Key Takeaways

  • Switch engaged Goldman Sachs and JPMorgan Chase to lead an IPO that may value the company near $80 billion.
  • The offering reflects escalating demand for AI-oriented data center capacity across hyperscale and enterprise markets.
  • Broader IPO momentum in the United States and rising infrastructure investment expectations underpin investor interest.

Switch's decision to advance preparations for a public listing arrives as market appetite for infrastructure tied to artificial intelligence runs high. The company hired Goldman Sachs and JPMorgan Chase to guide an initial public offering that could raise up to $10 billion in the fourth quarter. Sources familiar with the matter indicated the valuation being discussed is close to $80 billion including debt, although the numbers remain in motion. Confidentiality continues to shape many of the conversations, which is typical for large listings.

Investor attention centers not only on Switch but on nearly every business positioned to support large-scale AI compute. Global data center capital spending is forecast to reach roughly $400 billion in 2027, according to Gartner. This expansion is driven by hyperscale environments and AI infrastructure rollouts, aligning with Switch's core focus on power, cooling, and connectivity at scale.

Colocation and wholesale providers are experiencing broad tailwinds alongside this capacity growth. Worldwide revenues in those segments are projected to grow at a 9% to 10% compound annual rate through 2028, eventually exceeding $140 billion based on research from IDC. This growth rate provides a favorable backdrop for operators preparing to step onto public markets, as investors closely monitor long-term contract structures and utilization patterns.

The timing aligns with a rebounded U.S. IPO market, led by a pipeline of large technology and AI-related listings. High-profile companies like SpaceX, Anthropic, and OpenAI continue to shape investor expectations. The active market environment encourages infrastructure firms to secure capital for upcoming expansion phases.

Switch operates large campuses that cater directly to AI workloads, supporting the power and cooling levels needed for GPU-intensive computing. Customers include Nvidia, Dell Technologies, and FedEx. Demand from GPU clusters used for training and inference continues to scale, rippling into national energy considerations. The International Energy Agency reported that U.S. data centers consumed roughly 2.5% of total national electricity in 2022, and AI-related requirements could push consumption up by 30% to 40% by 2030. Facilities designed for efficiency and high availability, often benchmarking against Uptime Institute Tier classifications, face increased utilization to meet these demands.

Broader public listings across the sector demonstrate the depth of market activity. Companies including SoftBank-backed SB Energy are pursuing their own offerings, while Brookfield-supported Csquare and Cerebras Systems also draw attention. Investors are evaluating the entire ecosystem for exposure, rather than focusing solely on compute vendors or hyperscale operators.

The landscape of hyperscale facilities is expanding rapidly. Synergy Research Group reported that the global count of hyperscale data centers surpassed 1,000 sites in 2023. By 2030, this number is expected to more than double as cloud and AI deployments expand. Competitive differentiation hinges on interconnection density, access to power, and the ability to deliver reliable performance under increasing thermal loads. Switch's footprint in key U.S. regions aligns with these requirements as the company prepares its IPO disclosures.

Reuters has followed the evolution of Switch's plans since the early stages in 2024. The company's formal preparations indicate internal confidence in current market conditions. Whether the final valuation settles near $80 billion or shifts will depend on broader economic signals in the months ahead, including interest rate trends and investor risk appetite.

The long-term trajectory of AI infrastructure demand remains clear. Enterprises continue to accelerate model training, cloud providers expand GPU availability, and colocation operators vie for strategic partnerships. Switch's planned offering positions the firm to capitalize on this momentum, testing how public market investors value the physical infrastructure underlying AI expansion.