Markets · AI Chips & Compute · Companies
Cramer Says AI Data-Center Rebound Is Not Yet a Buy Signal
Cramer: AI data-center stocks still risky despite rebound; charts, earnings, election and Anthropic IPO are hurdles.
Latest news, developments and reporting on AI Chips & Compute.
Cramer: AI data-center stocks still risky despite rebound; charts, earnings, election and Anthropic IPO are hurdles.
AWS released TorchNeuron, letting PyTorch AI apps run on Trainium with a one-line change, sharpening its challenge to Nvidia.
Tenstorrent raised $1.5B at a $6.2B valuation to scale chips for open-weight AI and video generation.
Nvidia launched open-weight Nemotron 3.5 Super VL for enterprise agents and Bolt models for local AI PCs.
Google says a 780 MW Minnesota AI data center can finance more clean power than it uses; critics question grid costs.
Tenstorrent raised $1.5B at a $6.2B valuation to expand its open RISC-V AI compute platform and challenge Nvidia.
Chinese AI talent pay reaches 5 million yuan amid salary inversion; firms warned against top models for every task.
China's GDP grew 4.5% in Q3, beating forecasts as AI-related exports lifted output, The Guardian reported.
Six hyperscalers are set to spend over $700 billion on computing this year, double 2025's record, as chip scaling laws fade.
Turing Quantum announced photonic-chip supply-chain cooperation, an expanded computing alliance and a Greater Bay Area hybrid quantum-classical center, as it pushes photonic quantum computing toward commercial deployment.
MemSilicon has raised more than RMB400 million in a Series A round and appointed former Graphcore president Lu Tao as CEO and co-founder, as the startup expands from RRAM storage into LPU inference chips.
Seoul's KOSPI closed up 1.76% as institutions bought shares, while the won hit a two-year high and investors weighed returns on heavy AI spending.
DeepSeek released V4.1-Flash-1017, Nvidia improved Kimi K3 inference by 1.68 times, and Chinese chip and embodied-AI startups announced funding, while JD and OpenAI outlined robotics plans.
CXMT shares fell 6.5% despite a forecast swing to profit for the first nine months of 2026.
LS Electric wins a 170 billion-won U.S. switchgear deal for a hyperscale AI data center, with deliveries due next year.
Microsoft has ended a U.K. AI data center lease with Nscale, hurting the Nvidia-backed neocloud’s planned U.S. IPO.
Arm’s AGI CPU nearly doubles AMD EPYC’s modelled agentic AI rack performance; production tests remain.
Nscale's £2bn Essex data centre is delayed to the 2030s by grid constraints; Microsoft exits as anchor tenant.
A CNBC columnist argued that Friday’s sharp selloff in AI and semiconductor shares may reflect forced deleveraging rather than the end of the data-center buildout, citing CoreWeave’s bond yields and Nvidia’s financing role.
Belgian startup Corbenic is proposing that servers use their existing SSDs as memory for AI workloads, with TechRadar reporting a claimed 14x speedup that it says owes more to reading less data than to raw speed.
Intel introduced the GPU D-series at the Open Compute Project Global Summit, an AI inference accelerator with up to 480GB of LPDDR5X memory and support for 300-billion-parameter models on a single card.
Rivian CEO RJ Scaringe said the automaker is developing AI models, an autonomy stack and custom silicon, and is clearing a path to license its technology to other brands.
2026 WESEMiBAY drew 156,700 visitors and 1,000+ exhibitors in Shenzhen; organizers plan a 100,000 sqm 2027 edition.
South Korea's finance minister says chip-led growth will strengthen as the IMF lifts its 2026 forecast to 3.3 percent.