| 1.1 |  The Physics Are Clear

Water has roughly 3,500 times the volumetric heat capacity of air. At the rack densities driven by modern AI accelerators -- 30 kW, 50 kW, even 70 kW per rack -- air cooling simply cannot remove heat fast enough without impractical volumes of airflow. The industry is moving to liquid cooling not as an option, but as a necessity.

| 1.2 |  How Rack Densities Have Changed

A decade ago, 5-8 kW per rack was standard. Traditional CRAC/CRAH units handled cooling adequately. Today, GPU-dense AI training clusters routinely exceed 40 kW per rack. At these densities, hot aisle temperatures spike, air mixing undermines efficiency, and the physical space required for adequate airflow becomes prohibitive.

| 2.1 |  Liquid Cooling Architectures

Several liquid cooling approaches have emerged: direct-to-chip (cold plates on processors), immersion (servers submerged in dielectric fluid), and facility-level systems like coolant distribution units (CDUs) and active rear-door heat exchangers (RDHx). All of these architectures share one thing in common: a heat exchanger sits at the core of the system, transferring heat between the server-side coolant loop and the facility rejection loop.

| 2.2 |  The Heat Exchanger Is the Bottleneck

The performance of the entire liquid cooling system is bounded by its heat exchanger. A higher heat transfer coefficient means more cooling capacity in a smaller unit. Lower pressure drop means less pumping energy. Better thermal effectiveness means tighter approach temperatures and more efficient heat rejection. Off-the-shelf plate exchangers are not optimized for these specific conditions.

| 3.1 |  Purpose-Built for the Problem

Phasic Energy designs heat exchangers for specific liquid cooling applications -- not generic catalog units repurposed from HVAC or industrial process lines. Each unit is matched to its fluid type, flow rate, temperature range, and physical constraints, delivering performance that general-purpose exchangers cannot achieve.