Our Analysis·May 31, 2026·9 min read

What OpenRouter’s $113M Series B Signals for AI Inference Gateways

A $113M Series B at a reported $1.3B post-money for the company trying to become the neutral control plane for AI inference traffic.

Context

Last week, OpenRouter announced a $113M Series B led by CapitalG, with participation from NVentures, ServiceNow Ventures, MongoDB Ventures, Snowflake Ventures, Databricks Ventures, AMP PBC, Pace Capital, a16z, and Menlo Ventures. The round reportedly values the company at about $1.3B post-money, more than 2.6x the roughly $500M valuation reported around its prior seed and Series A financing. OpenRouter says it now processes 25T tokens per week, up from 5T tokens per week six months earlier, across more than 8M users and developers and 400+ models.

The round is exciting because it prices a simple but powerful shift: AI inference is becoming too fragmented, expensive, and mission-critical to manage model by model. OpenRouter is not just selling access to many models. It is trying to sit between applications and model providers as the routing, billing, governance, observability, and optimization layer. That is why the cap table matters. Google, Nvidia, ServiceNow, MongoDB, Snowflake, and Databricks all showing up around the same neutral inference layer is a loud signal that enterprise AI will probably stay multi-model for a while.

The tension is that the category is both validated and vulnerable. OpenRouter has a large usage lead, a reported ~$153M in cumulative funding, and roughly 4.5x to 4.8x more capital than its direct competitors combined. But routing can also be bundled by cloud platforms, developer platforms, and enterprise software vendors. The core question is whether OpenRouter’s traffic creates a compounding data advantage, or whether AI gateway functionality becomes a feature inside bigger platforms.

In our AI infrastructure market deck, you will find everything you need to understand the market

Q1Does OpenRouter have enough usage data to build a model-routing moat?

OpenRouter may have one of the strongest real-world model usage datasets outside the major labs and cloud platforms.

The company reported 25T tokens per week, or roughly 100T tokens per month, across 8M+ users and developers and 400+ models. That scale matters because token volume is the closest available proxy for actual usage, user demand, and future revenue potential.

The most important metric is token volume. A router with massive traffic can become a data layer: it can learn which models are cheaper, faster, more reliable, or better for specific tasks. If that learning loop compounds, routing quality could become a moat, not just a feature.

Methodology note This answer uses company-reported usage metrics from OpenRouter’s Series B materials: 25T weekly tokens, about 100T monthly tokens, 8M+ users and developers, and 400+ models. These are treated as company-claimed platform metrics, not audited revenue or customer-validated contract data. See full methodology below.

Q2Is OpenRouter’s $1.3B valuation supported by usage growth?

OpenRouter’s valuation has not increased as aggressively as its usage curve.

We do not have a confirmed revenue number, but we do have token volume. OpenRouter’s reported token volume grew 5x in six months, from 5T tokens per week to 25T tokens per week. Over a similar period, its valuation increased from roughly $500M around the prior financing to about $1.3B, or about 2.6x.

The usage metric is company-reported platform activity. The valuation comparison uses the reported prior valuation around the seed or Series A and the reported Series B post-money valuation.

This directional check suggests investors may believe the usage is real and valuable. But the valuation did not fully chase the usage curve, likely because investors still worry about take rate, margins, competition, and whether routing can become a defensible moat.

For more data on this, please check full memo.

Methodology note The valuation comparison uses the roughly $500M prior valuation reported around OpenRouter’s seed or Series A and the reported $1.3B Series B post-money valuation. Usage growth is measured from OpenRouter’s company-reported move from 5T to 25T weekly tokens over six months. See full methodology below.

Q3How fast has OpenRouter’s funding increased from seed to Series B?

OpenRouter’s funding trajectory accelerated sharply across rounds.

The company went from a $12.5M seed to a $28M Series A to a $113M Series B. That is 2.2x from seed to Series A, 4.0x from Series A to Series B, and 9.0x from seed to Series B.

The timing matters too. Seed to Series A took roughly 2 months, while Series A to Series B took around 13 to 14 months.

The check size did not just grow. The growth rate of the check size expanded. That usually means the company gathered stronger proof before raising the next round. In simple terms: investors saw more evidence and wrote a much bigger check.

Methodology note Funding acceleration is calculated from disclosed round amounts: $12.5M seed, $28M Series A, and $113M Series B. Timing comparisons are based on the disclosed or reported financing chronology, with round windows treated by announcement date unless a financing date was separately disclosed. See full methodology below.

Q4Is OpenRouter’s Series B large compared with AI infrastructure Series B rounds?

OpenRouter’s $113M Series B is huge for the AI gateway category, but normal for broad AI infrastructure.

We compared OpenRouter not only to direct AI gateway competitors, but also to broader AI infrastructure Series B rounds. That broader set included AI inference clouds, AI-native cloud infrastructure, chips, agent infrastructure, and evaluation or observability companies.

Examples in the benchmark set included Nscale, MatX, Together AI, Fractile, Panthalassa, DeepInfra, Parallel Web Systems, Modal, and Braintrust. Across that group, the median Series B round was about $110M, very close to OpenRouter’s $113M Series B.

That matters because OpenRouter is not raising like a small developer tool. Its round size places it near the middle of a high-quality AI infrastructure benchmark set, signaling that investors are treating it like a serious infrastructure company.

Methodology note The AI infrastructure benchmark includes recent Series B rounds from Nscale, MatX, Together AI, Fractile, Panthalassa, DeepInfra, Parallel Web Systems, Modal, and Braintrust. OpenRouter’s $113M Series B is compared against that broader benchmark median of about $110M. See full methodology below.

Q5Does OpenRouter now have a funding moat in the model-routing market?

Yes. OpenRouter has roughly 4.5x to 4.8x more funding than all direct competitors combined.

After the Series B, OpenRouter has raised about $153M. The five direct competitors combined have raised roughly $32M to $34M. That gives OpenRouter a 4.5x to 4.8x cumulative funding advantage versus the whole direct competitor group.

This funding gap matters because routing infrastructure benefits from scale. More capital can fund model integrations, enterprise features, reliability, pricing experiments, and go-to-market. It does not guarantee defensibility, but it gives OpenRouter a much larger strategic budget than the rest of the direct market.

It’s actually something we elaborate on in our latest market report.

Methodology note The direct competitor set used here includes Portkey, Martian, LiteLLM, Unify, and Helicone, following the Q&A funding table. Funding totals are treated as estimates where public data is conflicting or incomplete, especially for LiteLLM, Unify, and Helicone. See full methodology below.

Q6Is the AI gateway market fragmented or winner-concentrated?

The AI gateway category is already highly concentrated.

There are multiple companies in the space, but only a few have meaningful investor backing. More importantly, most of the top-three funding belongs to OpenRouter.

OpenRouter, Portkey, and Martian together account for about 97.8% of tracked last-24-month capital in the AI inference gateway category: $177M out of roughly $181M. OpenRouter alone accounts for about $153M of that total.

This signals that the market is not a broad funding war. It looks more like one giant, OpenRouter; one smaller challenger, Portkey; and a few early bets, such as Martian. That concentration could help OpenRouter compound faster if usage, data, and enterprise demand keep moving toward the largest router.

Methodology note Category concentration is calculated by summing disclosed or estimated funding rounds in the AI inference gateway and model exchange category over the last 24 months. OpenRouter, Portkey, and Martian are the top-three funded companies in this retained set, with approximately $177M out of $181M in tracked category capital. See full methodology below.

In our AI infrastructure market deck, you will find everything you need to understand the market

Read more

§

Methodology, Sources & Disclosure

All timing comparisons in this note are measured as of May 31, 2026. Funding-round time windows refer to announcement dates, not legal close dates, unless a close date is separately disclosed.

The retained investment thesis behind OpenRouter’s Series B is that AI inference is becoming too fragmented, expensive, and strategically important for companies to manage model by model or provider by provider. This thesis was retained because the round was framed around routing, governance, billing, observability, optimization, enterprise controls, and the idea that traffic across many models can become a control-plane and data advantage.

The category used for market-activity analysis is AI inference gateway and model exchange. It includes companies that provide a unified interface for developers and enterprises to access, route, govern, observe, optimize, and pay for AI model inference across multiple model providers. It excludes pure foundation model companies, pure inference hosting companies, cloud GPU providers, vector databases, generic API gateways, and AI application companies that only use multiple models internally.

The direct competitor set used for funding comparisons in the Q&A includes Portkey, Martian, LiteLLM, Unify, and Helicone. In the broader analysis materials, Not Diamond is also treated as a direct routing competitor, while pure model providers, inference clouds, hyperscaler platforms, and broader AI platforms such as OpenAI, Anthropic, Google, xAI, DeepSeek, Together AI, Fireworks AI, Baseten, Groq, Replicate, AWS Bedrock, Azure AI Foundry, and Google Vertex AI are excluded from direct competitor rankings. Competitor funding rankings include only private or venture-backed companies with comparable disclosed financing data, so acquired units, public-company divisions, and companies with insufficiently comparable financing data are discussed qualitatively or treated as estimates where needed.

Category capital concentration is calculated by summing disclosed funding rounds in the retained category set over the relevant period. When round amounts are disclosed as estimates, conflicting tracker values, or ranges, concentration figures are treated as approximate and use the clearest conservative estimate. In this note, OpenRouter’s roughly $153M in tracked funding represents about 84.6% of the roughly $181M tracked over the last 24 months in the retained category set.

We selected these sources because they come either from direct company announcements, which are the primary source for funding, product, usage, and corporate milestones, or from tier-1 / authoritative publications, which provide independent validation, valuation context, competitor context, and comparable market signals: OpenRouter Series B announcement, Business Wire Series B release, CapitalG investment thesis, Wall Street Journal coverage of OpenRouter’s seed and Series A, TechCrunch coverage of OpenRouter’s Series B valuation, Andreessen Horowitz note on OpenRouter, Orrick note on OpenRouter’s seed and Series A financing, Portkey Series A announcement, Martian funding coverage, LiteLLM GitHub repository, Helicone website, Nscale Series B announcement, TechCrunch coverage of MatX Series B, Together AI Series B announcement, DeepInfra Series B announcement, Modal Series B announcement, FinSMEs coverage of Braintrust Series B.

We are not affiliated with OpenRouter, its investors, or the named comparable companies. No payment, consideration, or commitment of future business has been received from OpenRouter, its investors, or any named comparable company in connection with this note. Nothing herein constitutes investment advice or an offer to transact in any security.

In our AI infrastructure market deck, you will find everything you need to understand the market