Articles

NVIDIA says Vera is now in full production and can complete agentic workloads 1.8x faster than x86 CPUs. OpenAI, Anthropic, SpaceXAI, ByteDance, CoreWeave, and OCI are among the names tied to adoption or evaluation.

The thread focused on a concrete supply-chain link: HBM demand for AI racks can squeeze DDR and LPDDR supply for everyday devices.

Anthropic has agreed to spend $200 billion on Google Cloud services and Broadcom-built TPUs over five years, beginning 2027. The deal accounts for over 40% of Google's disclosed cloud revenue backlog; Alphabet also plans up to $40B in additional investment in Anthropic.

Meta reported Q1 2026 revenue of $56.3 billion (+33% YoY), with net income rising 61%. But investors pushed the stock down 6–10% after Meta raised its full-year AI capex forecast to $125–145 billion, citing higher component costs and new data center buildouts.

Alphabet reported Q1 2026 revenue of $109.9 billion, up 22% year over year. Google Cloud surged 63% to $20.03 billion — its first $20B quarter — while net income jumped 81% to $62.6 billion. The stock rose 6% after the report.

Meta's new spending range says the hyperscaler arms race is getting more expensive, not calmer. Reuters reports the company raised 2026 capital expenditure guidance to $125 billion-$145 billion, and the stock fell more than 6% after hours.

AI infrastructure is moving upstream into energy. On April 27, 2026, Meta said it reserved up to 1 GW of space solar with Overview Energy and up to 1 GW/100 GWh of long-duration storage with Noon Energy for its data center buildout.

South Korea is no longer treating AI infrastructure as a private-sector side quest. A ₩400 billion loan from the Financial Services Commission will expand Naver’s Gak Sejong facility, bankroll GPU deployment and give HyperCLOVA X a bigger domestic base just as AI sovereignty becomes industrial policy.

Alphabet just rewired the AI capital race: $10 billion goes to Anthropic now at a $350 billion valuation, with another $30 billion tied to performance targets. Coming days after Amazon’s own pledge, the deal shows that frontier labs are no longer raising money in rounds so much as pre-buying compute at planetary scale.

This is less about one more cloud partnership and more about the infrastructure shape of the next agent wave. NVIDIA and Google Cloud say A5X Rubin systems can scale to 80,000 GPUs per site and 960,000 across multisite clusters, while cutting inference cost per token and boosting token throughput per megawatt by up to 10x versus the prior generation.

HN treated TPU 8t and 8i as more than giant datacenter numbers. The thread focused on the bigger shift: agent-era infrastructure is splitting training and inference into separate hardware bets.

HN latched onto the RAM shortage because the uncomfortable link is physical: HBM demand for AI data centers is now shaping prices for phones, laptops, and handhelds.