#10 of 30 · AI lab + cloud
Tencent (Hunyuan)
Half its electricity now green, PUE capped at 1.25, GPUs rationed by export rules.
In plain words
How scores workTencent scores 58 out of 100 — middle of the pack. Some disclosure or clean-energy commitments, offset by rising emissions or blind spots. Confidence in this score is good confidence (B).
What helps
- Efficiency · Average PUE 1.246 under a 1.25 cap; modular T-Block design reaches 1.06.
- Carbon · Renewable electricity 48.5% in 2025, up from 22% in 2024, and 82.9% at owned data centres; 6.5 billion kWh of green power procured; 80 MW of on-site generation built or under way.
- Commitments · 70% absolute Scope 1+2 cut and 30% Scope 3 cut by 2030 against 2021; 100% green power by 2030.
What hurts
- Efficiency · No per-prompt figures; Hunyuan is a 389B MoE with 52B active parameters.
- Carbon · Remaining load on China's coal-heavy grid; the last full inventory in English (2022) was 5.74 MtCO₂e.
- Water · No WUE or consumption figures published.
What nobody knows
- Per-prompt energy disclosed
- Per-prompt water disclosed
- Training emissions in model cards
- Water-use effectiveness reported
- Hardware reuse or e-waste reported
Tencent's Hunyuan models serve WeChat and Yuanbao, and its data centres run at an average PUE of 1.246 against a self-imposed 1.25 cap. Renewable electricity jumped from 22% in 2024 to 48.5% in 2025 (82.9% at owned data centres), backed by 6.5 billion kWh of green power bought since 2021. Capex fell 24% in 2025 because it could not buy the GPUs it wanted. Targets: a 70% absolute Scope 1+2 cut and 100% green power by 2030.
Key figures
Per-prompt and fleet metrics. Hover a title for what it means; open “How we got this number” for the working.
ecogpt.co estimate; no company figure.
How we got this number
Hunyuan's flagship is a 389B mixture-of-experts model with 52B active parameters; a little below a dense frontier model's 0.3 Wh. Uncertainty roughly ±3×.
On-site cooling only, so comparable with Google and OpenAI's figures.
How we got this number
0.25 Wh per prompt × 0.8 L/kWh (estimated WUE for T-Block sites with PUE 1.25). Counting the water evaporated to generate the electricity adds ≈0.45 mL.
Location-based: what the grid actually emitted. Renewable certificates would bring the market-based figure near zero.
How we got this number
0.25 Wh × 580 gCO₂/kWh (Chinese grid, location-based). Owned data centres are 82.9% renewable, so Tencent's market-based figure would be far lower.
Latest full inventory published in English.
Company characterisation; no percentage printed.
How we got this number
No disclosure. T-Block sites use indirect evaporative cooling with a PUE around 1.25; 0.8 L/kWh assumed.
Hardware & materials
The chips, power, copper and future e-waste behind the fleet — the part sustainability reports leave out
GPUs, TPUs or custom accelerators the company runs. Analyst figures are in H100-equivalents, company figures in physical chips.
How we got this number
Tencent publishes no fleet size. Capex reached $10.6bn in 2024, then fell 24% in 2025 because export rules limited GPU purchases; the company says its reserves are 'sufficient' for training. Uncertainty ±2×.
Megawatts the AI fleet pulls. One megawatt runs roughly 750 US homes.
How we got this number
150,000 accelerators × 1.3 kW each, all-in. Fleet power implied by the chip count.
Terawatt-hours per year. One TWh is a year of electricity for about 95,000 US homes.
How we got this number
195 MW × 8,760 h × 80% utilisation. Compute fleet only.
CO₂e emitted making the chips and servers before they answer a single prompt.
How we got this number
150,000 accelerators × 514 kg CO₂e each (chip plus share of the host server). Emitted in manufacturing, before the first prompt is served.
Tonnes of server hardware that will be retired in three to five years.
How we got this number
150,000 accelerators × 17 kg of server hardware each. Tomorrow's e-waste unless reused; generative AI could produce 2.5 Mt of it a year by 2030.
Tonnes of copper in cabling, busbars and transformers.
How we got this number
195 MW × 27 t of copper per MW (cabling, busbars, transformers).
Diesel backup units or gas turbines at the data centres, permitted or observed.
How we got this number
195 MW ÷ 2.6 MW per generator, the average size permitted at US data centres. Permit counts are not published fleet-wide; this is the number a fleet this size normally carries.
Share of retired servers and parts given a second life.
Kilograms of antimony-equivalent, the life-cycle unit for scarce minerals consumed.
Score breakdown
Each pillar is 0–100; weights are fixed across the index
- No WUE or consumption figures published. Tencent
- 70% absolute Scope 1+2 cut and 30% Scope 3 cut by 2030 against 2021; 100% green power by 2030. Tencent
- Slowed GPU deployment after DeepSeek's efficiency gains rather than overbuilding, and says it has enough chips for training. Data Center Dynamics
- No enforcement actions on record. Tencent
Models
What this company ships, and what each prompt costs
| Model | Weights | Energy / prompt | Water | Training CO₂e |
|---|---|---|---|---|
| Hunyuan T1 / Turbo S 389B total / 52B active (MoE) | Closed | 0.25 Wh Estimateecogpt.co estimate. Our own estimate from published factors. The working is shown next to the number; treat it as an order of magnitude, not a measurement. | 0.200 mL Estimateecogpt.co estimate. Our own estimate from published factors. The working is shown next to the number; treat it as an order of magnitude, not a measurement. | 4,000 t Estimateecogpt.co estimate. Our own estimate from published factors. The working is shown next to the number; treat it as an order of magnitude, not a measurement. |
Where the compute lives
Named sites from company disclosures and press reporting
- Shenzhen HQGuangdong, China
- TianjinTianjin, China · T-Block modular campus.
- QingyuanGuangdong, China
- ChongqingChongqing, China
Timeline
Incidents and developments involving this company