Two-Phase Cooling Capital-Risk Board Adviser

The mandate

The board repeatedly returns to whether direct-to-chip and immersion systems should become the default thermal architecture for three campuses designed above one hundred kilowatts per rack. Engineering presents attractive power-density gains, operations warns about fluid stewardship and maintenance access, customers want incompatible service envelopes, and the investment case assumes supply and insurance conditions that have not survived independent challenge.

The cadence is two working days each month, a monthly ninety-minute session with the design and operations leads, attendance at four capital-project committee meetings, and one visit to each campus development team. The adviser must respond to a genuinely time-critical design query within two business days; ordinary papers receive a five-business-day review window. All scheduled participation and travel preparation are included in the retainer.

The appointment runs for twelve months and ends automatically. At month ten, the capital-project committee may decide whether an entirely new scope is justified after final investment decisions; there is no automatic renewal, evergreen clause or implied continuation. The current term cannot be extended beyond its stated year merely because a campus gate moves.

The adviser holds no line authority and carries no executive, design-certification, procurement or operating responsibility. Management owns specifications and supplier choices, engineers sign designs, and the board retains investment authority. The adviser may challenge evidence, request comparisons and recommend stage gates, but cannot instruct staff, negotiate vendors or represent the operator to customers or regulators.

Concurrent non-competing board work is expected, limited to three other material appointments. A role with a cooling-fluid producer, data-centre equipment vendor, engineering bidder, campus competitor, insurer underwriting these assets or customer seeking reserved capacity would create a conflict requiring disclosure and possibly recusal. Transaction-contingent fees and supplier-sponsored research are incompatible with this mandate.

Why the board wants this voice

The room contains deep facility knowledge but no director who has lived through high-density cooling from design promise to maintainable multi-tenant operation. Current debate treats efficiency, availability, serviceability, water, fluid, warranty and residual-value risks in separate papers. The chair wants an independent operator-investor perspective before architecture choices become irreversible civil works.

What you will own

• Press the board to distinguish proven thermal performance from assumptions about maintenance labour, fluid life, component warranty and customer behaviour.
• Test campus density forecasts against deployment timing, rack heterogeneity, partial loading, stranded mechanical capacity and realistic customer adoption.
• Challenge comparisons between air, direct-to-chip, rear-door and immersion designs using common availability, energy, water and lifecycle-cost boundaries.
• Examine supplier concentration across cold plates, manifolds, distribution units, fluids, seals, controls, commissioning skills and replacement inventory.
• Shape staged investment gates that preserve architectural options until performance, operability, insurability and contracted demand are evidenced.
• Probe failure and maintenance scenarios involving leaks, contamination, mixed fluids, pump loss, heat-rejection limits and occupied-rack intervention.
• Frame the closing board view on which cooling choices are strategic commitments, reversible experiments or customer-specific exceptions.

Candidate qualifications

• Governed or operated high-density data-centre cooling beyond pilot stage, including direct liquid or immersion technology in occupied facilities.
• Challenged a thermal business case using maintainability, spares, warranties, fluid management, insurance and residual infrastructure value.
• Participated in capital gates where rack-density forecasts and customer contracting materially affected mechanical and electrical design choices.
• Evaluated competing cooling architectures under consistent energy, water, reliability, service-access and lifecycle boundaries rather than vendor claims.
• Managed or independently reviewed a serious liquid-cooling commissioning, contamination, leak or supply-chain failure and its commercial consequences.
• Served boards without accepting contingent compensation or allowing a supplier relationship to shape supposedly independent technical advice.

Non-negotiables

• Can attend all four committee meetings and complete the Tokyo, Osaka and Chiba campus visits within the stated cadence.
• Will disclose investments, retainers, research funding and close relationships across cooling, data-centre, engineering and insurance markets.
• Accepts that management owns engineering and procurement decisions and may reject the adviser’s recommendation with recorded rationale.
• Brings operational evidence from liquid-cooled production capacity; laboratory prototypes and sustainability policy alone are insufficient.

1. 49 words maximum. Which liquid-cooling assumption most often survives design review but fails after a data hall becomes occupied?
2. 49 words maximum. List the supplier, customer, investment or research relationships this board would need you to disclose.
3. 49 words maximum. Confirm the twelve-month Tokyo cadence and describe one capital gate you would insist remain reversible.