Meta Signs Space-Based Solar Deal as AI Power Demands Push Infrastructure Boundaries

Meta Platforms signed an agreement with space energy startup Overview Energy on April 27, 2026, to secure up to 1 gigawatt of power from space-based solar energy for its data centers. The partnership marks the first commercial commitment to space-based solar power in the data center industry, as hyperscale operators grapple with power constraints limiting AI infrastructure expansion.

The deal gives Meta early access to capacity from Overview Energy's planned constellation of satellites in geosynchronous orbit that collect solar power and transmit it to Earth using near-infrared wide-beam lasers. Overview Energy founder and CEO Marc Berte's company emerged from stealth mode with backing from EQT Foundation and plans to demonstrate end-to-end functionality through a low Earth orbit mission in 2028.

Technical Architecture and Power Transmission

Overview Energy's system converts solar energy captured by satellites into near-infrared light, which is beamed to existing terrestrial solar facilities on the ground. The approach allows conventional solar installations to generate power when they would otherwise be idle, effectively extending their operational window beyond daylight hours and weather constraints.

The company has already conducted proof-of-concept testing, transmitting multiple thousands of watts of power from an aircraft to a ground receiver where it was converted into electricity. The near-infrared laser transmission uses wavelengths that are invisible and considered safe for atmospheric transmission, avoiding the technical and regulatory challenges associated with microwave power beaming approaches.

According to the Department of Energy, satellite-based solar panels can capture and transmit substantially more energy than terrestrial installations due to the absence of atmospheric interference and the ability to maintain continuous solar exposure. Overview Energy is building what it describes as the first satellite system designed for gigawatt-scale energy generation.

Infrastructure Adaptation Under Pressure

Meta has simultaneously begun deploying compute infrastructure in temporary structures at sites where power is available but construction timelines remain lengthy. The company is using tent deployments to reduce time-to-deployment at locations with secured electrical capacity, reflecting the acute pressure on data center construction schedules as AI workload demands outstrip traditional infrastructure planning cycles.

The space solar agreement sits alongside other power procurement strategies. Earlier this year, Vistra signed a 20-year nuclear power purchase agreement with Meta for 2,609 MW of capacity, indicating a diversified approach to securing baseload power for AI infrastructure.

Looking at the broader context here, this combination of unconventional infrastructure deployment and exotic power sourcing reflects the unprecedented speed at which AI compute demands are outpacing traditional utility and construction industry response times. The pattern recalls the early cloud buildout period, when hyperscalers similarly pushed beyond conventional data center approaches to meet rapidly scaling demand.

Space Solar Power: From Concept to Commercial Reality

NASA has published comprehensive analysis on space-based solar power technology, and the concept has moved from theoretical research to commercial development. Overview Energy's approach of transmitting power to augment existing ground-based solar installations sidesteps some of the regulatory and infrastructure challenges that have historically limited space solar adoption.

The geosynchronous orbital positioning allows continuous solar collection without the day-night cycle limitations of terrestrial installations. By targeting existing solar facilities as receivers, the system can leverage established grid interconnections and power conversion infrastructure rather than requiring entirely new ground-based receiving stations.

We have seen this pattern before, when urgent infrastructure needs accelerated adoption timelines for technologies that had languished in research phases. The early commercial internet drove fiber deployment schedules that had previously been measured in decades down to years. Similarly, mobile data growth compressed the normally deliberate pace of cellular infrastructure investment into rapid nationwide buildouts.

Commercial and Technical Implications

The 1 GW capacity commitment represents a significant scale for a first-generation space solar system. For perspective, that capacity matches the output of a large nuclear reactor or natural gas plant, though the actual delivered energy will depend on system efficiency and operational uptime factors that remain to be proven at commercial scale.

The partnership structure gives Meta early access rather than exclusive capacity, suggesting Overview Energy plans to serve multiple customers from the same satellite constellation. This approach distributes the substantial capital costs across multiple revenue streams while providing customers with some assurance of access to limited initial capacity.

Worth flagging: the 2028 demonstration timeline places commercial deployment several years out, meaning Meta's immediate AI infrastructure needs will still require conventional power sources. The space solar agreement appears positioned as a medium-term solution rather than a response to current capacity constraints.

Infrastructure Evolution Under AI Pressure

The convergence of tent deployments and space-based power agreements illustrates how AI infrastructure demands are pushing both construction and energy sectors beyond established practices. Traditional data center development cycles, measured in years from site selection to operation, conflict with the quarterly deployment schedules that hyperscalers now require for competitive AI service delivery.

Meta's approach combines immediate tactical solutions—temporary structures where power exists—with longer-term strategic investments in novel power sources. This parallel track strategy acknowledges both the urgency of current capacity needs and the likelihood that conventional power grid expansion will continue to lag infrastructure demand.

The space solar partnership also signals recognition that AI power requirements may exceed what terrestrial renewable sources can reliably provide at the scale and consistency required for training and inference workloads. As transformer architectures continue scaling and inference demand grows, baseload power availability becomes a competitive differentiator rather than a standard operational requirement.

The successful deployment of gigawatt-scale space solar power would mark a fundamental shift in how hyperscale infrastructure approaches energy sourcing, potentially opening pathways for data center placement that are decoupled from traditional grid constraints and geographic limitations.