Railway vs Render pricing

The two pages

Tier by tier

Rows are lined up by position on each page, not by tier name, because the names never match across two vendors. Row 1 is the top tier on each page, row 2 the one under it. Where one page has fewer tiers than the other, the row is left blank rather than paired with something it is not.

1. Row 1
   Railway

   FreeFree

   Render

   HobbyFree
2. Row 2
   Railway

   Hobby$5/mo

   Render

   Pro$25/mo
3. Row 3
   Railway

   Pro$20/mo

   Render

   Scale$499/mo
4. Row 4
   Railway

   EnterpriseCustom

   Render

   EnterpriseCustom

The entry tier

Changes we recorded

Every change we have written up for either page, newest first.

• Render simplifies sandbox pricing presentation and removes per-instance features, shifting to workspace-level concurrency limits

  Render·13 Nov 2026

  Render simplifies sandbox pricing presentation and removes per-instance features, shifting to workspace-level concurrency limits Why it matters: • Render is de-emphasizing isolated VM guarantees and indefinite runtime-suggesting sandbox positioning is moving away from long-running/stateful workloads toward ephemeral compute • The new concurrency limit model (1,000 default on Pro/Scale) introduces a new dimension of capacity gating that competitors should monitor for upsell/upgrade pressure patterns • Copy consolidation from 'microVMs' to 'VMs' and removal of '40 GB disk included' messaging simplifies positioning but may indicate less differentiation on resource guarantees Positioning: - Changed 'Isolated microVMs for agents, apps, and untrusted code' to 'Isolated VMs for agents' (narrower use case framing) - Removed 'Can run indefinitely' claim entirely - Removed '40 GB of disk included per sandbox' specification - Removed 'Pay only for the CPU you use' summary tagline Features & Limits: - Added new workspace-level limit: 'Default limit of 1,000 concurrent sandboxes on Pro, Scale workspaces' - Added 'Higher limits on request' with link to docs (introduces sales friction for scale) - Changed 'Automatic suspend and resume when idle' to 'Automatic suspend when idle' (removed resume mention) Pricing Presentation: - Reorganized CPU pricing table: separated 'active CPU hour' rate from memory provisioning rate into distinct columns for clarity - Memory pricing display changed from bundled format (e.g., '$0.038/hour for memory') to standalone column header 'provisioned memory' with same hourly rates - CPU rate remains $0.12/active CPU hour across all tiers (no change) - Memory rates unchanged: $0.038/hour (2GB), $0.076/hour (4GB), $0.152/hour (8GB), $0.304/hour (16GB) Navigation & UX: - Added jump links to pricing sections: workspace plans, compute pricing, FAQs (quality-of-life improvement) - Converted Spanish cookie/dialog copy to English (likely site-wide localization effort, not material to pricing strategy) What to do: • Audit your own sandbox/ephemeral compute messaging to ensure clarity on runtime guarantees and resource limits-Render's removal of 'indefinite' runtime suggests this is a differentiator if you support it • Monitor Render's sales patterns around the 1,000 concurrent sandbox cap over next 2-4 weeks; if they're converting upgrades at that threshold, consider your own concurrency/quota model • Review your memory pricing display for clarity-Render's table redesign suggests they heard friction on pricing comprehension; test your own pricing table UX with users

• Railway pricing page changed

  Railway·13 Nov 2026

• Render restructures database pricing with storage-inclusive tiers, introduces Sandboxes product, and significantly raises high-end instance pricing

  Render·8 Nov 2026

  Render restructures database pricing with storage-inclusive tiers, introduces Sandboxes product, and significantly raises high-end instance pricing Why it matters: • Render is repositioning database as a managed service with predictable storage costs (bundling 256 MB-50 GB storage per tier) rather than compute-only, making direct price comparison more complex and potentially advantageous for their mid-market positioning • New Sandboxes product (microVMs for agents/untrusted code at $0.12/CPU hour + memory) is a greenfield offering targeting AI agent and edge workload use cases-a new competitive surface outside their traditional developer infrastructure • Pricing floor moves up on database instances (Standard CPU $10→$10 at 256MB, but new $20 tier at 1GB replaces old $32 tier at 1GB), while max scale jumps from 40GB/10GB RAM to 50GB/32GB RAM-positioning them higher in the infrastructure stack Pricing & Plans: - Database instances now include storage bundled per tier (256 MB-50 GB storage depending on plan), displayed in new column - Standard CPU $10/month: RAM unchanged at 256 MB, but connections drop from 250 to 200; now includes 256 MB storage - Standard CPU $20/month: NEW tier, 1 GB RAM (replaces old $32/month tier which had 1 GB RAM) - Standard CPU $32/month → $60/month: RAM increases from 1 GB to 3 GB, storage 3 GB included, connections rise from 1,000 to 3,000 - Standard CPU $135/month → $100/month: RAM decreases from 5 GB to 5 GB (unchanged), but price drops; storage 5 GB included, connections drop from 5,000 to 5,000 - Standard CPU new $200/month tier: 10 GB RAM, 10 GB storage, 10,000 connections (replaces old $250/month High CPU tier which had 10 GB RAM) - High CPU $250/month → $350/month: RAM increases from 10 GB to 20 GB, storage 20 GB included, connections rise from 10,000 to 20,000 - High CPU $550/month → $600/month: RAM increases from 20 GB to 40 GB, storage 40 GB included, connections rise from 20,000 to 20,000 - High CPU new $750/month tier: 50 GB RAM, 50 GB storage, 50,000 connections (maximum scale cap increased from 40 GB to 50 GB) - Task compute: new $2.50/hour tier added for 8 CPU, 32 GB RAM - Database max scale increases from 40 GB RAM to 50 GB RAM; compute max scale increases from 4 CPU/16 GB RAM to 8 CPU/32 GB RAM New Product: - Sandboxes (beta): new isolated microVM offering for agents, apps, untrusted code - Sandboxes pricing: $0.12/active CPU hour + memory-based rate ($0.038-$0.304/hour depending on configuration); up to 8 CPU, 16 GB RAM per instance - Sandboxes features: indefinite runtime, automatic suspend/resume when idle, 40 GB disk included, pay-per-active-CPU model, snapshot storage at $0.08/GB per month - Sandboxes plan tiers include concurrent sandbox limits (100 default up to 5,000-10,000 on request) and task compute increments (10,000 CPU/40,000 GB RAM base) Features: - Database: 'Dedicated technical account manager' changed to 'Dedicated technical account team' - Database: 'Logical backup' changed to 'Logical exports' for paid instances - Database: new 'Connection pooling' feature added - Database support for 'Free(limitations apply)' tier now explicitly includes 50 connections (previously shown without connection count) Removed Content: - Migration credits offer ($10K cap) removed entirely - Migration credits application link removed - Direct navigation links to workspace plans, compute pricing, FAQs removed from main copy What to do: • Conduct detailed cost-comparison modeling across 5-50 GB RAM range: Render's storage bundling and tiered restructuring may make them appear cheaper on mid-market workloads; identify pricing positioning gaps for your own database product • Assess Sandboxes as a competitive threat in AI/agent infrastructure: their $0.12/CPU hour + memory model targets a different buyer persona (ML/automation platform builders) than traditional app hosting-evaluate whether to expand your own serverless compute or agent-adjacent offerings • Audit your own feature parity on managed database features (logical exports, connection pooling, dedicated account teams) and messaging around support tier upgrades, as Render is bundling these more aggressively into paid plans

• Railway introduces Sandboxes & VMs as a new product offering with per-second metering at $50/GB-month and $50/vCPU-month, removes Cloud Regions from Enterprise tier.

  Railway·8 Nov 2026

  Railway introduces Sandboxes & VMs as a new product offering with per-second metering at $50/GB-month and $50/vCPU-month, removes Cloud Regions from Enterprise tier. Why it matters: • Railway is expanding beyond traditional container deployment into infrastructure primitives (VMs/sandboxes), signaling a shift toward broader compute workloads and agent/code execution use cases that could capture workloads currently served by competitors like AWS Lambda or specialized VM providers. • The removal of Cloud Regions from Enterprise tier (now listed only as 'Higher Resource Limits') suggests Railway is either consolidating features or pricing strategy; competitors should clarify their own cloud region strategy and pricing transparency. • The sub-second metering model ($0.00001929 per GB/sec, $0.00001929 per vCPU/sec) and idle-only billing positioning is a direct cost-efficiency message; Railway is emphasizing pay-for-actual-use over wall-clock time, which could appeal to cost-conscious workloads. Product/Features: - New Sandboxes & VMs offering launched: isolated Linux VMs for agents, builds, and untrusted code execution. - Sandbox billing model: metered per second at $0.00001929 per GB/sec and $0.00001929 per vCPU/sec, approximately $50 per GB-month and $50 per vCPU-month for active use only. - Sandbox lifecycle features added: checkpoints, forks, templates, private networking, and public domains. - Idle timeout feature: configurable up to 5 minutes (default), up to 2 hours, or never, with automatic destruction of idle sandboxes. - Sandbox resource tiers: default 2 vCPU/2 GB, scaling to maximum 32 vCPU/32 GB per sandbox. Pricing: - Sandbox pricing: $50 per GB of memory per month and $50 per vCPU per month, metered per second, plus $0.05 per GB egress. - Pricing model emphasizes actual usage: customers pay only for CPU actually used and memory held while running, not for idle wall-clock time. Plans: - Enterprise tier: Cloud Regions removed from listed benefits; now shows 'Enterprise Support, Higher Resource Limits' plus prioritized support with SLOs, dedicated Slack Connect channel, and services with 48vCPU/48GB RAM per replica. Positioning: - Railway is positioning sandboxes as a cost-efficient alternative for waiting/idle workloads ('a sandbox waiting on a model or a user costs almost nothing'). - Emphasis on flexible, per-second metering and automatic cleanup (destroyed sandboxes cost nothing) versus traditional VM hourly billing. What to do: • Audit your own VM/sandbox offering (if any) and cost messaging: clarify whether you meter per-second or per-hour, and whether you charge for idle time. Railway's positioning on 'pay-only-for-active-use' is now a competitive liability if you don't match it. • Evaluate whether Railway's removal of Cloud Regions from Enterprise signals a pricing restructuring or de-emphasis; confirm your own multi-region story with customers and sales to prevent churn justifications. • Monitor Railway's Sandboxes adoption and feature velocity; if this becomes a revenue driver, assess whether your product roadmap should include similar compute-primitive offerings or deeper integrations with agent/LLM workflows that depend on untrusted code execution.

• Railway adds explicit usage-based pricing transparency and enterprise demo booking to pricing page

  Railway·3 Sept 2026

  Railway adds explicit usage-based pricing transparency and enterprise demo booking to pricing page Why it matters: • Railway is emphasizing pay-for-what-you-use cost predictability as a competitive differentiator, signaling confidence in their resource efficiency to undercut traditional flat-rate competitors • The addition of enterprise demo booking directly on the pricing page suggests they're optimizing conversion for high-value deals and moving away from pure self-serve positioning • New messaging around '0% idle cost' and 'only pay for active CPU/memory' indicates Railway is repositioning against competitors with wasteful resource billing or hidden charges Positioning: - Added explicit messaging: 'Only pay for active CPU and memory' (suggests focus on resource efficiency vs. competitors) - Added '0% idle cost' messaging (new value prop differentiator) Pricing: - Introduced or emphasized usage-based billing transparency with specific metrics (CPU/memory-based rather than generic 'compute credits') What to do: • Audit your own pricing model messaging-if usage-based, ensure it matches or exceeds Railway's transparency on what 'active' means and how costs scale • Review enterprise sales funnel: Railway's demo booking link suggests they're converting high-touch deals; ensure your enterprise onboarding process is equally frictionless • Test messaging that counters 'pay for idle time'-either highlight your own utilization efficiency or introduce usage-based tiers if you haven't already

• Render pricing page adds Spanish language support for EU/LATAM market expansion.

  Render·3 Sept 2026

  Render pricing page adds Spanish language support for EU/LATAM market expansion. Why it matters: • Indicates expansion beyond English-speaking markets, potentially signaling growth into Spanish-speaking regions (EU, LATAM). • May reflect increased international customer base or strategic localization push to compete globally. • No pricing or plan changes detected-purely a UX/localization enhancement. Localization: - Added Spanish translations for cookie consent dialog and privacy messaging. - Existing English content preserved; no removal of English language option. - No changes to pricing structure, plan names, features, or terms detected. What to do: • Monitor Render's localization roadmap-check if additional languages roll out to assess geographic expansion priorities. • Audit your own pricing page localization strategy; ensure parity or leadership in key markets Render is targeting. • Watch for any regional pricing adjustments or market-specific promotions that may follow this language addition.

• Railway pricing page changed

  Railway·1 Sept 2026

• Render pricing page changed

  Render·1 Sept 2026

• Render pricing page changed

  Render·29 Aug 2026

• Render pricing page changed

  Render·26 Aug 2026

• Render pricing page changed

  Render·23 Aug 2026

• Render introduces new pay-per-use Flex tier for Workflows and removes Beta label, repositioning around migration credibility while maintaining existing tiered pricing.

  Render·21 Aug 2026

  Render introduces new pay-per-use Flex tier for Workflows and removes Beta label, repositioning around migration credibility while maintaining existing tiered pricing. Why it matters: • Flex tier targets cost-sensitive or bursty workload users with transparent per-CPU and per-GB billing, potentially competing on unit economics for non-production or experimental tasks • Removal of 'Beta' from Workflows signals production readiness and confidence, likely accelerating adoption among enterprises evaluating infrastructure migration • Shift from 'switching clouds' to 'migrating production infrastructure' in credit messaging indicates Render is targeting higher-stakes, larger deal enterprise migrations rather than general cloud shopping Pricing & Billing Model: - New Flex tier added at $0.20/active CPU hour + $0.05/active GB hour (pay-per-use) vs. existing tiers which are provisioned capacity billing - Flex tier specs: up to 1 CPU, up to 4GB RAM, billed only for active compute time (new 'Billed for' column introduced across all tiers for clarity) - Existing Starter through Pro Max tiers unchanged in price ($0.05-$2.00/hour) but now explicitly labeled 'Provisioned capacity' billing Positioning: - Workflows promoted from Beta to GA (label removed) - Marketing copy evolved from 'Switching clouds?' to 'Migrating production infrastructure?' - narrower, higher-value segment focus - Workflows now explicitly claim 'Up to 10,000 CPUs for concurrent tasks' (previously 'Run 1000s of parallel, async tasks') - quantified ceiling signals scalability confidence Tier Removal: - Pro Ultra tier ($7.00/hour, 8 CPU, 32GB) removed from pricing table; all other existing tiers remain with identical pricing What to do: • Audit your own Workflows or task scheduling offering: assess whether a pay-per-use Flex option is missing from your pricing and whether it could unlock a cost-conscious segment Render now targets explicitly • Review messaging around production-grade readiness and migration support; if you lack a clear 'migration credits' or 'migration support' program, prioritize one to compete on enterprise land • Model the unit economics of Render's Flex pricing ($0.05/GB-hour, $0.20/CPU-hour) against your own offerings to ensure competitive positioning on bursty, non-production workloads

• Render updates privacy notice cookie disclosure with American English spelling

  Render·18 Aug 2026

  Render updates privacy notice cookie disclosure with American English spelling Why it matters: • Cosmetic localization change only-no material impact to pricing, features, or positioning • Reflects minor editorial refresh; no strategic or competitive implications Positioning: - Changed 'utilises' to 'utilizes' (British to American English) - Changed 'personalisation' to 'personalization' (British to American English) - Privacy policy link and functionality remain unchanged What to do: • No action required-this is a cosmetic copy update with zero competitive impact

• Render updates cookie notice with British English spelling (minor localization adjustment).

  Render·15 Aug 2026

  Render updates cookie notice with British English spelling (minor localization adjustment). Why it matters: • Indicates possible UK/EU market localization efforts, though no pricing or feature changes present. • No competitive threat or strategic shift-purely a terminology/regional language update. Messaging: - Changed 'utilizes' to 'utilises' (US to British English) - Changed 'personalization' to 'personalisation' (US to British English) - Cookie notice text remains functionally identical in meaning and transparency What to do: • Monitor for follow-up localization changes (currency, regional pricing tiers, compliance messaging) that might signal Render's expansion into UK/EU markets. • No immediate action required unless Render introduces region-specific pricing or compliance changes.