NVIDIA is expanding its AI reach across data centers, personal computers, cars, and robots. On August 31, 2026, the company and MediaTek announced a broader partnership to build AI platforms that connect edge devices with cloud infrastructure.
The agreement includes a $3.5 billion NVIDIA investment in convertible bonds issued by MediaTek. It also brings MediaTek into NVIDIA’s NVLink Fusion ecosystem, where the two companies will work on custom AI infrastructure and other computing platforms.
“AI is transforming every computing platform — from the world’s largest AI factories to the PC and the car,” NVIDIA CEO Jensen Huang said.
One Partnership, Three Areas of Work
NVIDIA and MediaTek outlined three areas for the expanded agreement: AI infrastructure, local AI computing, and automotive systems. Within the NVLink Fusion ecosystem, MediaTek will help develop custom AI infrastructure, extending the partnership beyond individual chips and into larger systems.
For local AI computing, the companies will collaborate on chips for NVIDIA RTX Spark and DGX Spark PCs. These systems are aimed at bringing AI computing closer to users instead of placing every task inside a distant data center.
The automotive work will focus on platforms for AI-powered, software-defined vehicles. That effort places MediaTek and NVIDIA together in a market where vehicle functions depend more on software, computing hardware, and AI systems.
MediaTek Vice Chairman and CEO Rick Tsai joins Huang as a key figure in the expanded relationship. Together, the companies are linking semiconductor design with NVIDIA’s broader platform approach.
New Inference Hardware Joins the Expansion
NVIDIA’s Groq 3 LPX inference accelerator entered full production on August 24, 2026. In benchmarking with Gemma 4 31B, the accelerator delivered 3,400 output tokens per second, setting a record for the test.
Nebius is the first AI cloud provider to adopt NVIDIA Groq 3 LPX. NVIDIA says the accelerator enables four-times-faster responsiveness for agents and other workloads where response delays matter. The figures tied to the system also include a 100,000-token context for agentic systems.
“Generation is the phase of inference that determines how responsive an AI system actually is, and that’s exactly what NVIDIA Groq 3 LPX is built to accelerate,” Danila Shtan said.
The focus on inference reflects a shift in AI infrastructure. Training models requires enormous computing capacity, but the systems people use must also produce answers, actions, and outputs without long delays. NVIDIA is building products for both sides of that demand.
Amazon Adds Millions More NVIDIA GPUs
Amazon announced on August 26, 2026, that it will add another 2 million NVIDIA GPU chips to its data centers in 2027 and 2028. The chips will include NVIDIA Blackwell Ultra, Rubin, and Rubin Ultra GPUs.
The new plan follows an earlier agreement to deploy more than 1 million NVIDIA GPUs across AWS infrastructure starting in 2026. Demand from Amazon has exceeded expectations since that agreement, creating another large expansion for NVIDIA’s data center business.
NVIDIA’s technology will spread across AWS, including networking hardware, open models, CPUs, data processing software, and its robotics platform. NVIDIA also plans to send an unspecified number of Vera CPUs to Amazon, with some integrated with Rubin and others used as standalone processors.
NVIDIA expects Vera to be deployed by every major hyperscaler, neocloud, AI lab, and system OEM, with shipments already underway. Amazon AI chief Peter DeSantis is overseeing an environment where Amazon is also building its own chips, including Trainium and Arm-built Graviton CPUs, to lessen its dependence on NVIDIA.
Amazon’s AI chip business has crossed a $25 billion annualized revenue run rate, driven by $225 billion in total commitments from AI labs. That business gives Amazon another path for meeting AI demand while its AWS infrastructure continues to add NVIDIA hardware.
Revenue Growth Meets a Huge Supply Commitment
NVIDIA reported $96.2 billion in sales for the second quarter, including $89 billion in data center revenue. Data center revenue rose 117% from a year earlier, and NVIDIA expects total revenue to reach $108 billion in the third quarter, with some revenue coming from its next-generation Rubin GPUs.
To support future data center projects, NVIDIA has committed $279 billion to secure supply and manufacturing capacity. The company projects $92 billion in spending for the rest of the fiscal year and another $87 billion for fiscal year 2028.
Jensen Huang described the economic logic behind that investment in direct terms: “AI is generating profitable tokens … If we had more compute, we could generate more profitable tokens, which results in more profit for all of the services.”
That same push is visible in work beyond NVIDIA’s own hardware. OpenAI said, “We designed Jalapeño to minimize data movement and communication delays,” adding another example of the industry’s focus on moving information through AI systems with fewer delays.
The next stage of this expansion will also reach the startup community. An event is scheduled for October 13–15, 2026, in San Francisco. Across cloud systems, local computers, vehicles, and robots, NVIDIA’s partnerships show how the company is positioning AI as a connected computing layer rather than a single type of chip.
Based on
- NVIDIA and MediaTek Expand Partnership to Build AI Edge-to-Cloud Platforms — unite.ai
- NVIDIA Groq 3 LPX Now in Full Production With World-Class Speed for Agentic AI | Markets Insider — markets.businessinsider.com
- Nvidia’s AI advantage is moving beyond the GPU | TechCrunch — techcrunch.com
- Amazon just tripled its order of Nvidia chips over ‘surging demand’ | TechCrunch — techcrunch.com
