Vercel Pricing (2026): The $20 Plan Is Only the Floor

Vercel Pro is $20 a month, but seats, usage, add-ons, v0, and AI costs stack separately. See every live rate and break-even point.

Published Updated

Tools
Vercel Pricing (2026): The $20 Plan Is Only the Floor

Vercel Pro costs $20 a month for one deploying developer, but that is the floor, not the ceiling. The same team can also pay for extra developer seats, a paid CDN capacity tier, metered infrastructure, fixed add-ons, v0, and AI usage, while the one $20 infrastructure credit does not multiply with headcount.

Prices and limits were re-verified against Vercel's live pricing page, its documentation, and the other vendors' linked pricing pages on October 17, 2026. Prices are in USD before applicable taxes. Rates marked "starting at" vary by region, so treat them as the lowest published rate rather than a universal quote.

Vercel live pricing page showing Hobby, Pro, and Enterprise plans
Vercel pricing page, September capture. The table below carries the current rates.

Vercel pricing at a glance

Vercel has three hosting plans: Hobby at $0, Pro at $20 per month for the first deploying developer, and Enterprise at a custom contract price.

PlanCurrent priceIncluded core valueThe line that matters
Hobby$0/month200 projects, 100 GB transfer, 1M CDN requests, capped compute and storagePersonal, non-commercial use only; no paid overages
Pro$20/month1 deploying seat, $20 infrastructure credit, entry Flat Rate CDN tier: 1M requests and 1 TBExtra deployers cost $20/month; higher CDN tiers and usage can add to the bill
EnterpriseCustomPro features plus advanced controls, multi-region failover, support, and 99.99% SLAPricing, usage, and support terms are negotiated

The choice is simpler than the rate sheet makes it look. Keep Hobby for a personal, non-commercial project that stays inside its hard caps. Move to Pro when the project is commercial or needs paid overages and collaboration. Talk to Enterprise when a contract, 99.99% SLA, managed security controls, directory sync, or multi-region failover is a procurement requirement.

Decision flow routing personal projects to Hobby, commercial apps to Pro, and SLA requirements to Enterprise
The plan decision is driven by use rights first, then operational requirements.

What Vercel Pro really costs

Vercel Pro costs $20 for the platform, not $20 for an all-inclusive bundle. That fee includes one deploying developer seat and one $20 monthly credit for managed infrastructure across the entire team.

The credit belongs to the Pro plan, not to each paid seat. A three-developer team pays $60 a month but still receives one $20 infrastructure credit. A five-developer team pays $100 and still receives $20. Unlimited Viewer seats are free, but Viewers can inspect deployments and analytics and comment on previews; anyone who configures or deploys needs an Owner or Member seat at $20 a month. Pro also includes one free Billing seat for managing invoices and payment methods, with read-only access to projects.

The practical bill formula is:

$20 platform fee + $20 for each additional deploying seat + any paid Flat Rate CDN tier + fixed add-ons + max($0, eligible metered usage minus the $20 team credit).

AI Gateway credits, v0, marketplace charges, domains, and taxes need their own budget lines.

The Pro credit offsets managed infrastructure usage from the first billable unit. It does not pay the fixed subscription for a higher CDN tier. Unused credit expires monthly and does not roll over.

Since September 8, 2026, Flat Rate CDN has been generally available, enabled by default for new Pro teams; existing teams can opt in. The included capacity is 1 million CDN requests and 1 TB per month, replacing the old default-plan calculation based on 10 million included requests. Daemons joined the pricing page with its October 15, 2026 public beta, adding CPU and memory costs for background workers.

Flat Rate CDN charges for a capacity tier shared across the team's projects:

Monthly CDN fee, on top of ProCDN request capacityData transfer capacity
Included1 million1 TB
$2010 million50 TB
$10050 million50 TB
$300150 million50 TB

The current CDN terms cover CDN Requests, Fast Data Transfer, Blob Data Transfer, Sandbox Data Transfer, and Observability events generated by CDN requests. A one-day spike does not trigger a tier upgrade. Sustained usage can move the team to a higher tier next cycle, and exceeding the highest tier can move it to on-demand pricing. Service terms require at least seven days' notice of a higher tier fee; tiers do not downgrade automatically.

This is capacity for normal web-app delivery. Bulk downloads, file hosting, and workloads dominated by media or file delivery are outside its eligibility rules. Materially excessive or ineligible use can lead to Flex CDN, a different plan, or service limits. The Flat Rate CDN guide covers the settings and traffic-policy details.

Headcount changes the value quickly:

  • One deploying developer pays $240 a year and has $20 of monthly infrastructure credit per developer.
  • Three deploying developers pay $720 a year, while the credit works out to $6.67 per developer.
  • Five deploying developers pay $1,200 a year, while the credit works out to $4 per developer.
  • Ten deploying developers pay $2,400 a year, while the credit works out to $2 per developer.

That is the first break-even rule: Vercel's seat price matches Netlify's $20 entry Pro team plan at one deployer and loses on base price from the second deployer onward. The platform can still justify the difference through included traffic or workflow fit, but the extra seat must earn its $240 annual cost.

Is Vercel free? The Hobby plan truth

Vercel Hobby is enough indefinitely for a personal, non-commercial project that stays below its monthly and operational limits. A portfolio, personal blog, learning project, or open experiment can remain at $0 if it does not become commercial and does not need more capacity.

The live plan includes 1 million CDN Requests and 100 GB of Fast Data Transfer each month. Functions include 4 Active CPU hours, 360 GB-hours of provisioned memory, and 1 million invocations. Storage and optimization include 1 GB of Blob storage, 10,000 simple Blob operations, 2,000 advanced Blob operations, 10 GB of Blob transfer, 5,000 image transformations, 300,000 image cache reads, and 100,000 image cache writes.

Hobby also includes 1 million ISR reads, 200,000 ISR writes, 100,000 Global Config reads, 100 Global Config writes, 50,000 Workflow events, 1 million Queue operations, 500 Vercel Connect token requests, 1,000 Vercel Connect triggers, 10,000 Vercel Flags requests, and 50,000 Web Analytics events per month. Free Speed Insights includes 10,000 events across the team over a rolling 30-day window. The operational limits include 200 projects, 100 deployments per day, one concurrent build, a 45-minute build ceiling, and 100 MB of source files per CLI deployment.

Those are hard allowances, not a small pay-as-you-go account. Hobby users cannot buy more usage. A project pauses or hits the relevant limit instead of converting itself into an overage invoice.

The decisive limit is not technical. Vercel's pricing terms reserve Hobby for personal, non-commercial use. A tiny company site with almost no traffic belongs on Pro because it is commercial. A busy personal project can remain on Hobby only while it stays inside the caps. Who never needs to pay: the owner of a personal, non-commercial project whose traffic, compute, storage, deployments, and collaboration needs stay within those published limits.

The non-price limits that can force an upgrade

Security rules, deployment controls, concurrency, and retention can force a higher tier before traffic does. Hobby allows up to 3 custom firewall rules and 3 IP blocks. Pro raises those to 40 and 100, while Enterprise raises both to 1,000. System bypass rules move from none on Hobby to 25 on Pro and 100 on Enterprise.

Sandbox concurrency is 10 on Hobby, 10,000 on Pro, and custom on Enterprise. Custom environments move from none on Hobby to one on Pro and 12 on Enterprise. Rolling Releases move from none to one included Pro project and custom Enterprise terms.

The collaboration limits follow the same pattern. Hobby has one developer seat. Pro includes one paid deployer, charges $20 a month for each additional deployer, adds unlimited free Viewers, and includes one free Billing seat. Enterprise negotiates the seat structure.

Retention is the quiet operational limit. Runtime Logs and Session Traces last one hour on Hobby, one day on Pro, and three days on Enterprise. Observability Plus can extend paid-plan Runtime Logs to 30 days. If an incident investigation regularly begins days after the event, retention can decide the plan before compute does. Preview retention is a separate constraint: see why Hobby previews can disappear before their expected retention date.

Every current Vercel Pro overage and add-on

Vercel Pro turns almost every resource into a named meter. The following rates come from the current pricing matrix; product-specific additions are linked separately. The $20 credit offsets eligible usage, while fixed subscriptions remain separate.

Delivery, firewall, and caching

Delivery now belongs to the selected Flat Rate CDN tier when it is enabled. Do not add the old per-request and per-GB charges on top of covered traffic. If you opt out, on-demand pricing starts at $2 per million CDN Requests and $0.15 per GB of Fast Data Transfer, with regional differences; check your team's billing mode before estimating that alternative.

Caching can avoid a function execution, but the request and transferred bytes still count toward CDN capacity. Fast Origin Transfer, the traffic between Vercel's CDN and compute, remains a separate meter.

Firewall rate limiting starts at $0.50 per million allowed requests. BotID Deep Analysis costs $1 per 1,000 checks. These remain separate security meters; Flat Rate CDN does not make deeper traffic analysis free.

Incremental Static Regeneration, which rebuilds cached pages after deployment, bills reads at starting at $0.40 per million and writes at starting at $4 per million. A content site with frequent revalidation can have modest transfer and still create a noticeable write meter.

Blob, images, configuration, and microfrontends

Vercel Blob charges $0.023 per GB-month for storage, $0.40 per million simple operations, and $5 per million advanced operations, while Blob Data Transfer is included in Flat Rate CDN when enabled. Uploading a file can still create storage and operation charges.

Image Optimization costs starting at $0.05 per 1,000 transformations, $0.40 per million cache reads, and $4 per million cache writes. A catalog that generates many sizes and formats may spend more on transformations and writes than a page-count estimate suggests.

Global Config costs starting at $3 per million reads and $1 per 100 writes. Microfrontends include two projects, then cost $250 per month for each additional project; routing costs $2 per million requests. That additional-project fee is a fixed commitment, so the $20 infrastructure credit does not soften it.

Functions, sandboxes, workflows, Daemons, and agent infrastructure

Vercel Functions bill starting at $0.128 per Active CPU hour, $0.0106 per provisioned-memory GB-hour, and $0.60 per million invocations. Active CPU means code execution time is metered separately from the memory held while the function handles work. A slow external API call may use little CPU but still accrue memory duration. Concurrent requests can share an instance, so estimate billed instance GB-hours rather than blindly multiplying every request by its wall time.

Vercel Services, currently beta, starts at $0.50 per million requests. Vercel Sandbox starts at $0.128 per Active CPU hour, $0.0212 per provisioned-memory GB-hour, $0.60 per million sandbox creations, and $0.08 per GB-month of snapshot storage; Sandbox Data Transfer is included in Flat Rate CDN. The Sandbox rate sheet also lists Drive storage at $0.05 per GB-month, reads at $0.0015 per GB, and writes at $0.004 per GB in its default Washington, D.C. region. Container Registry image storage is $0.10 per GB-month, while Fast Origin Transfer starts at $0.06 per GB.

Daemons, background processes that pull work and stay running between jobs, add starting at $0.128 per Active CPU hour and $0.0212 per provisioned-memory GB-hour. Hobby lists 5 CPU hours and 420 GB-hours per month; Pro meters those resources against its shared infrastructure credit. A worker can accumulate memory-hours even when its CPU demand is modest.

Workflows cost $20 per million events, $0.50 per GB written, and $0.50 per GB-month retained. Queues start at $0.60 per million API operations. Vercel Connect costs $3 per 1,000 token requests and $0.95 per 1,000 triggers. Those are requests for provider credentials and forwarded webhook events, not AI inference tokens, even though the pricing page places these rows in its AI section. These meters matter for agents because one user outcome may span many workflow events, queue operations, tool connections, sandbox starts, and model calls.

Builds, collaboration, observability, and security

Builds cost $0.007 per minute on Basic machines, $0.014 per minute on Standard machines, $0.028 per minute on Enhanced, $0.105 per minute on Turbo, or $0.0035 per CPU-minute on Elastic. A Preview Deployment Suffix costs $100 per month. Each additional deploying seat costs $20 per month.

Vercel Flags includes 150 Explorer overrides a month, then charges $250 for unlimited overrides; flag traffic costs $0.03 per 1,000 requests. Observability Plus is $1.20 per million events. Free Speed Insights includes 10,000 events per team over a rolling 30 days. Speed Insights Plus is $10 per project per month plus $0.65 per 10,000 events, with no included Plus event allowance on Pro. Web Analytics is $3 per 100,000 events, Web Analytics Plus is $10 per month, Tracing Spans beta costs $0.50 per million span units, and Drains Volume costs $0.50 per GB. CDN-request Observability events are covered by Flat Rate CDN; other billable Observability events still need a budget.

The large fixed add-ons are easy to miss because they do not look like infrastructure usage. SAML Single Sign-On costs $300 per month, a HIPAA BAA costs $350 per month, Password Protection costs $20 per protected project per month, and shared-pool Static IPs cost $100 per project per month plus starting at $0.15 per GB. Bulk redirects include 1,000 per project, then cost $50 per month for each additional 25,000. Pro includes one custom environment per project; packs of five more cost $50 per month, up to 16 per project.

AWS PrivateLink is available to Pro and Enterprise teams through Advanced Networking. The first connection is included with Advanced Networking, each additional connection costs $30 per month, and transfer costs $0.04 per GB. It gives Vercel Functions and builds a private route to supported AWS-hosted services; Vercel's launch page and docs do not publish a separate base price for Advanced Networking, so confirm that price in the dashboard before treating the first connection as free. The PrivateLink explainer covers the network and region limits.

Vercel cost calculations at working scale

Vercel is inexpensive when the chosen CDN tier covers normal traffic and the credit absorbs the remaining infrastructure usage. It gets expensive when one user outcome fans out across several billable meters or when headcount grows faster than traffic. These are modeled monthly bills, not measured production invoices.

Cost per million dynamic requests

Take 1 million function invocations, each using 100 milliseconds of Active CPU, with 27.7778 total billed GB-hours of memory. That is the memory usage of a 1 GB instance handling the requests sequentially for the same duration, with no extra waiting time. The workload uses 27.7778 CPU-hours. At the Washington, D.C. rates, CPU costs $3.56, memory costs about $0.29, and invocations cost $0.60, for $4.45 per million requests before the monthly credit.

That is a clean baseline, not a universal function price. Longer wall time raises memory cost. More CPU raises CPU cost. Each response also counts toward CDN requests and transfer capacity. Image work, storage, queues, workflows, or AI tokens add their own meters.

CDN-request cost at 10M, 50M, and 100M

For one deploying developer, assume eligible traffic, no other billable resources, and no more than 50 TB of transfer. At sustained 10 million requests, the $20 Pro base plus the $20 CDN tier totals $40 per month. At 50 million, the $100 CDN tier makes the total $120. At 100 million, the 150-million-request tier costs $300, making the total $320. These totals use the published capacity tiers, not the old request-overage formula.

The $20 infrastructure credit cannot cancel a fixed CDN subscription. A temporary spike can leave the current month's tier unchanged; these examples price capacity for a normal month at the stated volume, rather than assuming every spike causes an immediate upgrade.

Transfer cost at 2 TB and 5 TB

For an eligible web app with no more than 10 million CDN requests, both 2 TB and 5 TB fit the $20 CDN tier's 50 TB capacity. The one-developer total is $40 per month in either case, before compute and other usage. At no more than 1 million requests and 1 TB, the included CDN tier keeps that subtotal at $20.

Transfer volume alone no longer determines the tier: an app with small responses can cross the request capacity first. A bulk-download service also cannot assume these web-app rates apply to its traffic.

A worked bill with Functions, Sandbox, Workflows, and Daemons

Take one developer on the $20 CDN tier, with 10 million CDN requests and 2 TB of eligible transfer. Suppose its separately metered work in the month is:

  • Functions: the 1 million calls above cost $4.45.
  • Sandbox: 10 CPU-hours, 100 billed GB-hours, and 10,000 creations cost $3.406 at the listed starting rates.
  • Workflows: 100,000 events, 1 GB of writes, and 1 GB-month retained cost $3.00.
  • Daemons: 10 CPU-hours and 720 billed GB-hours cost $16.544 at the listed starting rates.

The metered subtotal is $27.40. The shared $20 credit leaves $7.40, so the modeled subtotal is $20 Pro + $20 CDN + $7.40 usage = $47.40. With three deploying developers, it becomes $87.40, because the additional seats do not add credit.

This is a subtotal for those named resources. Add actual builds, origin transfer, container images, storage operations, observability, fixed add-ons, and any other resources used. AI Gateway model usage is charged to its separate credit balance; Pro infrastructure credit cannot cover it.

Team-seat break-even

The seat calculation runs the opposite way. One deployer is $20, five deployers are $100, and ten are $200 before usage. If most collaborators only need to inspect a preview, keep them as free Viewers. Paying for a developer seat that rarely deploys turns an occasional action into a recurring $240 annual cost.

v0, Vercel Agent, and AI Gateway are separate bills

v0 is not included in the $20 Vercel Pro platform fee. A builder paying for both Vercel hosting and v0 can have two subscriptions plus token usage before the deployed app serves its first customer.

v0 live pricing page with Free, Plus, Business, and Enterprise plans
v0 pricing is separate from Vercel hosting

v0 Free is $0 with $5 of one-time credit for new users and a seven-message daily limit. v0 Plus is $30 per user per month, with $30 of monthly credit and $2 of daily login credit per user. Business is $100 per user per month, includes the same $30 monthly credit and $2 daily login credit, and opts data out of training by default. Enterprise is custom.

The current v0 models range from $0.20 to $10 per million input tokens and $1.20 to $50 per million output tokens, depending on whether you choose Mini, Pro, Max, or Max Fast. If you use the v0 API, budget its generation usage separately from the app's hosting costs.

For one person, Vercel Pro plus v0 Plus starts at $50 per month. Pro plus v0 Business starts at $120 per month. Both totals exclude additional infrastructure and token usage.

Vercel Agent adds another meter at $0.25 per million tokens plus the pass-through token cost. AI Gateway has its own credit balance: its free tier includes $5 per month for eligible models at provider list prices with zero markup. Purchasing credits moves the team to the paid tier and ends that monthly free credit. Paid accounts can bring provider keys without a Gateway markup; model choice and input/output token counts determine inference cost.

Optional AI Gateway controls have their own meters:

  • Custom Reporting: $0.075 per 1,000 tag, user-ID, or quota-entity writes, plus $5 per 1,000 reporting queries.
  • Team-wide provider allowlist: $0.10 per 1,000 successful requests.
  • Team-wide zero data retention: $0.10 per 1,000 requests.
  • Trace Drains: $0.05 per 1,000 traces plus $0.50 per GB of trace egress.

Reporting, allowlist, and retention surcharges draw from AI Gateway credits. Trace Drains bill through the hosting plan's Drains usage, with no included Pro allowance for those meters. For budget controls rather than another pricing ladder, use the AI Gateway per-user spend guide.

Token Pass, in early access for Pro teams, costs $30 per month for $80 of monthly credit on its selected models.

The decision rule is simple: price hosting, code generation, production-agent assistance, and model inference as separate products. The shared Vercel brand does not turn them into one allowance.

Vercel alternatives, normalized to the same outcomes

Netlify wins when deploying headcount dominates, Cloudflare wins when bandwidth and request volume dominate, and AWS Amplify can win at low usage when a team prefers pure consumption pricing. None is uniformly cheaper.

Netlify: cheaper teams, weaker bandwidth allowance

Netlify Pro starts at $20 a month with unlimited members and 3,000 credits, so it beats Vercel's base seat price from the second deploying developer onward.

Netlify live pricing page with credit-based plans
Netlify pricing

Netlify spends 20 credits per GB of bandwidth, 10 credits per GB-hour of compute, 2 credits per 10,000 web requests, and 15 credits per production deploy. Pro auto-recharge, disabled by default, costs $10 for 1,500 credits.

Using Netlify's entry Pro tier and recharge packs, if every credit goes only to bandwidth, 1 TB costs $140 with 12 whole recharge packs; 2 TB costs $280 with 26 packs. This uses 1 TB = 1,024 GB and does not claim recharge is cheaper than every higher Netlify tier. For eligible delivery, Vercel costs $20 at 1 TB with at most 1 million requests, or $40 at 2 TB with at most 10 million requests, before other usage. Netlify therefore wins a five-developer, low-traffic team at $20 versus Vercel's $100 base, while Vercel wins a one-developer, bandwidth-heavy app. The broader product differences are covered in Vercel vs Netlify.

Cloudflare: the traffic-cost winner

Cloudflare Workers Paid starts at $5 per account per month, includes 10 million requests and 30 million CPU milliseconds, then charges $0.30 per million requests and $0.02 per million CPU milliseconds. It has no data-transfer or bandwidth charge, and static-asset requests are free and unlimited.

Cloudflare Workers pricing documentation
Cloudflare Workers pricing

Cloudflare's own 100-million-request example at 7 milliseconds of CPU per request totals $45.40. For sustained traffic at that request count, Vercel's one-seat base plus the appropriate Flat Rate CDN tier totals $320, with up to 50 TB of eligible transfer but before function compute and other meters. This is a comparison of the bill components, not an identical-runtime benchmark. That makes Cloudflare the clear cost choice for request-heavy edge code when its runtime and operational model fit. The tradeoffs, including its platform limits and split product billing, are detailed in the Cloudflare review.

AWS Amplify: low-usage pay-as-you-go

AWS Amplify charges no seat fee. Outside its Free Tier, Standard builds cost $0.01 per minute, storage costs $0.023 per GB-month, transfer costs $0.15 per GB, server-side-rendered requests cost $0.30 per million, and SSR duration costs $0.20 per GB-hour.

AWS Amplify live pricing page
AWS Amplify pricing

Looking only at transfer, AWS Amplify breaks even with Vercel's $20 one-seat base at about 133 GB per month. Below that, AWS transfer costs less than $20. Above it and up to Vercel's included 1 TB, Vercel is cheaper if requests stay within the included 1 million. At 1 TB, AWS transfer alone is $153.60 versus Vercel's $20 under that assumption; at 2 TB it is $307.20 versus Vercel's $40 with at most 10 million requests. The Vercel figures assume eligible web-app delivery; both sides exclude other resources.

AWS can still win when several developers would create Vercel seat charges or when the workload uses little transfer. Add the rest of the stack before deciding: Amplify builds, storage, SSR, and a $15 per-app WAF fee plus WAF usage can move the result.

The September comparison below used the previous request-overage calculation. Its Vercel column counted requests only: the displayed 7 ms CPU label was not included in that $180 amount. Use the current $320 CDN-and-seat subtotal above, then add compute, when estimating Flat Rate CDN today.

Historical September comparison showing the superseded Vercel request-only subtotal and Cloudflare's CPU-inclusive example
Historical pricing illustration: Vercel's $180 excluded compute despite the 7 ms label. Current Flat Rate CDN plus one seat at sustained 100M requests is $320 before compute.

Trial, billing cadence, refunds, and price history

Vercel Pro is billed monthly. The live self-serve pricing and Pro documentation do not publish an annual Pro discount, so there is no annual core-plan commitment to justify with a discount. Enterprise terms are negotiated separately.

Domains are the annual renewal to watch. Vercel may charge a domain up to 30 days before its purchase anniversary. Each eligible paid Pro team can claim one eligible first-year domain; Pro trials do not qualify, and the domain renews later at the standard rate.

The Pro trial lasts 14 days or until a trial usage cap is reached, whichever comes first. The trial includes a $20 credit. Selected caps include one owner, 10 total members, 200 projects, 16 CPU-hours, 1,440 GB-hours of provisioned memory, 1 million function invocations, and 10,000 image transformations. You get one trial per user account. Without a payment method, the team reverts to Hobby and is not charged.

Refund terms are strict. Vercel's terms say fees are non-refundable unless the agreement expressly says otherwise. A cancellation or downgrade takes effect at the next renewal period. If Vercel terminates a self-service subscription without cause, the terms provide a refund of the pro-rated, unearned prepaid amount.

Vercel's current self-serve hosting price page lists no student-specific Pro rate. Budget the published price unless Vercel has confirmed a separate offer for your account.

The price structure has changed in ways that explain conflicting numbers elsewhere. On April 4, 2024, Vercel announced granular infrastructure meters and a Fast Data Transfer rate reduction from $0.40 to $0.15 per GB. On September 9, 2025, Pro moved to the flexible $20 team credit, free Viewer seats, self-serve enterprise add-ons, and default spend alerts. Flat Rate CDN is the current delivery model for new Pro teams; legacy on-demand calculations should not be mixed into its tier fees.

Who should pay for Vercel, and who should skip it

Pay for Vercel Pro when the app is commercial, one or two people deploy it, its traffic fits an eligible CDN tier, and the delivery capacity and deployment workflow justify the seat fee. Configure Spend Management before a public launch, keep non-deploying collaborators on free Viewer seats, and price every fixed add-on separately.

Stay on Hobby when the work is genuinely personal and non-commercial and every cap is comfortable. Do not use Hobby as a permanent production tier for a business simply because the traffic is small.

Skip Vercel for Netlify when a larger team needs deployment access but usage is modest. Skip it for Cloudflare when edge requests or bandwidth dominate and the Workers model fits. Consider AWS Amplify when usage is low, the team is already AWS-native, and a no-seat consumption bill matters more than an integrated allowance.

The upside
What it does well
4 points

  • One deploying developer, the entry CDN tier with 1 TB and 1 million requests, and $20 of flexible infrastructure credit fit inside the $20 Pro base.
  • Unlimited Viewer seats avoid charging for people who only inspect previews and analytics.
  • Hobby has useful hard caps and cannot generate paid overages.
  • The live page exposes detailed per-resource rates, making a modeled bill possible.
The downside
Where it falls short
5 points

  • Hobby is restricted to personal, non-commercial use.
  • Each additional deploying developer adds $20 a month without adding another $20 infrastructure credit.
  • Unused credit expires monthly, while usage can spill into many separate meters.
  • Paid CDN tiers, fixed add-ons, v0, and AI usage sit outside the simple $20 headline.
  • Regional "starting at" rates mean a global workload needs a region-aware estimate.

Vercel pricing FAQ

Is Vercel 100% free?

Vercel Hobby costs $0 and cannot incur paid overages, but it is limited to personal, non-commercial use. It includes 100 GB of Fast Data Transfer, 1 million CDN Requests, capped functions and storage, 200 projects, and 100 deployments per day.

How much does Vercel cost?

Hobby is $0. Pro starts at $20 per month with one deploying seat, one $20 team infrastructure credit, and the included Flat Rate CDN tier with 1 million requests and 1 TB. Additional deploying seats cost $20 each; paid CDN tiers and other usage can add to the bill. Enterprise is custom.

What is Flat Rate CDN?

It is Pro's tier-based CDN billing option, with temporary spike protection. The entry tier is included; higher tiers add $20, $100, or $300 per month. Sustained usage can change the next cycle's tier, and eligibility restrictions apply. It does not cover Functions, Daemons, or AI model usage.

Can I buy additional usage?

Yes on Pro: eligible infrastructure usage beyond the $20 monthly credit is billed on demand. Hobby cannot purchase overages. Fixed CDN tiers, seats, and add-ons are separate from that credit.

Does Vercel offer a student discount?

The current self-serve hosting price page lists no student-specific Pro rate. Use the published Pro price unless Vercel confirms another offer for your account.

Does Vercel offer refunds?

Vercel's terms state that fees are non-refundable unless expressly stated otherwise. Cancellation takes effect at the next renewal period; a limited pro-rated refund applies if Vercel terminates a self-service subscription without cause.

Can I pay for Vercel Pro annually?

The live self-serve Pro pricing is monthly, and Vercel does not publish an annual Pro discount. Enterprise contracts are negotiated; domains renew on their own annual schedule.

Has Vercel pricing changed?

Yes. Pro now lists Flat Rate CDN for covered delivery traffic, and the pricing page adds Daemons CPU and memory meters. The full page was re-verified on October 17, 2026; the older blanket 10-million-request Pro allowance is not the default Flat Rate CDN tier.

Last Updated
Category
Build
Related Articles
Vision Language Models: What Builders Can Trust

Vision Language Models: What Builders Can Trust

What vision language models can read, where they fail, and how to choose API or self-hosted models with a test plan for your own images.Oct 26, 2026Build
Multi Agent Systems: When One Agent Beats a Team

Multi Agent Systems: When One Agent Beats a Team

When should you split one LLM agent into several? Compare supervisors, hierarchies, handoffs and group chat, with a task-based decision table.Oct 26, 2026Build
Prompt Injection: Direct and Indirect Attacks in AI Agents

Prompt Injection: Direct and Indirect Attacks in AI Agents

Understand direct and indirect prompt injection, how agents expose business data, and the safeguards to use before connecting an inbox or CRM.Oct 26, 2026Build
GraphRAG: How It Works and When to Use It

GraphRAG: How It Works and When to Use It

How Microsoft GraphRAG answers questions across documents, what drives indexing work, and when to use it instead of vector, hybrid or agentic RAG.Oct 26, 2026Build
AI Agent Architecture: The Production Map

AI Agent Architecture: The Production Map

Map the model, tools, memory, orchestration, approvals and evaluation behind a production AI agent, with a support workflow and buy-versus-build decisions.Oct 25, 2026Build
Cursor Models: What You Can Use After the October Change

Cursor Models: What You Can Use After the October Change

Which Cursor models you can use after October, how Auto routes tasks, and how to choose across Grok, Composer, Claude, GPT and Gemini.Oct 25, 2026Build
LLM Parameters: Model Weights and Generation Settings

LLM Parameters: Model Weights and Generation Settings

Understand model size, active MoE weights and API controls, with provider-specific rules and starting settings for extraction, drafting and code.Oct 25, 2026Build
Continuous Batching: How It Improves LLM Throughput

Continuous Batching: How It Improves LLM Throughput

How continuous batching improves shared LLM throughput, changes latency, and works with KV cache in vLLM, SGLang, TensorRT-LLM and TGI.Oct 25, 2026Build
Newsletter

One letter, every Sunday.Working systems, not hot takes.

Weekly. No spam. Unsubscribe anytime.