Silicon Data Raises $30.5M to Price AI Compute

Silicon Data, the New York company building market intelligence for AI compute, has raised a $30.5M Series A. Valor Atreides AI Fund led the round, with a broad group of financial, strategic, and venture investors participating.

The financing matters because Silicon Data is attacking an expensive contradiction inside the AI economy. Compute has become one of the industry's most important inputs, but the market still lacks the pricing transparency, benchmarks, and risk tools that financial markets take for granted. Silicon Data wants to turn GPU capacity from an opaque operating cost into something buyers, sellers, and investors can measure with more discipline.

The round follows Silicon Data's verified $4.7M seed from March 2025, bringing disclosed funding across the two rounds to $35.2M. No valuation or detailed allocation of the Series A proceeds was disclosed.

What Happened

Valor Atreides AI Fund led the $30.5M Series A, joined by ME Group, DRW, F-Prime, Samsung, VanEck, Further, Jump, Tectonic, Wintermute, Breed, Hack, Blank VC, Sancus Ventures, and SoGal Ventures.

Several names represent continuity, not a first encounter. DRW and Jump co-led Silicon Data's $4.7M seed round in March 2025. Wintermute Ventures, Sancus Ventures, and SoGal Ventures also participated in that earlier financing, giving the Series A a meaningful layer of returning support.

Founder and CEO Carmen Li started Silicon Data around the idea that compute should be analyzed more like a market than treated as a stack of hardware invoices. Li remains the company's founder and chief executive. No current CTO was identified, so none is included here.

Building the Price Layer for AI Compute

Silicon Data's product portfolio reflects how quickly the compute problem has expanded. The company publishes daily GPU rental-price indices, builds performance benchmarks through SiliconMark, offers predictive pricing through SiliconPriceIQ, tracks compute carbon through SiliconCarbon, and models forward prices through a GPU Forward Curve. Its newer SiteIQ product focuses on underwriting the deployable GPU capacity and economics of AI data centers.

The distinction is important. A dashboard can tell a buyer what a GPU provider is charging today. A market-data layer attempts to normalize differences across providers, regions, hardware, contract structures, and time horizons so buyers can compare the economics underneath the quote. Silicon Data's stated ambition is closer to financial infrastructure than price scraping.

The company says its indices cover 95% of neo-cloud GPU providers and all major hyperscalers. That reach helps explain why the product could become useful to several constituencies at once: AI builders managing infrastructure budgets, cloud providers pricing capacity, data-center operators underwriting projects, and financial institutions evaluating compute exposure.

Why the CME Group Partnership Matters

The strongest validation of Silicon Data's market thesis arrived before the Series A. In May 2026, CME Group and Silicon Data announced plans to launch compute futures based on Silicon Data's daily GPU benchmarks later in the year, pending regulatory review.

If launched, the contracts would give market participants a way to manage volatility in GPU rental costs. That matters because AI compute pricing can vary across cloud providers, regions, chip generations, cluster configurations, and contract durations. Without a trusted reference price, procurement teams negotiate against a moving target while finance teams struggle to model the risk.

CME Group brings market structure and distribution. Silicon Data brings the underlying benchmarks. The partnership does not guarantee that compute becomes a mature commodity market overnight, but it shows that an established derivatives exchange sees enough economic exposure to build tools around it.

What the Series A Signals

The investor list is unusually aligned with the company's product direction. Trading firms understand volatility and benchmarks. Asset managers understand price discovery and risk. Strategic technology investors understand how fast AI infrastructure spending is expanding. Venture firms understand the value of owning a data standard early.

That mix makes the round more interesting than the amount alone. Silicon Data is not merely selling analytics into a hot infrastructure cycle. It is positioning itself at the point where operational data becomes a reference layer for commercial contracts, asset underwriting, and potentially financial products.

Carmen Li reported in June 2026 that Silicon Data had grown revenue 5x since January. The figure is company-reported and not independently audited, but it adds a commercial signal to the partnership and product activity. Customers appear to be looking for ways to understand compute economics before another large infrastructure commitment lands on the balance sheet.

The Bigger Industry Shift

AI infrastructure has spent several years being discussed primarily in physical terms: chips, power, cooling, networking, and data-center capacity. The next phase adds a financial layer. Buyers need benchmarks, forward views, utilization models, and instruments that help them understand not only whether capacity exists, but what it should cost and how that cost might move.

Silicon Data's Series A is a bet that the company defining those measurements can become part of the market's plumbing. The opportunity is substantial, but so is the execution burden. Benchmarks must remain credible, representative, and useful across a market that changes whenever a new GPU generation ships or a hyperscaler changes its terms.

The funding gives Silicon Data more room to build that layer. The market will decide whether its indices become durable standards, but the company's path is now clear: make compute prices visible, make performance comparable, and make AI infrastructure risk easier to understand before the invoice becomes the strategy.