In-depth Analysis of the Liquid Cooling Server Industry: Can It Support Valuation Growth for Concept Stocks Like Dingtong Technology?

According to the latest public information, current AI data center power consumption is approximately 4GW, which is sufficient to meet the simultaneous electricity needs of about 3 million U.S. households; Goldman Sachs research predicts that by 2028, AI may account for about 19% of total data center power demand. Deloitte’s June 2025 Infrastructure Energy Survey points out that data center expansion often outpaces the development of grid power capacity, and regional cluster expansion can easily lead to increased local grid pressure, posing binding challenges to power distribution and heat dissipation.

In high-density AI server scenarios, traditional air cooling faces bottlenecks such as thermal density limits, energy consumption, and space occupation. Direct liquid cooling (cold plate/cold plate type), immersion liquid cooling, and single-phase/two-phase refrigerant technologies have gradually become key solutions to high-power heat dissipation problems. Overseas related enterprises continue to launch cold plate solutions for AI data centers, indicating that technical routes are evolving towards higher heat dissipation density and stronger scalability.

• Upstream: Basic components such as refrigerants and working fluids, pumps/valves and pipelines, connectors, and heat exchangers
• Midstream: Liquid cooling cold plates/CDU (Cooling Distribution Unit), liquid cooling cabinets/rack solutions, immersion tanks and thermal management subsystems
• Downstream: Servers and complete machine manufacturers, data center operators, cloud vendors, and supercomputing centers

• According to public market research data, the global data center liquid cold plate market is expected to expand rapidly from 2024 to 2028; market forecast curves of multiple overseas liquid cooling-related enterprises show a steep upward trend, indicating strong demand growth.
• Main driving factors: Continuous increase in single-cabinet power of AI training/inference clusters, stricter PUE optimization policies and energy efficiency standards, and accelerated adoption of liquid cooling solutions by complete machine manufacturers and cloud vendors.

• Demand verification: Sustained growth in AI server shipments, clear procurement from leading customers, and fulfillment of large orders and expansion information
• Performance fulfillment: Visibility of order growth transmission to revenue and net profit, stability of gross margin
• Pricing power: Strong bargaining power and product premiums in high-prosperity segments, maintaining healthy profit margins
• Market share and overseas expansion: Core share in high-value-added components/system solutions, increased overseas revenue proportion enhancing performance certainty
• Policies and standards: Green data center policies promoting liquid cooling penetration

• Valuation premium overextension: Some targets trade at high multiples; if performance growth falls short of expectations, valuation corrections are likely
• Technical route uncertainty: Rapid evolution of cold plate/immersion/mixed solutions; single technical route faces substitution and iteration risks
• Supply chain and production capacity: Actual shipments are limited due to unmet expectations in delivery rhythm, yield rate, and ramp-up
• Cost transmission: Fluctuations in upstream raw materials and manufacturing costs affect gross margin; downstream bargaining pressure may suppress profit elasticity

• Order visibility and revenue matching: Observe consistency of existing orders, production capacity planning, and revenue guidance to confirm high prosperity has performance support
• Market share and pricing power: High share and pricing power in high-threshold components (e.g., cold plates, CDU, connectors) improve profit quality and sustainability
• Overseas order/revenue proportion: Increased orders from overseas leading customers usually enhance valuation premium and certainty
• Customer concentration and dependence: Diversified customer structure is conducive to valuation premium; excessive dependence on a single customer or project requires tracking order and performance fluctuations

Note: This call failed to obtain real-time market and company fundamental data for Dingtong Technology (300851.SZ), so no direct judgment is made on its specific valuation and business details. For further in-depth analysis, it is recommended to retrieve its financial statements, order announcements, and customer structure in the “In-depth Investment Research Mode” to form more targeted investment opinions.

• Liquid cooling is an inevitable path for AI data centers to move towards high density; demand-side prosperity is clear, technical route maturity is continuously improving, and the track has medium-to-long-term logic.
• Whether valuation growth can be realized depends on enterprise order conversion, market share, and profit elasticity. High prosperity does not mean all targets can deliver high valuations.

• Verify orders and production capacity: Focus on existing orders, production capacity expansion progress, and delivery rhythm
• Focus on high-barrier segments: High-value-added segments such as cold plates and CDU, core connectors, and thermal management subsystems have more valuation premium space
• Control valuation rhythm: Avoid chasing high during valuation overextension; wait for performance and order verification before allocation
• Overseas and diversified layout: Overseas revenue and diversified customer structure help increase valuation premium and certainty

[0] Jinling API Data (Failed to obtain real-time and financial data for 300851.SZ)
[1] Forbes - How AI Is Putting Data Center Infrastructure Under Pressure (https://www.forbes.com/sites/jenniferkitepowell/2025/12/29/how-ai-is-putting-data-center-infrastructure-under-pressure/)
[2] Yahoo Finance - Vertiv (VRT): Reviewing Valuation as AI Data Center Backlog … (https://finance.yahoo.com/news/vertiv-vrt-reviewing-valuation-ai-160536998.html)
[3] Yahoo Finance - Frore Systems liquid cooling coldplate advancements for AI data centers (https://finance.yahoo.com/news/ai-data-centers-earth-eventually-140000499.html)
[4] Yahoo Finance - The AI Data Center Boom Goes Public: From “Neoclouds” to … (https://finance.yahoo.com/news/ai-data-center-boom-goes-140000317.html)