LIVE · WED, JUN 24, 2026 --:--:-- ET
Issue Nº 64 COST TOTAL $14510.26 ARTICLES TODAY 14 TOKENS TOTAL 9.10B
Podcast
aiexpert
Subscribe · $8/mo
Running the wire
Market Micron posts $41.5B Q3 revenue, guides $50B for Q4 on AI memory supercycle Funding Qualcomm acquires Modular for ~$4B to build hardware-agnostic AI stack against NVIDIA CUDA Market AWS launches EC2 G7 instances with NVIDIA RTX PRO 4500 Blackwell; 4.6x inference gains Chips Qualcomm unveils Dragonfly C1000 data-center CPU; Meta commits to 2028 production volumes Chips OpenAI unveils Jalapeño inference chip with Broadcom, targets late-2026 deployment Breaking Huang tells shareholders black-market data centers from smuggled chips are a "dead end" Research Google integrates computer use natively into Gemini 3.5 Flash for agentic automation Research Google OpenRL: Self-hosted Kubernetes API for LLM post-training; decouples RL from infrastructure Market Micron Q3 earnings beat on record DRAM margins; HBM supply fully allocated through 2026 Policy US secures Netherlands for Pax Silica chip alliance; ASML tensions persist over MATCH Act export restrictions Chips OpenAI & Broadcom unveil Jalapeño: Custom LLM inference chip targets gigawatt-scale deployment by end of 2026 Breaking Gemini 3.5 Flash adds native computer use; agent framework now default across Search Research AI rapidly designs novel radio-frequency chips beyond human intuition, reducing years of work to hours Chips China's LineShine supercomputer tops TOP500 with 2.198 exaflops CPU-only, ending US El Capitan's reign Market Cerebras stock plummets 17% after margin-guidance miss as CEO says warning was 'misunderstood' Market Sunrun, Tesla, Renew Home form 16GW virtual power plant for AI data centers; RUN +31% Breaking Amazon Zoox unveils redesigned robotaxi, planning paid service launch in late 2026 Funding XCures closes $46M Series B round at $127M post-money valuation Funding Qualcomm acquires Modular for ~$4B to bolster AI software stack and data center play Chips OpenAI & Broadcom unveil Jalapeño, custom LLM inference chip with 9-month design cycle Market Micron posts $41.5B Q3 revenue, guides $50B for Q4 on AI memory supercycle Funding Qualcomm acquires Modular for ~$4B to build hardware-agnostic AI stack against NVIDIA CUDA Market AWS launches EC2 G7 instances with NVIDIA RTX PRO 4500 Blackwell; 4.6x inference gains Chips Qualcomm unveils Dragonfly C1000 data-center CPU; Meta commits to 2028 production volumes Chips OpenAI unveils Jalapeño inference chip with Broadcom, targets late-2026 deployment Breaking Huang tells shareholders black-market data centers from smuggled chips are a "dead end" Research Google integrates computer use natively into Gemini 3.5 Flash for agentic automation Research Google OpenRL: Self-hosted Kubernetes API for LLM post-training; decouples RL from infrastructure Market Micron Q3 earnings beat on record DRAM margins; HBM supply fully allocated through 2026 Policy US secures Netherlands for Pax Silica chip alliance; ASML tensions persist over MATCH Act export restrictions Chips OpenAI & Broadcom unveil Jalapeño: Custom LLM inference chip targets gigawatt-scale deployment by end of 2026 Breaking Gemini 3.5 Flash adds native computer use; agent framework now default across Search Research AI rapidly designs novel radio-frequency chips beyond human intuition, reducing years of work to hours Chips China's LineShine supercomputer tops TOP500 with 2.198 exaflops CPU-only, ending US El Capitan's reign Market Cerebras stock plummets 17% after margin-guidance miss as CEO says warning was 'misunderstood' Market Sunrun, Tesla, Renew Home form 16GW virtual power plant for AI data centers; RUN +31% Breaking Amazon Zoox unveils redesigned robotaxi, planning paid service launch in late 2026 Funding XCures closes $46M Series B round at $127M post-money valuation Funding Qualcomm acquires Modular for ~$4B to bolster AI software stack and data center play Chips OpenAI & Broadcom unveil Jalapeño, custom LLM inference chip with 9-month design cycle
Market

AWS launches EC2 G7 instances with NVIDIA RTX PRO 4500 Blackwell; 4.6x inference gains

Amazon Web Services announced general availability of EC2 G7 instances powered by NVIDIA RTX PRO 4500 Blackwell Server Edition GPUs on June 18, 2026. G7 is the first public cloud instance type to feature the Blackwell-generation server GPU, delivering up to 4.6x AI inference performance and up to 2.1x graphics performance compared to prior-generation G6 instances. Instances support up to 8 GPUs per node with 32 GB of memory per GPU, totaling 256 GB of GPU memory, paired with custom Intel Xeon Scalable 6th-generation processors, up to 700 Gbps of EFA-enabled networking (7x vs. G6), and up to 7.6 TB of NVMe SSD storage.

G7 comes in 7 sizes supporting up to 192 vCPUs and is optimized for AI inference workloads (language translation, video/image analysis, speech recognition, recommendation systems), professional graphics rendering, VDI, and GPU-accelerated analytics on Amazon EMR. AWS achieved NVIDIA Exemplar Cloud status on NVIDIA GB300 training workloads, confirming AWS infrastructure meets NVIDIA's reference performance thresholds. G7 instances are available in US East (Ohio) and US West (Oregon) with plans for regional expansion, and can be purchased via On-Demand, Savings Plans, and Spot options.

The launch reflects hyperscaler demand for scaled GPU capacity: G7 provides faster vector indexing (up to 10x faster at 1/4 cost versus CPU-only OpenSearch via NVIDIA cuVS), lower-latency multi-GPU communication via GPUDirect P2P and RDMA, and the networking throughput needed for distributed inference. The combination of Blackwell compute, high-bandwidth memory (2.45x vs. G6), and optimized interconnect targets production-scale AI deployment where latency, throughput, and per-inference cost drive architecture decisions.

For cloud architects deploying inference at scale, G7 validates Nvidia's Blackwell timeline in customer hands and signals a tightening race on cost-per-inference: OpenAI/Broadcom's Jalapeño and Qualcomm's Dragonfly target similar efficiency gains, but G7's immediate availability, AWS scale, and Blackwell's maturity offer hyperscalers a trusted baseline. Monitor G7 adoption curves and vector-search performance gains as indicators of whether general-purpose cloud GPUs remain cost-competitive with custom ASICs for high-volume inference.

Sources

Get the signal before the noise.