Orient: In the short term, computing-power coordination will focus mainly on "power ensuring computing"; in the medium to long term, it will move toward mutual support between computing and power.

Orient released a research report stating that the short-term contradiction in computing-power coordination is "supply assurance and power quality." Computing centers are not only high-energy-consumption loads but also high-value loads, and power supply assurance takes priority over electricity cost optimization. In the short term, computing-power coordination will focus mainly on purchasing electricity from the grid and having electricity serve computing, with emphasis on reliable power supply, voltage sag mitigation, and backup power and distribution upgrades. In the medium to long term, "new energygrid-forming energy storagecomputing power" is a potential direction for integration. Orient's main points are as follows: The short-term contradiction in computing-power coordination is "supply assurance and power quality." IT equipment in intelligent computing centers accounts for 45%60% of total electricity consumption, with AI accelerators being the largest single power-consuming source. Electricity costs can account for about 40% of IDC business costs and about 10% of AIDC costs. However, AI servers are extremely sensitive to voltage sags. In actual tests, a voltage drop of just tens of milliseconds on the grid can cause GPU training tasks to report errors and exit, resulting in loss of intermediate data and extremely high restart costs. Therefore, computing centers are not only high-energy-consumption loads but also high-value loads, and power supply assurance takes priority over electricity cost optimization. In the short term, the opportunity cost of computing power is far higher than electricity price fluctuations, and computing power will not make concessions for the grid. On the training side, the fixed-cost loss from idle GPUs far exceeds short-term electricity price fluctuations, and model capability takes priority over electricity cost optimization. On the inference side, rigid tasks such as L0 and L1 are latency-sensitive. Although L2 and L3 can shift load away from peak periods, their current share is limited, and pricing guidance mechanisms have not yet been systematically formed. At the same time, the grid as a whole provides reliable backstop support, and user-side electricity bills do not fully reflect grid congestion costs. Therefore, in the short term, computing-power coordination will focus mainly on purchasing electricity from the grid and having electricity serve computing, with emphasis on reliable power supply, voltage sag mitigation, and backup power and distribution upgrades. The power supply assurance system is under pressure from the bottom up. Cabinet power distribution, system backup power, and campus power supply all face systemic upgrades. On the cabinet side, single-cabinet power has risen from 25 kW in traditional IDCs to 2050 kW, and the GB200 supernode full cabinet has power consumption of 120 kW, with 800V HVDC becoming the future route. On the backup power side, traditional UPS plus diesel generators are transitioning toward BESS and grid-forming energy storage. On the campus side, large intelligent computing centers are moving from 10 kV access to 110 kV and even 220 kV direct connection, grid operation and maintenance assurance costs are rising, and direct green power connection has become an important alternative. Spatiotemporal matching is subject to both economic and physical constraints, with economic constraints taking priority in the short term. The adjustability of computing loads is graded from L0 to L3: traditional IDC loads are stable but have weak adjustability; training can in theory resume from checkpoints and be frequency-reduced, but the opportunity cost is high; online inference has weak adjustability, while offline inference and some training can shift load away from peak periods. Fluctuations in new energy output are transmitted as high-frequency electricity prices through the spot market, but current peak-valley price spreads and pricing mechanisms are still insufficient to drive large-scale concessions by computing power. In the short term, computing-power matching is more about "electricity ensuring computing" rather than "computing yielding to electricity." In the medium to long term, after the cost of computing facilities declines, computing power will gradually become more sensitive to electricity prices, and "new energygrid-forming energy storagecomputing power" is a potential direction for integration. As GPU depreciation pressure eases, market-based peak-valley electricity price spreads widen, and the costs of grid-forming energy storage and direct green power connection decline, the adjustability of computing power will gradually be released. On the path forward, advanced enterprises should first pilot computing-power price linkage and dynamic pricing, then promote the integration of direct green power connection and grid-forming energy storage, and ultimately move from "electricity serving computing" to "mutual support between computing and electricity."