How could the co-design of efficient AI data centers affect Google's operations?
Optimized AI data centers could lower Google's operational costs and improve capacity The collaboration between Nvidia and Vertiv aims to address the significant inefficiencies in AI data centers, where only 70% of the power drawn from the grid is used for AI processing. By optimizing infrastructure and power distribution, these new designs could allow hyperscalers like Google to reduce the 30% of energy currently lost to cooling, power conversion, and system failures. This increased efficiency and optimized capacity could translate into lower operational costs and better utilization of resources for Google's massive AI workloads.
- Effect
- Mild positive
- How direct
- Brind's reasoning
- When
- Within weeks
- The story
- Gone quiet
How it reaches Google
-
Rod Evans, Vice President of Supercomputing, is co-designing next-generation AI data centers with Nvidia and Vertiv. Nvidia stated that only 70% of the power used in an AI data center is dedicated to AI processing. The remaining 30% is attributed to infrastructure and system issues, including cooling, power distribution, and restarts.
The full event1independent outlet -
Vertiv, as the infrastructure provider, is developing sustainable facilities that maximize the output from every watt of power. These optimized designs directly address the power and cooling challenges faced by hyperscalers, allowing Google to manage its large-scale AI workloads with greater efficiency and potentially lower energy costs.
No report states this step directly. Brind drew it from the reporting on each side of it, so treat it as an informed guess.
-
American multinational technology company, a subsidiary of Alphabet Inc.
Everything about Google
Tap any step to see the evidence behind it.
The facts so far
As reported. Each one links to where it comes from.
- Only 70% of power used in an AI data center is used for AI work, with 30% representing inefficiencies.techradar.com
- For a one-gigawatt data center, 20% is lost through cooling, power distribution, and conversion.techradar.com
- An additional 10% is lost to incorrectly configured racks, failures, and restarts.techradar.com
Why it matters
The efficiency of AI data centers is a critical factor in the long-term viability and cost structure of hyperscale cloud computing. As AI workloads continue to grow, the demand for power and cooling capacity increases exponentially. If the industry can successfully recover a significant portion of the 30% inefficiency, it could fundamentally change the economic model of running large AI infrastructure.
This development is part of a broader industry shift where the bottleneck for AI capacity is moving from chip shortages to the availability and efficiency of power and grid connections. Companies like Vertiv and Nvidia are leading the charge to solve this infrastructure challenge, which is vital for all major cloud providers, including Google.
What we don't know yet
- How quickly will these optimized designs be adopted by major hyperscalers?
- What specific cost reductions could Google realize by implementing these new infrastructure standards?
What would change this answer
— Other considerations
Beyond this explainer
Explore connected ideas
Each opens in Related concepts: what the web connects to it, with Brind's news on it.
The ideas come from how millions of web pages place them near each other. The questions are Brind's AI reasoning, written only from this explainer and Brind's own reporting.
Who else could feel it
Other possible ripples from the same event.
- VistraLikelyVistra may see increased demand for reliable electricity supply from AI data centers
- Schneider ElectricPossibleNvidia and Vertiv's focus on efficiency may challenge Schneider Electric
- Navitas SemiconductorPossibleNew AI data center designs may boost demand for efficient power semiconductors
- GE VernovaPossibleDemand for energy equipment may rise as AI data centers prioritize efficiency
Reporting
- techradar.comOct 14
Keep going
Brind's analysis is written by AI from the reporting linked above and can be wrong. This explainer is a possibility worth considering, not a forecast: steps marked as Brind's reasoning are not stated in the reporting. It is not investment advice; do your own research.