Equinix Hits $100B Valuation on AI Inference Strategy

Colocation veteran stakes AI inference claim

Equinix has reached a $100 billion market capitalization, making it the most valuable data center real estate investment trust as the 1998-era colocation company positions itself for enterprise AI workloads. The milestone comes as shares climbed more than 33% this year, outpacing Digital Realty's $68 billion valuation.

The company announced an expanded Nvidia partnership at its first Equinix Horizon event, according to CNBC, which first reported the developments on September 2. The centerpiece is the Equinix Inference Exchange, which enables enterprise customers to run AI models through Nvidia's Enterprise Reference Architectures and Together AI's open-source inference platform across Equinix's global footprint.

Network scale meets AI demand

Equinix operates more than 280 data centers spanning 77 metropolitan areas with over 10,500 interconnected customers. The facilities are optimized for Nvidia's B300 Blackwell Ultra GPUs, with select liquid-cooled sites supporting the newer Vera Rubin chips.

Second-quarter results showed 16% revenue growth to $2.625 billion, with adjusted funds from operations per share rising 19% to $11.78. Adjusted EBITDA margin reached a record 53%, and management issued its largest guidance increase in company history.

The company's expansion strategy concentrates on proven markets—more than 80% of planned capacity additions through 2029 target its strongest 25 global metros. Customer churn dropped to 1.8%, below the historical 2% to 2.5% range.

Why it matters

Equinix is threading a strategic needle in the AI infrastructure buildout. Rather than competing for massive training clusters that hyperscalers are building, the company is targeting enterprise inference workloads where its connectivity fabric and distributed footprint create differentiation. Goldman Sachs Research projects hyperscalers will spend more than $5 trillion on AI infrastructure by 2030, and Equinix's positioning suggests colocation providers can capture meaningful share without head-to-head competition for the largest deals.

Capital intensity rises sharply

The AI opportunity comes with steep costs. Equinix nearly doubled its annual capital expenditure guidance to $5 billion to $7 billion, up from $3 billion to $4 billion previously. The increased spending creates greater financing requirements and potential balance sheet pressure, though management has emphasized protecting the company's investment-grade credit rating.

The infrastructure spending surge reflects broader industry dynamics as data center operators race to meet AI compute demand. Equinix's approach focuses on inference rather than training, potentially requiring less capital intensity per customer than hyperscale training facilities while still demanding substantial investment in power, cooling, and GPU-optimized infrastructure.

Details were first reported by CNBC.