E2B’s Above-$100M Run Rate Needs a Definition Before a Sandbox Buying Decision

E2B says it has passed $100 million in annualized run rate in the body of a November 5 Business Wire release carried by FinancialContent. The headline instead says $120 million. The defensible summary is above $100 million with an unresolved headline/body discrepancy, not audited $120 million recurring revenue.

The company also reports close to 200,000 teams and 94 Fortune 100 companies signed up. It says its computer count moved from 1 billion in June to 2 billion in October, with 3 billion in the next 30 days presented as a forecast. These are vendor claims about adoption and activity, not independent evidence of sandbox security.

The release supports taking demand for agent execution infrastructure seriously. It does not settle the revenue definition or tell a buyer whether a sign-up became a paying production customer. Keep those distinctions intact before using the announcement in a supplier-selection document.

A run-rate claim needs a consistent definition

The release’s conflicting $100M and $120M figures cannot be reconciled from the supplied facts. Do not choose the larger amount because it appears in the headline, or silently rename annualized run rate as audited recurring revenue. State the inconsistency and ask the company which measure and reporting point it intends.

E2B also reports $8 million a year earlier. That comparison describes the company’s reported trajectory, but it does not establish whether the definitions are consistent across the periods. Avoid calculating a growth multiple when the underlying revenue basis remains unresolved.

For buyers, the practical issue is not to adjudicate a financial statement from a press release. It is to avoid importing unsupported precision into a supplier-risk review. A fast-growing vendor claim may justify deeper diligence without serving as proof of durable revenue or future service continuity.

Sign-ups describe reach, not paid production usage

The reported 94 Fortune 100 sign-ups are a specific adoption claim. They are not a statement that those organizations pay for the service or use it in production. Keep the word signed up attached whenever the number appears in an internal summary.

The same rule applies to close to 200,000 teams. The figure can indicate broad interest in the offering while leaving paid conversion and retained usage unanswered. It does not establish how many teams have workloads resembling yours or what those teams require from their sandbox provider.

Ask for evidence relevant to the proposed purchase rather than an implied endorsement from a recognizable cohort. A useful reference conversation would concern the workload and operating requirements you actually have. That is a suggested diligence step, not a claim that the release supplies such references or validates their suitability.

Activity volume is not an isolation test

E2B reports growing from 1B computers in June to 2B in October. The supplied facts do not define those units in enough detail to turn them into a measure of distinct customers or production workloads. Preserve the company’s terminology rather than substituting a more convenient operating metric.

The forecast of 3B within the next 30 days belongs in a separate category from the historical milestones. It describes what the company expects, not what has already happened. Combining the forecast with observed activity would make the adoption story more certain than the release supports.

Even a well-defined activity count would not independently establish isolation quality. A buyer still needs to assess what an agent may execute and where its outputs can go. The release does not provide independent security tests, so neither the activity total nor the sign-up count can substitute for that review.

The commercial trade-off includes operational dependence

E2B reports $38 million raised in total, including a $21 million Series A in July 2025. Financing and revenue are different evidence categories. The funding figures do not establish the terms of your future contract or prove that the service meets your availability and incident-handling requirements.

A sandbox can become a dependency in the agent workflow, so the acceptance decision should include what happens when execution fails or access must be removed. These are questions for the buyer to investigate, not deficiencies demonstrated by the announcement. The available facts leave their answers open.

The honest cost of evaluation is doing that work despite impressive adoption claims. Have engineering examine the execution boundary while procurement asks for the commercial definitions. A financial clarification cannot settle technical isolation, and a successful contained test cannot reconcile the release’s revenue discrepancy. Both reviews should retain their own evidence and unresolved items.

What To Do About It

• Clarify the financial label: This week, have procurement request the definition and reporting basis of annualized run rate, preserving the unresolved headline figure of $120M.
• Separate adoption categories: Have the supplier-review owner record sign-ups independently from paid references and production usage, including the reported Fortune 100 sign-ups of 94.
• Test the execution boundary: Have the engineering owner define permitted actions and data access for a contained evaluation rather than accepting activity volume as assurance from E2B.
• Keep forecasts separate: Have the document owner label the next milestone as an expectation, not an achieved result, preserving the forecast horizon of 30 days.

The Bottom Line

E2B’s reported above-$100 million run rate is a substantial demand signal with an unresolved definition and conflicting headline. Its sign-ups and computer counts do not establish paid production adoption or independently tested security. Evaluate sandbox suitability against your execution boundary while asking for financial clarification. A buying decision needs evidence that the service fits the workload, not a more flattering rewrite of the release.