Equinix는 Nvidia 및 Together AI와 파트너십을 맺어 모델 배포를 간소화하는 AI Inference Exchange를 출시했다.
On 2 September 2026, Equinix (NASDAQ: EQIX) announced plans to launch a distributed AI inference service for global enterprises, developed in partnership with Nvidia and newcomer Together AI. The offering, named the Equinix Inference Exchange, will become available beginning in the first quarter of 2027.
The service integrates Nvidia Enterprise Reference Architectures with Together AI’s inference platform, deployed across Equinix’s network of more than 280 data centres spanning 77 metropolitan areas and over 230 cloud on-ramps.
This partnership marks Equinix’s third publicly announced AI-infrastructure collaboration with Nvidia in 2026. It follows an April announcement regarding an AI Discovery Hub with HPE and Nvidia in Hong Kong, as well as a June disclosure of a secure AI factory initiative with Cisco and Nvidia, according to Equinix’s newsroom. Together AI has not previously appeared on the company’s list of disclosed AI partners. The sequence of deals suggests a strategic approach of layering multiple AI-inference partnerships atop Equinix’s core colocation and interconnection infrastructure, rather than relying on isolated announcements.
As of publication, no Form 8-K or other regulatory filing referencing the Inference Exchange had been submitted to the SEC’s EDGAR database, which aligns with standard practice for product launches that do not trigger material-event disclosure requirements.
Following the announcement, EQIX shares rose approximately 2%, according to a Reuters wire report distributed via Investing.com. Prior to the news, the stock had been trading near a 20-day low of $1,010.515, reflecting a 4.76% decline over the preceding 20 trading sessions based on consolidated exchange data. Short-selling activity had also increased leading up to the announcement. FINRA data shows the daily short-volume ratio for EQIX climbed from 0.443 on 21 August to 0.662 by 1 September. While the ratio alone cannot determine whether the subsequent price rally stems from genuine buying conviction or short covering triggered by the positive news, the accumulation of bearish positioning beforehand provides important context for the magnitude of the move.
Equinix’s underlying business continues to expand steadily. According to SEC filings, quarterly revenue increased from $2.127 billion in the first quarter of 2024 to $2.625 billion in the second quarter of 2026. Over the same period, net income grew from $231 million to $479 million. The financial trajectory has remained broadly consistent rather than volatile. Filings indicate that net income reached $301 million and $297 million in the second and third quarters of 2024, respectively, before rising to $343 million, $368 million, and $374 million across the first three quarters of 2025. Diluted earnings per share mirrored this upward trend, increasing from $2.43 in the first quarter of 2024 to $4.83 by the second quarter of 2026, with each successive quarter outperforming the previous one. This consistent growth in revenue and profitability provides Equinix with the financial flexibility to continue funding AI-focused partnerships without undue strain on its balance sheet, although the SEC filings do not specifically detail the commercial terms of the Together AI or Nvidia arrangements.
The announcement arrives amid a macroeconomic environment characterized by stable yet elevated U.S. borrowing costs. According to Federal Reserve data, the 10-year Treasury yield stood at 4.75% on 31 August 2026, a marginal increase from 4.73% the previous session. The 2-year yield remained at 4.34%, maintaining a 10-year/2-year spread of 0.40 percentage points—a level that has shown minimal fluctuation throughout the period. Separately, U.S. unemployment improved to 4.1% in July 2026 from 4.2% the previous month, according to Bureau of Labor Statistics data compiled by the Federal Reserve Bank of St. Louis, while headline consumer prices experienced a modest uptick. Although these economic indicators do not directly impact Equinix’s announcement, they contextualize the broader financing landscape in which data centre operators are evaluating additional capital expenditures for AI infrastructure.
Equinix has not yet disclosed pricing, capacity commitments, or anchor customers for the Inference Exchange ahead of its Q1 2027 launch. Investors will receive their next scheduled update on the initiative’s commercial progress when the company releases its third-quarter earnings later this year.
This article is provided for informational purposes only and does not constitute investment advice or a recommendation to purchase or sell any asset. Market conditions evolve rapidly, and all figures are accurate as sourced at the time of publication. Readers should conduct their own independent research before making any financial decisions.
Martin Holloway spent nearly two decades on bank trading floors in London’s financial district before transitioning to journalism. He studied Economics at a traditional British university, joined a graduate program at a major UK clearing bank in the late 1990s, and specialized in fixed income early in his career. Most of his trading tenure was spent on macro desks at European banks, managing short-end and cross-asset positions through the 2008 financial crisis and the extended post-crisis interest rate environment. He left active trading toward the end of the previous decade, completed a brief consulting stint on the buy side, and began writing full-time in 2020. His coverage focuses on markets he understands deeply: interest rates, foreign exchange, UK banking, gilt issuance, and the corporate developments that drive them. He maintains a detailed thesis on why sell-side analysts failed to anticipate the 2022 gilt crisis and is willing to elaborate at length. Martin resides in the South East of England with his family. He reads the Financial Times each morning out of routine rather than passion, and regards most market commentary as promotional material authored by individuals who have never managed overnight risk.