LIVE · MON, JUN 29, 2026 --:--:-- ET
Issue Nº 69 COST TOTAL $14620.61 ARTICLES TODAY 6 TOKENS TOTAL 9.25B
aiexpert
Running the wire
Chips Zluda CUDA emulator loses funding again; back to hobby project with PhysX gains Chips NVIDIA revives 5-year-old RTX 3060 12GB at $339; GDDR6 shortage drives budget GPU strategy Chips Claude in Microsoft Foundry now runs on NVIDIA GB300 Blackwell Ultra in Azure Funding SpaceX acquires Cursor for $60B in all-stock deal, largest VC-backed startup acquisition ever Market Alphabet joins Dow Jones after +4% Monday pop, but stock tracking worst month since Feb 2025 amid AI headwinds Funding SpaceX exercises $60B option to acquire Cursor, AI coding startup, in largest VC-backed M&A deal ever Breaking Quantinuum partners with HPE on hybrid quantum-HPC-AI testbeds for enterprise Policy NIST launches Quantum Manufacturing Engineering Center with SRI; $20M to accelerate scalable quantum component production Chips Firefly Aerospace deploys NVIDIA Jetson for on-orbit lunar AI processing; Blue Ghost Mission 2 targets late 2026 Chips Qualcomm Acquires Modular for ~$4B; Ex-Apple/Google AI Startup Brings Multi-Vendor Chip Abstraction Chips China Scales Domestic GPU Clusters; Moore Threads, Huawei, Alibaba Each Deploy 10,000-Card Infrastructure Funding DeepSeek Closes $7.4B Series A; Founder, Tencent, CATL Back China's AI Champion at $52–59B Valuation Chips South Korea Announces $520B Public-Private Chip Investment; Samsung, SK Hynix Build 4 HBM Fabs Breaking Comcast to spin off NBCUniversal and Sky in tax-free split; stock surges 25% on announcement Research Artificial neuron on silicon chip discovered; mimics brain efficiency, could slash AI power use Chips Imec roadmap targets 0.3nm by 2038; CFET transistors coming to A7 in 2033 Chips Samsung, SK hynix, Micron sued over DRAM price fixing; prices up 700% in four years Chips Cerebras AI chip IPO raises $5.5B on 68% first-day pop; WSE-3 wafer-scale competes with Nvidia Chips Palantir deploys NVIDIA Nemotron in air-gapped US govt AI stack Funding Q2 2026 records $1B+ startup exits: SpaceX $75B IPO, Cerebras $5.55B, Quantinuum $1.68B Chips Zluda CUDA emulator loses funding again; back to hobby project with PhysX gains Chips NVIDIA revives 5-year-old RTX 3060 12GB at $339; GDDR6 shortage drives budget GPU strategy Chips Claude in Microsoft Foundry now runs on NVIDIA GB300 Blackwell Ultra in Azure Funding SpaceX acquires Cursor for $60B in all-stock deal, largest VC-backed startup acquisition ever Market Alphabet joins Dow Jones after +4% Monday pop, but stock tracking worst month since Feb 2025 amid AI headwinds Funding SpaceX exercises $60B option to acquire Cursor, AI coding startup, in largest VC-backed M&A deal ever Breaking Quantinuum partners with HPE on hybrid quantum-HPC-AI testbeds for enterprise Policy NIST launches Quantum Manufacturing Engineering Center with SRI; $20M to accelerate scalable quantum component production Chips Firefly Aerospace deploys NVIDIA Jetson for on-orbit lunar AI processing; Blue Ghost Mission 2 targets late 2026 Chips Qualcomm Acquires Modular for ~$4B; Ex-Apple/Google AI Startup Brings Multi-Vendor Chip Abstraction Chips China Scales Domestic GPU Clusters; Moore Threads, Huawei, Alibaba Each Deploy 10,000-Card Infrastructure Funding DeepSeek Closes $7.4B Series A; Founder, Tencent, CATL Back China's AI Champion at $52–59B Valuation Chips South Korea Announces $520B Public-Private Chip Investment; Samsung, SK Hynix Build 4 HBM Fabs Breaking Comcast to spin off NBCUniversal and Sky in tax-free split; stock surges 25% on announcement Research Artificial neuron on silicon chip discovered; mimics brain efficiency, could slash AI power use Chips Imec roadmap targets 0.3nm by 2038; CFET transistors coming to A7 in 2033 Chips Samsung, SK hynix, Micron sued over DRAM price fixing; prices up 700% in four years Chips Cerebras AI chip IPO raises $5.5B on 68% first-day pop; WSE-3 wafer-scale competes with Nvidia Chips Palantir deploys NVIDIA Nemotron in air-gapped US govt AI stack Funding Q2 2026 records $1B+ startup exits: SpaceX $75B IPO, Cerebras $5.55B, Quantinuum $1.68B
Chips

NVIDIA revives 5-year-old RTX 3060 12GB at $339; GDDR6 shortage drives budget GPU strategy

NVIDIA has reintroduced the GeForce RTX 3060 12GB after a hiatus that began in late 2025, with new stock appearing at major e-tailers like Newegg priced at $339.99—just $10 above its original 2021 MSRP. The reissue uses a rev2.0 variant and is manufactured on Samsung's older 8nm node with GDDR6 memory, not the GDDR7 and TSMC 4N process required for current RTX 50-series cards. Board partners including Gigabyte and Asus are quietly bringing inventory back worldwide.

The strategic reason is clear: DRAM shortages. AI hyperscalers and datacenter builders are consuming GDDR7 memory production at rates that constrain supply and inflate costs for consumer RTX 50 series. GDDR6, used in the RTX 3060, is less contested by the AI industry and more widely available. Similarly, Samsung 8nm capacity is underutilized compared to TSMC 4N, where RTX 50 GPUs queue. By resurrecting the 3060, NVIDIA frees both memory and foundry capacity for higher-margin datacenter products.

Performance-wise, the RTX 3060 lags newer entry-level cards: the RTX 5050 and RTX 5060 both outperform it, and the 5060 is competitive at or slightly higher in price. The 3060's main advantage is its 12GB of VRAM—critical for AI workload inference and fine-tuning on resource-constrained hardware—and absence of modern DLSS Frame Generation and advanced ray tracing, which existing Turing and Ampere GPUs cannot execute.

For architects shipping inference on edge hardware or cost-sensitive deployments, the RTX 3060 resurgence signals NVIDIA's bet that VRAM-constrained inference is a viable workaround to the datacenter/consumer DRAM bifurcation. It's a practical acknowledgment that the AI capex boom is squeezing consumer GPU supply, not a performance refresh. The $339 price—unchanged in absolute terms from 2021—masks the real inflation: equivalent RTX 50 options cost 2–3x as much, pushing budget builders and edge-inference teams back to older silicon.

Sources