Neon announced a $46 million Series B on August 1, 2023, led by Menlo Ventures, to expand its hosted, serverless PostgreSQL platform. Founders Fund, General Catalyst, GGV Capital, Khosla Ventures, Elad Gil, Snowflake Ventures and Databricks also participated, taking Neon’s disclosed funding to $104 million. The company’s pitch is not a new database language or an AI model: it is PostgreSQL redesigned around separated compute and storage, branching, autoscaling, scale-to-zero and vector search.
What Neon’s Series B funded
Neon’s announcement said the company would use the capital for serverless Postgres, edge-computing capabilities, vector search, open-source Postgres work and partnerships with Vercel, Replit, Hasura and Cloudflare. Menlo Ventures partner Tim Tully joined Neon’s board. Neon also said it planned to grow from roughly 50 employees to about 100 by the end of 2023; that was a forward-looking plan, not a confirmed result.
The funding was announced on August 1, 2023, and Neon published its follow-up strategy post on August 2. It is historical funding news, while the product and pricing context below reflects later service information.
Read the funding announcement and Neon’s strategy post.
Free tools Windows power users keep installed
One-click scans. No signup required.
#1 Best Overall
The announced deal
| Item | Detail |
|---|---|
| Round | Series B |
| Amount | $46 million |
| Lead investor | Menlo Ventures |
| Other participants | Founders Fund, General Catalyst, GGV Capital, Khosla Ventures, Elad Gil, Snowflake Ventures and Databricks |
| Total disclosed funding | $104 million at the time |
| Previous round | $30 million Series A |
| Board change | Tim Tully of Menlo Ventures joined Neon’s board |
What Neon is—and is not
Neon is a hosted, multi-cloud PostgreSQL service. It remains PostgreSQL underneath; it is not a replacement relational language, a foundation model, an inference engine or a complete AI platform. Neon manages infrastructure around PostgreSQL and adds workflows intended for modern application development.
Its product combines decoupled compute and storage, autoscaling, scale-to-zero for inactive compute, database branching, managed backups and restore/history features, connection pooling, serverless-oriented drivers and PostgreSQL extensions for vector search. Neon’s Vercel marketplace listing also highlights branching, read replicas, point-in-time recovery and time-travel queries.
Why separate compute from storage?
Traditional managed databases commonly attach storage to a provisioned database instance that runs continuously. Neon’s architecture treats storage as a separate layer, allowing compute to be started, stopped, resized or replicated without treating it as inseparable from one permanently running server.
Rank #2
| Conventional managed PostgreSQL pattern | Neon’s serverless model |
|---|---|
| Compute is usually provisioned for an instance | Compute can be allocated dynamically |
| Storage and compute are commonly tied together | Storage and compute are separated |
| Database process generally runs continuously | Inactive compute may scale down to zero |
| New environments are often created manually | Branches can provide disposable environments |
| Instance pricing is comparatively predictable | Usage-based billing varies with workload |
This can help preview deployments, development branches and applications with sharply varying traffic. It does not remove infrastructure complexity. Scheduling, startup latency, connection routing, replication, storage history and provider-specific limits still have to be managed.
What “serverless Postgres” means in practice
Serverless does not mean there are no servers. It means the customer does not directly provision the underlying database machines and that capacity can be allocated dynamically.
- Compute can scale with demand.
- Inactive databases may scale to zero.
- Billing is linked to usage rather than only to a permanently running instance.
- A database waking after inactivity can add latency.
- Serverless functions may create many short-lived connections, making pooling and connection-aware drivers important.
In 2023, Neon’s CEO told VentureBeat that cold-start time had been reduced from about three seconds to below 200 milliseconds. That was a company statement from that period, not a current service-level guarantee or universal result. Actual behavior depends on region, plan, workload and connection method. VentureBeat’s coverage provides the contemporary context.
Why investors connected Neon with AI
The AI connection is mainly about application data infrastructure. AI products still need a transactional system for users, permissions, billing and application state. They may also need to store embeddings, perform similarity searches and combine those searches with relational filters, joins and updates.
Four different technologies
- PostgreSQL as the system of record: structured, transactional application data.
- PostgreSQL with a vector extension: relational queries and vector retrieval in one database.
- A hosted PostgreSQL provider such as Neon: managed infrastructure and developer workflows around PostgreSQL.
- A dedicated vector database such as Pinecone: infrastructure optimized primarily for vector retrieval.
Neon’s 2023 announcement highlighted its pg_embedding extension and edge-aware driver. The proposition is that some teams can keep relational data and moderate vector workloads together instead of operating two stores. It does not make PostgreSQL a model-serving system, and it does not mean a vector extension will match every dedicated vector database.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Scan for outdated or missing drivers - takes under a minuteDriver Scan →What the investors were betting on
- PostgreSQL would remain a default open-source relational foundation for new applications.
- Developers would prefer managed infrastructure to operating database servers themselves.
- Serverless and edge applications would need autoscaling, low-latency connectivity and scale-to-zero behavior.
- AI applications would increase demand for embedding storage and vector search.
- Platforms such as Vercel and Replit could distribute Neon to large developer audiences.
Snowflake and Databricks participation may create strategic opportunities around data and AI infrastructure, but the investment alone does not prove a product integration or commercial partnership.
Current pricing and commercial context
Pricing context checked August 18, 2026: Neon’s pricing page lists a free plan and usage-based paid plans. Compute is billed in CU-hours; one compute unit is described as approximately one vCPU and 4 GB of RAM. Inactive compute can scale to zero on listed plans.
| Plan or item | Published detail | Important qualification |
|---|---|---|
| Free | $0; up to 100 projects, 100 CU-hours per month per project and 0.5 GB storage per project | Limits and features can change |
| Launch | Typical example of $15/month for intermittent load and 1 GB; $0.106 per CU-hour and $0.35 per GB-month | Example, not a universal bill |
| Scale | Typical example of $701/month for high load and 100 GB; $0.222 per CU-hour and $0.35 per GB-month | Example, not a universal bill |
Actual cost can also depend on compute size and runtime, storage, branches, history or restore-window settings, read replicas, network use and egress. Paid plans can bill extra branch-hours, so inactive branches should be cleaned up. Check the live Neon pricing page before committing.
When Neon is a good fit
- PostgreSQL-first applications with intermittent or highly variable traffic.
- Serverless APIs and edge-connected applications.
- Teams that need preview environments and disposable database branches.
- Early-stage products that want managed PostgreSQL without running database servers.
- AI applications needing relational data plus moderate vector search.
- Teams that value fast provisioning and workflow flexibility over a completely fixed monthly bill.
When another option may be better
- Always-on, high-utilization databases: scale-to-zero provides little benefit, while usage-based billing may be less predictable.
- Strict operational control: hosted services limit operating-system access, configuration, replication topology and maintenance choices.
- Specialized vector workloads: large indexes, demanding recall or latency targets may favor a dedicated vector database.
- Compliance or residency constraints: region, networking and procurement requirements may favor a major cloud’s native service.
- Unsupported extensions or settings: verify compatibility before migration.
- Very tight latency budgets: measure wake-up and scaling behavior in the intended region and plan.
Neon compared with common alternatives
Supabase
Supabase packages managed PostgreSQL with authentication, storage, APIs, realtime features and edge functions. Its pricing page lists compute beginning at $10 for a Micro instance, with paid plans receiving $10 per month in compute credits. Choose it when a broader backend platform is the priority; choose Neon when database branching, scale-to-zero and database-centric serverless workflows matter more. See Supabase pricing.
Recommended Free Tools
Pinecone
Pinecone is a dedicated vector database. Its listed plans include a free Starter tier, Builder at $20 per month, Standard with a $50 monthly minimum and Enterprise with a $500 monthly minimum. Those figures do not represent a complete application bill. Pinecone is a stronger candidate when vector retrieval is the primary workload and dedicated indexing and enterprise controls are required. See Pinecone pricing.
Amazon Aurora Serverless or RDS for PostgreSQL
AWS can be the better choice for organizations that need native IAM, networking, compliance, observability and procurement integration. Costs depend on region, capacity or instance selection, storage, I/O, backups and traffic, so a generic monthly comparison would be misleading. See Aurora and RDS for PostgreSQL.
Self-managed PostgreSQL
Self-management offers maximum control, but the team owns backups, upgrades, high availability, replication, monitoring, security patches, capacity planning, disaster recovery and connection management. It is sensible only when that operational responsibility is intentional.
Questions to answer before choosing Neon
- Is the database idle, bursty or continuously busy?
- What is first-query latency after inactivity?
- How many concurrent connections will serverless functions create?
- Will pooling or a serverless-compatible driver be used?
- Are every required extension and setting supported?
- What are the limits and charges for storage, compute, branches, replicas, history and egress?
- Are the region, compliance posture and private-networking options acceptable?
- Can the data be exported and restored elsewhere?
- Is vector search a secondary feature or the core database workload?
The practical verdict
The $46 million round reflected investor confidence in a developer-focused, cloud-native PostgreSQL model. Neon is most compelling when an application is PostgreSQL-first, bursty, serverless or preview-heavy and benefits from branches and managed operations. It is not automatically cheaper, faster or more capable than every provisioned database, backend platform or vector database. The correct decision depends on traffic shape, connection behavior, vector-search requirements, operational controls and tolerance for variable costs.
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Fix the driver behind crashes, sound loss and screen glitches3Clear out junk files and repair common Windows errorsQuick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
