NVIDIA Blackwell inference stack cuts DeepSeek V4 token costs by 80%.

NVIDIA's Blackwell inference software delivers 5x throughput improvement for DeepSeek V4 through co-designed hardware-software optimizations, reducing token costs to roughly one-fifth of previous levels and enabling up to 20x higher throughput on the same GPU.

Sources

X mentions

5.9k ▲

First seen

2Dago

Velocity

+5%/6h

CONTRIBUTING SOURCES

2 ARTICLES

X DISCOURSE

5.9k TOTAL · TOP 3

@nvidia1H · 28.5K

RT @NVIDIAAI: Most motion papers tailor one controller to one specific task. This year at SIGGRAPH, our research team asks: can motor contr…

@nvidia1D · 25.4K

We're working with @PalantirTech to bring NVIDIA Nemotron open models into secure, air-gapped environments for U.S. government agencies and critical infrastructure operators. The new engine lets teams train on their own data, run models on their own infrastructure and retain full model ownership.

@nvidia2D · 404.6K

Water usage has been a hot topic in the AI data center world, but the numbers may surprise you. According to the Manhattan Institute, data centers use 0.2 percent of daily water usage in the U.S. and that number has dramatically decreased in the past few years due to a new method: liquid cooling. By moving to 45°C liquid cooling, AI factories in favorable climates can use dry coolers instead of conventional cooling-tower-based systems, cutting facility cooling water use from roughly 2.6M gallons per MW per year to near zero. Liquid cooling enables AI factories to be both water and energy efficient, while creating opportunities for heat reuse and dispersal to local communities, allowing these factories to become energy grid assets. Learn more below ⬇️ https://t.co/7WanoPNKTR