6 billion, up 32% year over year, reflecting growth across EPYC CPUs and AMD Instinct GPUs. The press release is dated February 3, 2026.
- AMD · Feb 3, 2026 ·
- Financial
AMD releases $360 million of reserved MI308 inventory charges
AMD said it benefited from an approximate $360 million release of previously reserved MI308 inventory and related charges in the fourth quarter. The release concerned AMD Instinct MI308 inventory.
- AMD · Feb 3, 2026 ·
In the fourth quarter, AMD benefited from an approximate $360 million release of previously reserved AMD Instinct™ MI308 inventory and related charges.
result of Q2 2025: the $800 million MI308 charge lands; no China revenue in the outlook on US Export Controls on AMD's AI Chips in China: From the MI250 License to MI308 and MI325X.
Editor's note: The page states that AMD benefited from an approximate $360 million release of previously reserved AMD Instinct MI308 inventory and related charges in the fourth quarter. The press release is dated February 3, 2026.
- AMD · Feb 3, 2026 ·
- Announcement
AMD introduces MI440X and previews the MI500 series
AMD introduced the Instinct MI440X GPU for enterprise deployments. It also previewed the next-generation Instinct MI500 Series GPUs.
- AMD · Jan 5, 2026 ·
AMD expanded its AI portfolio with the introduction of the AMD Instinct MI440X GPU for enterprise deployments and previewed the next-generation Instinct MI500 Series GPUs
Editor's note: The page states that AMD introduced the Instinct MI440X GPU for enterprise deployments and previewed the next-generation MI500 Series GPUs. It is dated January 05, 2026, matching the node date.
- AMD · Jan 5, 2026 ·
20258 entries
- FinancialKey moment
Financial Analyst Day: data center revenue to grow more than 60 percent a year; Helios from Q3 2026
AMD set long-term targets of more than 35 percent annual revenue growth, a non-GAAP operating margin above 35 percent and earnings per share above US$20, with data center revenue growing more than 60 percent a year. It said Helios systems with MI450 GPUs would begin delivering in the third quarter of 2026.
- AMD Investor Relations · Nov 11, 2025 ·
AMD expects to deliver a greater than 60% revenue CAGR for its data center business.
result of OpenAI agrees to deploy 6 gigawatts of AMD GPUs; AMD issues a warrant for up to 160 million shares.
- AMD Investor Relations · Nov 11, 2025 ·
- Regulatory
AMD reports no MI308 shipment revenue to China in Q3
AMD said its third-quarter results included no revenue from MI308 GPU shipments to China. The disclosure concerned AMD Instinct MI308 products.
- AMD · Nov 4, 2025 ·
Our third quarter results did not include any revenue from shipments of AMD Instinct™ MI308 GPU products to China.
Editor's note: The page states that third-quarter 2025 results did not include revenue from shipments of AMD Instinct MI308 GPU products to China. It is dated November 04, 2025.
- AMD · Nov 4, 2025 ·
- Announcement
Oracle plans a public 50,000-GPU MI450 supercluster
Image: AMD, press image · source · editorial use AMD said Oracle would be the first hyperscaler to offer a publicly available AI supercluster powered by 50,000 MI450 Series GPUs. The offering was scheduled to begin in the third quarter of 2026.
- AMD Newsroom · Oct 14, 2025 ·
Beginning in the third quarter of 2026, Oracle will be the first hyperscaler to offer a publicly available AI supercluster powered by 50,000 AMD Instinct MI450 Series GPUs.
Editor's note: Quote confirmed verbatim on the AMD newsroom page dated October 14, 2025 (editor re-check); the automated pass misread the rendered text.
- AMD Newsroom · Oct 14, 2025 ·
- AnnouncementKey moment
OpenAI agrees to deploy 6 gigawatts of AMD GPUs; AMD issues a warrant for up to 160 million shares
AMD and OpenAI announced a multi-year agreement under which OpenAI will deploy 6 gigawatts of AMD Instinct GPUs across generations, starting with 1 gigawatt of MI450 GPUs in the second half of 2026. AMD issued OpenAI a warrant for up to 160 million shares at US$0.01 that vests with deployment milestones and AMD share-price targets.
- AMD Investor Relations · Oct 6, 2025 ·
The first 1 gigawatt deployment of AMD Instinct MI450 GPUs is set to begin in the second half of 2026.
led to Financial Analyst Day: data center revenue to grow more than 60 percent a year; Helios from Q3 2026.
- AMD Investor Relations · Oct 6, 2025 ·
- Announcement
Microsoft says MI300X powers production models on Azure
AMD reported that Microsoft said Instinct MI300X was powering proprietary and open-source models in production on Azure. The statement concerned Azure production workloads.
- AMD · Jun 12, 2025 ·
Microsoft announced Instinct MI300X is now powering both proprietary and open-source models in production on Azure.
Editor's note: The page states that Microsoft announced Instinct MI300X is powering both proprietary and open-source models in production on Azure. The press release is dated June 12, 2025.
- AMD · Jun 12, 2025 ·
- Announcement
Meta says MI300X is broadly deployed for Llama inference
AMD reported that Meta detailed broad MI300X deployment for Llama 3 and Llama 4 inference. The statement concerned AMD Instinct MI300X deployments.
- AMD · Jun 12, 2025 ·
Meta detailed how Instinct MI300X is broadly deployed for Llama 3 and Llama 4 inference.
Editor's note: The press release states that “Meta detailed how Instinct MI300X is broadly deployed for Llama 3 and Llama 4 inference.” It is dated June 12, 2025.
- AMD · Jun 12, 2025 ·
- LaunchKey moment
MI350 series rolling out at Oracle; Helios rack on MI400 previewed for 2026
At Advancing AI 2025 AMD said MI350 series accelerators were already deploying at Oracle Cloud Infrastructure with broad availability in the second half of 2025, and previewed Helios, a rack-scale system built on the coming MI400 series GPUs, Zen 6 EPYC "Venice" CPUs and Pensando "Vulcano" NICs.
- AMD Investor Relations · Jun 12, 2025 ·
AMD demonstrated end-to-end, open-standards rack-scale AI infrastructure—already rolling out with AMD Instinct MI350 Series accelerators, 5th Gen AMD EPYC™ processors and AMD Pensando™ Pollara NICs in hyperscaler deployments such as Oracle Cloud Infrastructure (OCI) and set for broad availability in 2H 2025.
- AMD Investor Relations · Jun 12, 2025 ·
- Financial
AMD reports more than $5 billion in 2024 Instinct revenue
AMD said it delivered more than $5 billion of AMD Instinct accelerator revenue during 2024. It said Data Center annual revenue nearly doubled as EPYC adoption accelerated.
- AMD · Feb 4, 2025 ·
Data Center segment annual revenue nearly doubled as EPYC processor adoption accelerated and we delivered more than $5 billion of AMD Instinct accelerator revenue.
Editor's note: The page states that AMD delivered more than $5 billion of AMD Instinct accelerator revenue during 2024 and that Data Center segment annual revenue nearly doubled as EPYC adoption accelerated. It is dated February 4, 2025.
- AMD · Feb 4, 2025 ·
- Nov 11, 2025Financial Analyst Day: data center revenue to grow more than 60 percent a year; Helios from Q3 2026
- Nov 4, 2025AMD reports no MI308 shipment revenue to China in Q3
- Oct 14, 2025Oracle plans a public 50,000-GPU MI450 supercluster
- Oct 6, 2025OpenAI agrees to deploy 6 gigawatts of AMD GPUs; AMD issues a warrant for up to 160 million shares
- Jun 12, 2025Microsoft says MI300X powers production models on Azure
- Jun 12, 2025Meta says MI300X is broadly deployed for Llama inference
- Jun 12, 2025MI350 series rolling out at Oracle; Helios rack on MI400 previewed for 2026
- Feb 4, 2025AMD reports more than $5 billion in 2024 Instinct revenue
20245 entries
- Launch
MI325X launched at Advancing AI 2024; MI350 set for the second half of 2025
AMD launched the Instinct MI325X, shared details of the MI350 series expected in the second half of 2025, and said the MI400 series based on CDNA Next remained planned for 2026.
- AMD Investor Relations · Oct 10, 2024 ·
AMD also shared new details on next-gen AMD Instinct MI350 series accelerators expected to launch in the second half of 2025, extending AMD Instinct leadership memory capacity and generative AI performance.
- AMD Investor Relations · Oct 10, 2024 ·
- Launch
Fireworks AI adopts OCI bare-metal MI300X instances
AMD said OCI bare-metal instances using MI300X accelerators had been adopted by Fireworks AI. The instances were part of Oracle Cloud Infrastructure's MI300X offering.
- AMD · Sep 26, 2024 ·
These OCI bare metal instances have already been adopted by companies including Fireworks AI.
- Advanced Micro Devices, Inc. (AMD) Investor Relations · Sep 26, 2024 ·
Designed to run demanding AI workloads including large language model (LLM) inference and training that requires high throughput with leading memory capacity and bandwidth, these OCI bare metal instances have already been adopted by companies including Fireworks AI.
Editor's note: The AMD press release states that OCI chose MI300X accelerators for its BM.GPU.MI300X.8 bare-metal instance and that these instances had already been adopted by companies including Fireworks AI. It is dated September 26, 2024. Official source added: https://ir.amd.com/news-events/press-releases/detail/1217/amd-instinct-mi300x-accelerators-available-on-oracle-cloud-infrastructure-for-demanding-ai-applications
- AMD · Sep 26, 2024 ·
- Financial
AMD attributes sequential revenue growth to Instinct GPU shipments
AMD reported that revenue increased 21% sequentially. It attributed the increase primarily to the ramp of AMD Instinct GPU shipments.
- AMD · Jul 30, 2024 ·
Revenue increased 21% sequentially primarily driven by the strong ramp of AMD Instinct GPU shipments.
Editor's note: The page is readable and states that Data Center segment revenue increased 21% sequentially, primarily driven by the strong ramp of AMD Instinct GPU shipments. It is dated July 30, 2024.
- AMD · Jul 30, 2024 ·
- AnnouncementKey moment
AMD moves Instinct to an annual cadence: MI325X in Q4 2024, MI350 in 2025, MI400 in 2026
At Computex in Taipei AMD laid out a yearly roadmap for its AI accelerators: the MI325X with 288 GB of HBM3E in the fourth quarter of 2024, the CDNA 4-based MI350 series in 2025, and the MI400 series on a CDNA "Next" architecture in 2026.
- AMD Investor Relations · Jun 2, 2024 ·
New AMD Instinct MI325X accelerator expected to be available in Q4 2024 with up to 288GB of HBM3E memory
- AMD Investor Relations · Jun 2, 2024 ·
- Launch
Azure ND MI300X V5 instances become generally available
AMD said Azure ND MI300X V5 instances were generally available. Hugging Face was identified as the first customer.
- AMD · May 21, 2024 ·
The new Azure ND MI300X V5 instances are now generally available, with Hugging Face as the first customer
- Advanced Micro Devices, Inc. (AMD) · May 21, 2024 ·
The new Azure ND MI300X V5 instances are now generally available, with Hugging Face as the first customer.
Editor's note: The AMD press release states that the new Azure ND MI300X V5 instances are generally available and names Hugging Face as the first customer. It is dated May 21, 2024. Official source added: https://ir.amd.com/news-events/press-releases/detail/1198/amd-instinct-mi300x-accelerators-power-microsoft-azure-openai-service-workloads-and-new-azure-nd-mi300x-v5-vms
- AMD · May 21, 2024 ·
- Oct 10, 2024MI325X launched at Advancing AI 2024; MI350 set for the second half of 2025
- Sep 26, 2024Fireworks AI adopts OCI bare-metal MI300X instances
- Jul 30, 2024AMD attributes sequential revenue growth to Instinct GPU shipments
- Jun 2, 2024AMD moves Instinct to an annual cadence: MI325X in Q4 2024, MI350 in 2025, MI400 in 2026
- May 21, 2024Azure ND MI300X V5 instances become generally available
20232 entries
- LaunchKey moment
MI300X and MI300A launched; Microsoft Azure, Oracle and the major OEMs line up
At Advancing AI 2023 AMD announced availability of the MI300X accelerator and the MI300A APU. Microsoft announced Azure virtual machines built on MI300X, Oracle Cloud planned MI300X offerings, and Dell, HPE, Lenovo and Supermicro showed systems using the parts.
- AMD Investor Relations · Dec 6, 2023 ·
Today, AMD announced the availability of the AMD Instinct™ MI300X accelerators – with industry leading memory bandwidth for generative AI and leadership performance for large language model (LLM) training and inferencing – as well as the AMD Instinct™ MI300A accelerated processing unit (APU).
- AMD Investor Relations · Dec 6, 2023 ·
- AnnouncementKey moment
AMD unveils the Instinct MI300X with 192 GB of HBM3, sampling in Q3
At its Data Center and AI Technology Premiere AMD introduced the MI300X, a GPU-only version of the MI300 family built on the CDNA 3 architecture with up to 192 GB of HBM3 memory, aimed at large language model training and inference, and said it would sample to key customers from the third quarter.
- AMD Investor Relations · Jun 13, 2023 ·
The MI300X is sampling to key customers starting in Q3.
- AMD Investor Relations · Jun 13, 2023 ·
Questions this timeline answers
Each one is an ordered set of dated entries, most of them from this page, with a line on why each step led to the next.
Cause and effect
- OpenAI agrees to deploy 6 gigawatts of AMD GPUs; AMD issues a warrant for up to 160 million sharesled to →Financial Analyst Day: data center revenue to grow more than 60 percent a year; Helios from Q3 2026The OpenAI agreement underpinned the 60 percent data center growth target set five weeks later.
- Q2 2025: the $800 million MI308 charge lands; no China revenue in the outlook on US Export Controls on AMD's AI Chips in China: From the MI250 License to MI308 and MI325Xled to →AMD releases $360 million of reserved MI308 inventory chargesAMD's roughly $800 million MI308 inventory and related charges in the second quarter of 2025 created the previously reserved MI308 inventory and related charges that AMD later released by about $360 million.
- Meta signs for up to 6 gigawatts of AMD GPUs with a custom MI450-based part and a matching warrantis contradicted by →Meta signs a multibillion-dollar deal to rent Google Cloud TPUs for its next models on Custom AI Silicon at the Hyperscalers: TPU, Trainium, Maia, MTIA and the OpenAI–Broadcom ChipThe entries describe Meta's February 2026 AMD agreement with different accelerator generations: Timeline B says its custom part is based on MI450, while Timeline A calls the announced purchase MI400 GPUs.
Where sources disagree or fall silent
- AMD has said MI450 production shipments and MI455X/Helios shipments have begun, but it has not disclosed shipped volume, recognized revenue, or deployed capacity under the OpenAI, Meta and Anthropic agreements.
- The sources give broad timing windows rather than precise operating dates for OpenAI's first Helios capacity in the fourth quarter of 2026 and Anthropic's first gigawatt in the first half of 2027.
- AMD's disclosures use MI400, MI450 and MI455X terminology for closely related 2026 products and deployments, but do not fully explain the product-family relationship among those names.
- AMD reported standalone Instinct revenue above $5 billion for 2024, but its 2025 and 2026 Data Center revenue figures also include EPYC CPUs, so the later accelerator-only contribution is not disclosed.
- The sources do not establish whether MI308 shipments to China resumed after AMD reported no third-quarter 2025 shipment revenue and released previously reserved inventory charges in the fourth quarter.
What to watch next
- Dec 2026 — OpenAI expects Helios to begin operating in the fourth quarter of 2026 This is the first operating milestone in OpenAI's planned 6-gigawatt AMD deployment; AMD says deployments are expected to accelerate during 2027. Dated to the end of the quarter rather than its start, because AMD says only "beginning in the fourth quarter of 2026" and a month-precision October would retire this row before the window closes.
- Jan 2027 — Anthropic's first 1-gigawatt MI450 Helios deployment is due to begin in the first half of 2027 It is the initial delivery milestone under Anthropic's agreement to deploy up to 2 gigawatts of MI450 GPUs in Helios systems.
Who is involved
- AMDcompany
Designer and supplier of Instinct accelerators and Helios systems, pursuing annual AI-product releases and large customer deployments.
- Lisa Superson
AMD chief executive who has described Instinct and Helios shipments, supply-chain preparations and Data Center growth expectations.
- Heliosproduct
AMD's rack-scale AI system combining Instinct GPUs, Zen 6 EPYC CPUs and Pensando networking; it is the platform for the planned large deployments.
- AMD Heliosproduct
AMD's rack-scale AI system combining MI400/MI455X GPUs with Zen 6 EPYC CPUs and Pensando networking; it is the delivery vehicle for the largest announced deployments.
- Instinct MI450product
The GPU inside Helios and the gigawatt deals.
- AMD EPYCproduct
AMD server CPUs used in Helios and a material contributor, alongside Instinct GPUs, to reported Data Center revenue growth.
- Microsoftcompany
Made MI300X Azure instances available, said MI300X powered production models on Azure, and committed to ramp Helios at scale for Azure.
- Oraclecompany
Offered MI300X cloud instances, deployed MI350-series accelerators in OCI, and plans a public 50,000-GPU MI450 supercluster.
26 more
- AMD Instinctproduct
AMD's data center AI accelerator line, progressing from MI300X through the MI400 generation and planned later generations.
- AMD Instinct MI300Xproduct
AMD's 2023 large-language-model accelerator, deployed through Azure and Oracle Cloud and used by Meta, Microsoft and Anthropic.
- Metacompany
Reportedly uses MI300X broadly for Llama inference and agreed to deploy up to 6 gigawatts of AMD GPUs, beginning with a custom MI450-based part; it also received a warrant for up to 160 million AMD shares.
- AMD Instinct MI350 seriesproduct
The 2025 generation, including MI355X and later MI350P, which AMD said was deploying at Oracle and already used by Anthropic.
- Anthropiccompany
Already used MI355X GPUs and agreed to deploy up to 2 gigawatts of MI450 GPUs in Helios systems; AMD also committed up to $5 billion for a future Anthropic equity investment.
- OpenAIcompany
Agreed to deploy 6 gigawatts of AMD Instinct GPUs, beginning with 1 gigawatt of MI450-class capacity, and received a performance-based warrant for up to 160 million AMD shares.
- AMD Instinct MI308product
China-oriented Instinct part for which AMD reported no third-quarter 2025 China shipment revenue and later released reserved inventory charges.
- AMD Instinct MI400 series and MI450 GPUsproduct
The MI400 generation launched in July 2026; MI455X is the Helios frontier-AI GPU, while the OpenAI, Meta and Anthropic agreements refer to MI450-series deployments or architecture.
- AMD Instinct MI500product
Next-generation GPU series previewed by AMD.
- AMD ROCmproduct
AMD software platform referenced in planned MI455X and Helios configurations.
- Hewlett Packard Enterprisecompany
Said it would supply Helios trays and announced its first Helios-system order, a $1.2 billion deployment for Vultr.
- Instinct MI300Xproduct
2023 generation, 192 GB HBM3.
- AMD Instinct MI350Pproduct
GPU launched for existing infrastructure.
- AMD Instinct MI355Xproduct
GPU that AMD said Anthropic was already using.
- AMD Instinct MI440Xproduct
An accelerator introduced for enterprise deployments in January 2026.
- AMD Instinct MI450product
The architecture and product designation used in the OpenAI, Meta and Anthropic deployment agreements and in AMD's production-shipment updates.
- AMD Instinct MI600product
Future GPU series AMD said was coming in 2028.
- AMD Pensandoproduct
AMD networking product line named in the Helios 600 roadmap.
- Chinaother
Market affected by U.S. restrictions on the MI308, with no MI308 shipment revenue reported by AMD in the third quarter of 2025.
- Fireworks AIcompany
Adopted OCI bare-metal instances using MI300X accelerators.
- Hugging Facecompany
Identified as the first customer for Azure ND MI300X V5 instances.
- Llama 3product
Model used for inference on broadly deployed MI300X systems.
- Llama 4product
Model used for inference on broadly deployed MI300X systems.
- Tritonproduct
OpenAI framework referenced in AMD's Helios workload-optimization statement.
- TSMCcompany
Supply-chain company named in AMD's Data Center growth discussion.
- Vultrcompany
Ordered $1.2 billion of HPE-built AMD Helios AI Rack systems for U.S. cloud data centers.
Changelog
- Added 1 entry: 2026-10-05 AMD says customers are testing Helios systems at their own sites. Added 1 hand-verified entry by the editor.
- Linked 1 entry with Custom AI Silicon at the Hyperscalers: TPU, Trainium, Maia, MTIA and the OpenAI–Broadcom Chip.
- Linked 1 entry with US Export Controls on AMD's AI Chips in China: From the MI250 License to MI308 and MI325X. Added 1 entry: 2026-09-30 HPE announces $1.2 billion Vultr order for AMD Helios systems.
- Added 1 entry: 2026-09-10 Hewlett Packard Enterprise says it will provide Helios trays.
- Added 2 entries: 2026-09-08 AMD says MI450 production shipments begin in Q3 2026; 2026-09-08 AMD expects Data Center sales to double in 2027.
- Added 19 entries: 2024-05-21 Azure ND MI300X V5 instances become generally available; 2024-07-30 AMD attributes sequential revenue growth to Instinct GPU shipments; 2024-09-26 Fireworks AI adopts OCI bare-metal MI300X instances; 2025-02-04 AMD reports more than $5 billion in 2024 Instinct revenue and 15 more. Added 11 hand-verified entries by the editor.
A research pass records each dated fact with the page it came from and a verbatim quote. A second, independent pass re-opens every cited page and rejects entries it cannot confirm. The editor reviews what remains before it is published, and every change is listed above. Full method.