OpenAI has unveiled its first custom AI chip, Jalapeno, developed with Broadcom to address computing power shortages and reduce reliance on Nvidia's GPUs. According to Reuters, the chip is designed for AI inference, performing data processing to answer user queries to chatbots like ChatGPT. Broadcom CEO Hock Tan confirmed that the Jalapeno chip is comparable to Nvidia's Blackwell chips and Google's tensor processing units, stating it will be performant on future iterations of large language models (LLMs). The chip took approximately nine months to design before manufacturing at Taiwan's TSMC, with OpenAI reporting samples are running in labs with their GPT-5.3-Codex-Spark AI model operating at target power and performance levels.
Nvidia has announced a strategic partnership with Amazon Web Services to accelerate AI deployment at production scale across multiple AWS services. The collaboration addresses critical infrastructure constraints including low-latency inference, fast vector search, strong GPU price-performance and infrastructure that can grow without multiplying operational complexity. EC2 G7 instances powered by NVIDIA RTX PRO 4500 Blackwell Server Edition GPUs expand compute capabilities for AI, graphics, video and data analytics workloads, while NVIDIA cuVS library accelerates vector indexing by making GPU-powered vector indexing the default in OpenSearch Serverless. AWS has achieved NVIDIA Exemplar Cloud status for NVIDIA GB300, ensuring customers receive peak optimized performance for training workloads through deep co-engineering efforts between AWS and NVIDIA teams.
Nvidia has achieved a 21.5% market share in the data centre Ethernet switching market during the first quarter of 2026, according to IDC data. The company generated $2.1 billion in data centre Ethernet switch revenue, representing a 193% increase from the previous year. This dramatic surge positions Nvidia as the leading vendor in a market that grew 39.8% year-over-year to $15.4 billion, with data centre switching accounting for $10 billion of that total, up 61%. The broader Ethernet switch market reflects the accelerating buildout of AI factories globally.
Nvidia CEO Jensen Huang announced that the artificial intelligence industry has entered a new phase with the arrival of agentic AI, which he described as AI systems capable of autonomously carrying out tasks and making decisions. According to reports from NDTV Profit, Huang stated that 'Agentic AI has now arrived. Every industry is racing to adopt agentic AI' during Nvidia's annual stockholder meeting. The CEO emphasized that 'Every enterprise is becoming an agent company' as businesses deploy intelligent agents across various business functions.
The Nvidia-AWS collaboration delivers significant performance improvements across multiple AI workloads. G7 instances deliver up to 4.6x AI inference performance and up to 2.1x graphics performance compared to G6 instances, while providing up to 7.6TB of local NVMe SSD storage and 700 Gbps of EFA-enabled networking. The next generation of Amazon OpenSearch Serverless achieves vector indexing up to 10x faster at a quarter of the cost compared to CPU-only builds, making billion-scale vector databases practical to build in under an hour. Support spans AWS Deep Learning Amazon Machine Images, Amazon Deep Learning Containers, Amazon EMR, Amazon EKS, Amazon ECS and graphics AMIs, with additional support coming soon to Amazon SageMaker AI.
Huang reported strong financial results for the company, stating that revenue increased 65% over the past year while operating income rose 60%. According to IDC, Nvidia's record data centre networking revenue reached $14.8 billion in Q1 fiscal 2027, up 199% from a year earlier. The company's overall data centre revenue reached a record $75.2 billion, up 92%. These figures reflect continued demand for AI infrastructure and computing platforms, with the CEO noting that 'The buildout of AI factories – the largest infrastructure expansion in human history – is accelerating at extraordinary speed'.
Despite the optimistic outlook presented by the CEO, Nvidia shares edged lower, falling 0.35% to $199.33 following the annual shareholder meeting. According to NDTV Profit, this market reaction suggests investor caution despite the positive financial performance and growth prospects outlined by Huang. The competitive landscape has intensified as Arista held a 20.7% share and Cisco accounted for 17.8% of the data centre Ethernet switching market during the quarter, according to IDC. The shift toward integrated AI platforms means vendors must compete not just on switch speeds or pricing, but through broader strategies for compute, networking, and software integration.
