Skip to main content

Greener AI
starts here

Every AI request uses energy, water and generates CO2. LexiCo's STONE compression reduces token usage by up to 79% — and with it, the environmental footprint. Small changes, big impact.

See the savings Get started

Fewer tokens = greener planet

Every token sent to an AI model requires compute power — GPU cycles, electricity and cooling. LexiCo's STONE compression eliminates up to 79% of redundant tokens before they reach the AI model. The result is dramatically lower energy consumption per request.

STONE (Semantic Token Optimization & Natural Encoding) semantically analyzes content and removes redundancy and noise - without losing meaning. Less data in means less energy used, less heat generated, and less cooling water consumed.

79% fewer tokensⓘ Each request is intelligently compressed. 79% fewer tokens means 79% less compute power — directly translated to lower electricity consumption.
Less data transfer Compressed requests mean less network traffic. Less data through the wires reduces energy consumption across the entire chain — from client to data center.
Faster processing Shorter requests are processed faster. The GPU is busy for less time per request, resulting in both lower energy use and higher throughput.

0ⓘ

Energy savings

0

Less water usage

0

CO2 reduced/day

0

Fewer GPU hours

Calculate your environmental savings

Drag the sliders and see how much the environment saves when you use LexiCo.

Daily AI requests 10 000
1001 000 000
Tokens per request 4 000
50050 000
0
Trees saved
0
Liters of water saved
0
kWh energy saved
0
kg CO₂ reduced
16 / 16
Congratulations! With LexiCo you save the equivalent of
0
trees per year

Calculations are based on published data from IEA and research papers on AI energy consumption.

Water & cooling

Data centers use enormous amounts of water for cooling. A single large data center can consume over 12 million liters of water annually — equivalent to a small town's water supply. Every working GPU generates heat that must be cooled.

When STONE reduces GPU load by 79%, heat production is reduced accordingly. Less heat means less cooling needed, which means thousands of liters of water saved — per customer, per month.

Reduced cooling needsFewer GPU cycles generate less heat. Data centers need less cooling capacity when the workload is lower.
Thousands of liters savedFor a medium-sized business, a 79% reduction in GPU usage means hundreds of liters of cooling water saved monthly.
Preserves local water sourcesIn water-scarce regions, this is critical. Responsible AI use starts with minimizing resource consumption.

CO2 & carbon footprint

Lower energy consumption means direct reduction in greenhouse gas emissions.

A single ChatGPT request uses roughly 10 times as much energy as a Google search. For businesses with thousands of daily AI calls, the carbon footprint becomes significant. STONE compression cuts this dramatically.

By eliminating 79% of tokens before they reach the GPU, LexiCo reduces the actual computation performed. Fewer calculations = lower electricity consumption = less CO2. For a typical enterprise customer, this can mean several tons of CO2 saved annually.

CPU & memory — efficient resource use

Every request to an AI model requires CPU cycles for tokenization, memory allocation for the context window and GPU time for inference. STONE compression reduces the number of tokens per request, directly cutting all three costs. When millions of requests are processed daily, this adds up to enormous savings in energy and server capacity.

LexiCo's O(1) Proxy uses constant memory and CPU regardless of request size. Where traditional proxies scale linearly with payload size (O(n)), our proxy keeps memory pressure flat. Combined with STONE compression, this delivers dramatically lower resource usage per request.

Up to 79% fewer tokensSTONE compresses away redundant tokens before they reach the GPU. A request that normally uses 10,000 tokens can be reduced to 2,100 - with identical response quality. Shorter input means shorter processing time, lower energy consumption and faster response for the end user.
O(1) memory usageTraditional proxies allocate memory proportionally to payload size - a 100K-token request uses 1000x more RAM than a 100-token request. LexiCo's O(1) Proxy keeps memory consumption constant regardless of size. The result is predictable performance with no risk of memory leaks or OOM crashes under peak load.
Constant resource usageThe O(1) architecture means the proxy handles a 100-token request and a 100,000-token request with exactly the same resource usage. This eliminates the need for autoscaling based on payload size, making capacity planning simple: requests per second is the only variable that matters.

Fewer data centers

Efficient use of existing capacity reduces the need for new, energy-intensive data centers.

The AI industry is building new data centers at record speed. Each new center requires enormous amounts of energy, water, steel and concrete. By making existing infrastructure more efficient, LexiCo reduces the pressure to build new ones.

When each request uses 79% less resources, one data center can do the job that previously required five. That's not just cost-effective — it's climate-smart.

Live climate tracking

LexiCo gives you a personal dashboard showing your environmental savings in real time. See exactly how much energy, water and CO2 your business saves — updated for every single AI request.

Use the data in your ESG reporting, sustainability report, or marketing. Show customers and investors that your business takes environmental responsibility seriously.

Real-time dataSee your energy, water and CO2 savings updated live in the dashboard for every request.
ESG reportingExport environmental data to your sustainability report. Document AI's climate impact with real numbers.
Climate goalsSet environmental savings goals and track progress over time. Show that your AI strategy is sustainable.

Start your green AI journey

Sign up for free and see your environmental savings in real time. Every request through LexiCo is a step toward a greener future.

Get started for free Contact us

Free registration — no credit card required