TakeMe2Space books Falcon9 for orbital data centre push

TakeMe2Space has announced plans for a Falcon9 launch in 2028, aiming to deploy what it describes as a networked orbital data centre. According to reports, each of its Generation-2 satellites will carry Nvidia GPUs, 1,500 watts of compute capacity, and 100 TB of storage, with optical inter-satellite links enabling coordinated operation. The Bengaluru-based startup appears to be positioning orbital infrastructure as a potential solution to the power, cooling, and land constraints facing terrestrial data centres.

If successful, this approach could leverage abundant solar energy and eliminate cooling costs, though the viability of such a model remains unproven. TakeMe2Space’s earlier Generation-1 satellites may have served as a proof-of-concept; Generation-2 is reportedly intended to function as a distributed, cloud-scale platform. Optical links between satellites could allow workloads to be distributed across the constellation, potentially turning a swarm of small satellites into a single logical data centre.

However, the feasibility of this vision is far from certain. Orbital compute remains a niche market, and its economic viability depends on whether customers will accept the trade-offs of space-based infrastructure. While terrestrial data centres in markets like India offer competitive pricing for GPU-hour usage, orbital infrastructure would need to approach similar cost structures to attract cloud providers and AI startups.

The inclusion of Nvidia GPUs aligns with trends in terrestrial data centres, where the company dominates AI training and inference workloads. TakeMe2Space has suggested these GPUs could be used for "AI training, scientific simulations, and secure data processing," indicating a focus on enterprise customers seeking to offload compute-heavy tasks from congested terrestrial infrastructure. Whether such demand exists at scale is unclear.

Regulatory challenges could also shape the project’s trajectory. While orbital infrastructure avoids land-use and power-grid constraints, it still requires spectrum licenses and orbital slots, both of which are increasingly competitive as India’s space sector expands. The lack of public fanfare around the SpaceX launch agreement may reflect the company’s efforts to secure capacity before slots become harder to obtain.

The 2028 launch timeline aligns with broader developments in space infrastructure. Recent milestones, such as Isar Aerospace’s first commercial orbital launch from Europe and Pixxel’s $100 million funding round, suggest growing commercial interest in orbital capabilities. TakeMe2Space’s bet is that compute, rather than imaging or communications, will be the next frontier. If successful, the constellation could serve as a model for other startups exploring space-based data centres. If not, the satellites may still have value as standalone compute nodes, though the broader vision of a networked orbital cloud would remain unrealized.

What to watch next: TakeMe2Space’s ability to secure anchor customers. Without commitments from cloud providers or AI startups, the 2028 launch risks being a technology demonstration rather than a commercial product. The coming months should clarify whether the company can translate its hardware ambitions into tangible demand.

Sources: economictimes.indiatimes.com