# Basiscooling

*/Startups/Basiscooling*

## Startup Overview

The platform connects directly to server hardware to modulate liquid cooling flow based on live silicon thermal telemetry. Instead of cooling entire data center halls based on generalized ambient temperatures, the system targets thermal loads precisely at the chip level. It ingests real-time temperature data from GPUs and CPUs to dynamically adjust pump speeds and valve positions across individual racks.

High-density compute operators and data center managers face significant energy waste when running static, room-level cooling configurations. As rack power densities climb to support intensive processing workloads, traditional systems overcool inactive racks while struggling to prevent localized thermal throttling on active nodes. The software eliminates the guesswork of aggregate room sensors by treating the processors themselves as the primary thermal triggers.

Compared to legacy building management systems and platforms like Vertiv Environet, the architecture shifts thermal regulation from a reactive facility function to a fully autonomous, silicon-precise operation. By responding directly to on-chip telemetry rather than ambient room sensors, it directs cooling capacity exactly where the hardware demands it. This direct hardware integration sustains maximum compute throughput and prevents the energy overhead associated with blanket facility cooling.

## Startup Founding Hypothesis

**Approach**: that modulates liquid cooling flow using live silicon thermal telemetry
**Competitors**:
- [Legacy BMS](/Competitors/Legacy_BMS)
- [Vertiv Environet](/Competitors/Vertiv_Environet)
- [Static room-level cooling](/Competitors/Static_room-level_cooling)
**Differentiator2x2**: silicon-level precise and fully autonomous rather than relying on room ambient sensors

## Startup Solution Coordinate

**Solution**: [Silicon Flow Director](/Software/Silicon_Flow_Director)

## Startup Position2x2

```mermaid
quadrantChart
    x-axis Room Ambient Sensors --> Silicon-Level Precise
    y-axis Manual/Static Adjustment --> Fully Autonomous
    Static room-level cooling: [0.15, 0.15]
    Legacy BMS: [0.25, 0.50]
    Vertiv Environet: [0.45, 0.70]
    Basiscooling: [0.85, 0.85]
```

## Startup Offer

**Proof**:
- Targeting 15-25% reduction in pumping power for high-density AI training clusters.
- Aiming to eliminate over-cooling deadbands in liquid-to-liquid heat exchangers.
- Designed to maintain GPU junction temperatures within 2 degrees Celsius of target under peak training loads.
**Tiers**:
- Name: Rack Pilot · Price: ~$300–$600/rack/mo · Inclusions: Real-time telemetry ingestion and cooling distribution unit (CDU) flow modulation for up to 5 high-density liquid-cooled racks.
- Name: Pod Scale · Price: ~$15–$30 per kW/mo · Inclusions: Row-level autonomous flow control, predictive load ramping, and intended integration with existing variable-speed pumps for up to 1 MW of IT load.
- Name: Facility Enterprise · Price: ~$50k–$120k/yr base + usage · Inclusions: Facility-wide deployment with intended on-premise air-gapped hosting and custom BMS connector mapping.
**Guarantee**: If the modulation software fails to reduce cooling energy usage by at least 10% without inducing thermal throttling during a 30-day pilot, the software fees are fully refunded.
**Business Function**: ProvideService
**Objection Handlers**:
- Will this conflict with our existing BMS? -> Designed to operate as a supervisory control layer above existing BMS, issuing flow setpoints rather than replacing underlying hardware safety loops.
- What happens if the telemetry feed drops? -> Automatically defaults to static maximum cooling flow to protect hardware until the telemetry connection is restored.
- Do we need proprietary sensors? -> Intended to pull existing internal temperature telemetry directly from the server baseboard management controller (BMC) via IPMI or Redfish.
**Pricing Architecture**: UsageMeter

## Startup Brand

**Voice**: Clinical and exact, anchored entirely in thermal engineering realities.
**Tagline**: Autonomous liquid cooling synchronized to live silicon temperatures.
**Icon Concept**: valve
**Palette Intent**: institutional-cool
**Visual Identity**: A palette of chilled silver and deep coolant blue pairs with technical monospaced typography to reflect exact thermal tolerances.
**Archetype Reference**: the-ruler

## Startup Buyer Chain

**Chain**: Basiscooling → Data Center Operator → High-Density Compute Tenant
**Gtm Motion**: Acquires initial customers through single-rack proofs of concept targeting thermal engineers tasked with deploying high-density AI clusters. Expands by scaling the telemetry-based modulation across entire data center halls and standardizing the software as a facility-wide cooling primitive.
**Agent Channel**: Intended to list within DMTF Redfish API catalogs and OpenBMC capability registries, enabling autonomous workload balancers and Kubernetes schedulers to discover cooling control endpoints and modulate flow ahead of heavy silicon utilization.
**Primary Channel**: Technical presentations at the Open Compute Project (OCP) Summit and targeted outreach to directors of data center infrastructure management (DCIM) researching PUE optimization for liquid-cooled racks.

## Startup Customer Journey

```mermaid
flowchart LR A[OCP Summit] --> B[Rack Pilot] --> C[BMC Telemetry] --> D[Cooling Distribution Unit] --> E[Pod Scale System] --> F[Facility Enterprise] --> G[Redfish API Catalog]
```

## Startup Proof Points

_Illustrative — target and order-of-magnitude estimate figures, not an achieved track record (this Thing is concept-stage)._

**Pilot Goals**:
- A 30-day Rack Pilot on up to 5 high-density racks aiming to achieve the guaranteed 10% reduction in cooling energy usage with zero thermal throttling, validating the core modulation algorithm.
- A 60-day Pod Scale deployment across a variable AI training cluster targeting the elimination of over-cooling deadbands and proving seamless integration with existing variable-speed pumps.
**Target Metrics**:
- Target: 15-25% reduction in cooling distribution unit (CDU) pumping power
- Aim: <2 degrees Celsius variance in GPU junction temperatures under peak training loads
- Target: 10%+ decrease in overall cooling energy usage compared to static flow baselines
- Aim: 0 thermal throttling events induced during predictive load ramping
**Target Case Studies**:
- A hyperscale AI data center (VP of Infrastructure) deploying the Pod Scale tier to autonomously modulate CDU flow across a 1 MW IT load, aiming to reduce pumping power by 15-25% during variable training workloads.
- A regional colocation provider (Director of Facilities) implementing the Rack Pilot on 5 high-density liquid-cooled racks to prove the elimination of over-cooling deadbands without requiring proprietary sensors.
- An enterprise HPC cluster (Data Center Operations Manager) utilizing the Facility Enterprise tier to establish a supervisory control layer above their legacy BMS, proving failsafe capabilities by maintaining zero thermal throttling events.
**Testimonial Targets**:
- VP of Data Center Engineering validating that the supervisory software successfully lowers energy costs without overriding or conflicting with underlying hardware safety loops.
- HPC Systems Administrator confirming that the software natively ingests existing baseboard management controller (BMC) telemetry via Redfish, requiring zero proprietary sensor installations.
- Director of Facilities expressing confidence in the fail-safe mechanism, noting that the system instantly defaults to static maximum cooling flow during simulated network drops.

## Startup Top Risks

**Risks**:
- Severity: existential · Description: Hardware OEMs restrict or lock down access to the live silicon thermal telemetry APIs required to run the flow modulation engine. · Mitigation Status: unmitigated
- Severity: high · Description: Data center operators refuse to grant an external autonomous system direct control over their mission-critical liquid cooling infrastructure. · Mitigation Status: in-progress
- Severity: high · Description: Incumbent facility management vendors bundle basic liquid loop controls into their existing enterprise contracts to block standalone deployments. · Mitigation Status: unmitigated
- Severity: moderate · Description: Lack of standardization across liquid cooling hardware valves requires building custom physical and software integrations for each new server rack deployment. · Mitigation Status: in-progress

## Startup Competitors

- [Legacy BMS](/Competitors/Legacy_BMS) — Status Quo
- [Vertiv Environet](/Competitors/Vertiv_Environet) — Incumbent
- [Static Room-Level Cooling](/Competitors/Static_Room-Level_Cooling) — Status Quo
- [CoolIT Systems](/Competitors/CoolIT_Systems) — Hardware Competitor
- [Schneider EcoStruxure](/Competitors/Schneider_EcoStruxure) — Incumbent

## Startup Solution Stack

- [Autonomous Cooling Service](/Services/Autonomous_Cooling_Service) — Service-as-Software
- [Flow Optimization Agent](/Agents/Flow_Optimization_Agent) — Agent
- [Telemetry Ingestion Worker](/Agents/Telemetry_Ingestion_Worker) — Agent
- [Silicon Thermal API](/Software/Silicon_Thermal_API) — Software
- [Pump Control Engine](/Software/Pump_Control_Engine) — Software

## Startup Story Brand

**Hero**:
- **Need**: to be the infrastructure architect who masters 100kW+ rack densities without overprovisioning energy
- **Want**: to match liquid cooling flow to real-time silicon thermal demand
- **Identity**: the data center facility engineer managing high-density AI clusters
**Plan**:
- Step: Deploy Pilot · Detail: Install the air-gapped controller to ingest telemetry from up to five high-density racks.
- Step: Check Performance · Detail: Verify that pumping power drops by 15-25% as the system eliminates over-cooling deadbands.
- Step: Scale Autonomously · Detail: Authorize pod-level flow control to sync facility-wide variable speed pumps with live compute loads.
**Guide**:
- **Empathy**: When your pumps draw 25% more power than necessary during training idle times, your cooling efficiency evaporates.
**Problem**:
- **Villain**: static flow margins
- **External**: Cooling distribution units run at maximum pumping power regardless of actual GPU load because Vertiv Environet cannot see inside the silicon.
- **Internal**: You feel like you are burning the facility's PUE budget just to play it safe.
- **Philosophical**: Every thermal engineer deserves precise control — not the waste of cooling an empty chip.
**Success**: Your cooling system breathes with your compute, slashing energy overhead while keeping GPUs perfectly chilled at peak training loads.
**One Liner**: Static room-level cooling costs data centers millions in wasted pumping energy. Basiscooling modulates liquid flow based on live silicon telemetry so you can slash PUE without thermal throttling.
**Positioning**:
- **So That**: you eliminate over-cooling and reduce pumping power by up to 25%
- **Unlike**: Legacy BMS and static cooling
- **For Whom**: data center engineers managing AI clusters
- **Category**: Autonomous Liquid Cooling Control Software
**Call To Action**:
- **Direct**: Launch Rack Pilot
- **Transitional**: View BMS Connector Schema
**Failure Stakes**:
- Wasted megawatt-hours in pumping energy
- Shortened hardware lifespan from thermal cycling
- Inability to meet aggressive PUE targets
**Transformation**:
- **To**: the architect who operates the world's most thermal-efficient AI pod
- **From**: the facility manager fighting static BMS setpoints
**Controlling Idea**: Cooling flow must mirror real-time silicon demand, not room-level averages.

## Startup Token Hero

**Genre**: founding-hypothesis
**Rendered**: Static room-level cooling costs data centers millions in wasted pumping energy. Basiscooling modulates liquid flow based on live silicon telemetry so you can slash PUE without thermal throttling.
**Mechanism**: spine-derived-v1
**Template Id**: spine-founding-hypothesis
**Vocab Fingerprint**: eba8a2fd0d727804

## Startup Token Positioning

**Genre**: moore-positioning
**Rendered**: Autonomous Liquid Cooling Control Software for data center engineers managing AI clusters. Unlike Legacy BMS and static cooling — you eliminate over-cooling and reduce pumping power by up to 25%.
**Mechanism**: spine-derived-v1
**Template Id**: spine-moore-positioning
**Vocab Fingerprint**: 8abfcc3d6e029299

## Startup Token Pitch Deck

**Genre**: pitch-deck
**Rendered**: Problem: Cooling distribution units run at maximum pumping power regardless of actual GPU load because Vertiv Environet cannot see inside the silicon.
Solution: Static room-level cooling costs data centers millions in wasted pumping energy. Basiscooling modulates liquid flow based on live silicon telemetry so you can slash PUE without thermal throttling.
Customer: data center engineers managing AI clusters
Unlike: Legacy BMS and static cooling
**Mechanism**: spine-derived-v1
**Template Id**: spine-pitch-deck
**Vocab Fingerprint**: 8e22dc965f8f947a

## Startup Token M E D D P I C C

**Pain**: Cooling distribution units run at maximum pumping power regardless of actual GPU load because Vertiv Environet cannot see inside the silicon.
**Metrics**: Target: Your cooling system breathes with your compute, slashing energy overhead while keeping GPUs perfectly chilled at peak training loads.
**Rendered**: Pain: Cooling distribution units run at maximum pumping power regardless of actual GPU load because Vertiv Environet cannot see inside the silicon.
Economic buyer: Data Center Operator
Metrics: Target: Your cooling system breathes with your compute, slashing energy overhead while keeping GPUs perfectly chilled at peak training loads.
Competition: Legacy BMS and static cooling
**Mechanism**: spine-derived-v1
**Competition**: Legacy BMS and static cooling
**Economic Buyer**: Data Center Operator
**Vocab Fingerprint**: b08a7f161a9db771

## Startup Token Cold Email

**Genre**: cold-email
**Rendered**: Subject: Autonomous Liquid Cooling Control Software for data center engineers managing AI clusters

data center engineers managing AI clusters — Cooling distribution units run at maximum pumping power regardless of actual GPU load because Vertiv Environet cannot see inside the silicon. Static room-level cooling costs data centers millions in wasted pumping energy. Basiscooling modulates liquid flow based on live silicon telemetry so you can slash PUE without thermal throttling.
**Mechanism**: spine-derived-v1
**Template Id**: spine-cold-email
**Vocab Fingerprint**: 4ea27e5fe0b2f0d0

## Startup Token Agent Spec

**Genre**: ai-agent-spec
**Rendered**: Autonomous Liquid Cooling Control Software. Static room-level cooling costs data centers millions in wasted pumping energy. Basiscooling modulates liquid flow based on live silicon telemetry so you can slash PUE without thermal throttling. Serves data center engineers managing AI clusters.
**Mechanism**: spine-derived-v1
**Template Id**: spine-ai-agent-spec
**Vocab Fingerprint**: e265ffebba16ca33

## Neighborhood

### Candidate solutions

- [Calculate Grower Liquidations](/Problems/Calculate_Grower_Liquidations) — candidate solution for · Problems

### What it offers

- [Silicon Flow Director](/Software/Silicon_Flow_Director) — offers · Software

### Composed of

- [Autonomous Cooling Service](/Services/Autonomous_Cooling_Service) — composes · Services
- [Flow Optimization Agent](/Agents/Flow_Optimization_Agent) — composes · Agents
- [Telemetry Ingestion Worker](/Agents/Telemetry_Ingestion_Worker) — composes · Agents
- [Silicon Thermal API](/Software/Silicon_Thermal_API) — composes · Software
- [Pump Control Engine](/Software/Pump_Control_Engine) — composes · Software

### Competitors

- [Vertiv Environet](/Competitors/Vertiv_Environet) — competes with · Competitors
- [Legacy BMS](/Competitors/Legacy_BMS) — competes with · Competitors
- [CoolIT Systems](/Competitors/CoolIT_Systems) — competes with · Competitors
- [Schneider EcoStruxure](/Competitors/Schneider_EcoStruxure) — competes with · Competitors
- [Static Room-Level Cooling](/Competitors/Static_Room-Level_Cooling) — competes with · Competitors

### Embodies

- [Software](/Theses/Software) — embodies · Theses

### Similar Startups

- [Coolerpark](/Startups/Coolerpark) — similar · Startups
- [Coolerengine](/Startups/Coolerengine) — similar · Startups
- [Coolerworks](/Startups/Coolerworks) — similar · Startups
- [Calorque](/Startups/Calorque) — similar · Startups
- [Rackourish](/Startups/Rackourish) — similar · Startups
- [Sensorsuite](/Startups/Sensorsuite) — similar · Startups
- [Valveforge](/Startups/Valveforge) — similar · Startups
- [Voltefect](/Startups/Voltefect) — similar · Startups
- [Biozone](/Startups/Biozone) — similar · Startups
- [Largefield](/CompanyTypes/Large-Scale_3PL_&_Cross-Docking_Hub/Problems/Cross-Dock_Throughput_Bottlenecks/Startups/Largefield) — similar · Startups
- [Heviv](/Occupations/Chemical_Equipment_Operators_and_Tenders/Problems/Pharmaceutical_Batch_Spoilage/Startups/Heviv) — similar · Startups
- [Clearimpact](/Startups/Clearimpact) — similar · Startups
- [Welderlane](/Startups/Welderlane) — similar · Startups
- [Zenith](/CompanyTypes/Regional_Independent_Bottlers/Startups/Zenith) — similar · Startups
- [Pyrometerquay](/Startups/Pyrometerquay) — similar · Startups
- [Crestensor](/Startups/Crestensor) — similar · Startups
- [Roileaf](/Startups/Roileaf) — similar · Startups
- [Voltoltage](/Startups/Voltoltage) — similar · Startups
- [Corrap](/Startups/Corrap) — similar · Startups

### Similar Agents

- [Thermal Load Agent](/Agents/Thermal_Load_Agent) — similar · Agents