Mux Costs $9,500/Month. Cloudflare Stream Costs $625 For The Same Workload. So Why Are Most Teams On Mux?
A growth-stage learning platform we worked with in early 2026 was paying $11,400/month for Mux to host their 800-hour course library and serve roughly 12,000 hours/month to learners. Their video bill was the third-largest engineering line item. The CEO had been told Mux was "the standard" by their CTO who had used it at a previous OTT startup. Nobody had benchmarked alternatives.
We modeled their workload across four platforms:
- Mux (current): $11,400/month
- AWS MediaConvert + CloudFront: $4,800/month (58% savings)
- Self-hosted FFmpeg + Bunny CDN: $1,200/month + 0.5 FTE engineer (~$10,000 fully loaded)
- Cloudflare Stream: $760/month (93% savings)
The platform migrated to Cloudflare Stream over 6 weeks. Their monthly bill dropped to $760/month. Annual savings: $127,680. Playback quality was indistinguishable to their learners. The premium analytics features they "needed" turned out to be features they had never actually used.
This pattern is consistent across 14 video streaming audits we ran in 2025-2026: teams pick a video platform once, usually based on what was popular when they started, and never benchmark alternatives. The wrong platform costs $50K-$500K/year for mid-volume deployments.
The reason for the pricing spread is structural. Mux pioneered the developer-friendly video API in 2018 and won mindshare among engineering-led products. Cloudflare entered the video market in 2021 and built it on top of their existing network advantage (zero egress, global POPs). AWS has had MediaConvert + CloudFront for years but the assembled-from-parts experience has not improved. Self-hosted FFmpeg + CDN has gotten dramatically easier with modern containers and managed CDN options.
This post is the workload-to-platform decision framework: which video platform wins for which use case, what each one actually costs at three scales, and the migration playbook.
The Four Platforms That Actually Matter in 2026
| Platform | Pricing Model | Sweet Spot | Operational Burden |
|---|---|---|---|
| Cloudflare Stream | $5/1000 min stored + $1/1000 min delivered | Most VOD use cases | Very low |
| Mux | $0.005/min encoded + $0.00096/min delivered | Premium video products | Low |
| AWS MediaConvert + CloudFront | $0.0075-0.0405/min encoded + $85/TB egress | AWS-native stacks, moderate volume | Medium |
| Self-Hosted (FFmpeg + Bunny CDN) | EC2 + $5/TB egress on Bunny | Extreme scale, custom needs | High |
The 9x cost spread across these platforms is real. Picking by familiarity (most teams default to AWS or whoever they used at their last job) leaves enormous savings on the table.
The Real 2026 Pricing (Detailed)
Cloudflare Stream
- Storage: $5 per 1,000 minutes stored per month
- Delivery: $1 per 1,000 minutes delivered
- Encoding: Free (included in storage cost)
- Adaptive bitrate: Automatic, no extra charge
- Built-in player: Free with the service
- Live streaming: $1 per 1,000 minutes delivered + same storage cost
- Free tier: None (small workloads still get the same low rates)
Mux
- Encoding: $0.005 per minute uploaded (one-time)
- Storage: $0.003 per minute stored per month
- Delivery: $0.00096 per minute delivered (~$0.96/1000 min)
- Live streaming: $0.04 per minute streamed in
- Mux Data analytics: $0.10 per 1,000 video views (separate add-on)
- Smart Encoding: Optional 2x cost for AV1 + HEVC support
AWS MediaConvert + CloudFront
- MediaConvert encoding: $0.0075-$0.0135/min (Basic) or $0.0150-$0.0405/min (Professional with HEVC, AV1)
- S3 storage: $23/TB/month for HLS/DASH segments
- CloudFront delivery: $85/TB egress to North America
- Live streaming: AWS Elemental MediaLive at $0.36-$8.31/hour depending on resolution
- Players: Build-your-own (Video.js, hls.js) or pay for a separate player service
Self-Hosted FFmpeg + Bunny CDN
- Encoding compute: EC2/Fargate Spot for FFmpeg (~$0.04-0.08 per video minute encoded depending on quality)
- Object storage: Bunny Storage at $5/TB or B2 at $6/TB
- Delivery: Bunny CDN at $5/TB (Standard) or $10/TB (Volume)
- Live streaming: Run nginx-rtmp or NodeMediaServer + Bunny Stream
- Engineering: 0.5-1.0 FTE dedicated to video infrastructure
Real-World Cost Modeling: Three Production Workloads
Let's model three realistic workload profiles. May 2026 pricing.
Workload A: SaaS Course Platform (Mid-Volume VOD)
A learning platform with:
- 800 hours of stored video content
- 12,000 hours/month delivered to learners
- Adaptive bitrate (1080p, 720p, 480p)
- HLS playback in modern browsers
Cloudflare Stream:
- Storage: 800 hours x 60 min x ($5/1000) = $240/month
- Delivery: 12,000 hours x 60 min x ($1/1000) = $720/month
- Total: $960/month
Mux:
- Encoding (one-time amortized over 12 months): 800 hours x 60 x $0.005 / 12 = $20/month
- Storage: 800 hours x 60 min x $0.003 = $144/month
- Delivery: 12,000 hours x 60 x $0.00096 = $691/month
- Mux Data (12K x 1.2 views/hour avg): $1.44/month
- Smart Encoding: Often add 2x to encoding ($40 amortized)
- Total: ~$896/month (without Mux Data Pro features)
Wait — that's actually competitive! The reason: Mux pricing has fallen substantially in 2024-2025 to compete with Cloudflare Stream. Let's recompute with realistic Mux Data Pro adoption (most production deployments enable it):
Mux with Mux Data Pro: $896 + $200/month base for Mux Data Pro = ~$1,100/month
Still close. The real Mux premium kicks in at higher volume.
AWS MediaConvert + CloudFront:
- Encoding: 800 hours x 60 min x $0.0135 = $648 amortized over 12 months = $54/month
- S3 storage: 800 hours of HLS segments roughly 50TB = $1,150/month (HLS is 3x raw video size)
- Delivery: 12,000 hours x ~3GB/hour = 36TB at $85/TB = $3,060/month
- Build-your-own player: 0 marginal cost
- Total: ~$4,264/month
Self-Hosted FFmpeg + Bunny CDN:
- Encoding compute (Fargate Spot): 800 hours encoding amortized = ~$80/month
- Bunny Storage: 800 hours of HLS roughly 4TB = $20/month (Bunny Storage is far cheaper than S3 for video)
- Bunny CDN delivery: 12,000 hours x 3GB = 36TB x $5 = $180/month
- Engineering: 0.25 FTE dedicated = $5,000/month fully loaded
- Total: ~$5,280/month
Verdict at this scale: Cloudflare Stream wins on direct cost at $960/month. Mux is a close second at $896-$1,100/month. AWS MediaConvert is 4-5x more expensive ($4,264). Self-hosted is dominated by engineering overhead and not cost-competitive at this volume.
Workload B: Consumer OTT Platform (High-Volume VOD)
A premium OTT streaming service:
- 2,000 hours of premium content
- 200,000 hours/month delivered globally
- Multiple bitrate ladders (4K, 1080p, 720p, 480p, 240p)
- DRM (Widevine + FairPlay)
- Detailed per-viewer analytics required
Cloudflare Stream:
- Storage: 2,000 x 60 x $5/1000 = $600/month
- Delivery: 200,000 x 60 x $1/1000 = $12,000/month
- DRM: Available but limited; may require separate DRM service
- Total: ~$12,600 + DRM service ~$500 = $13,100/month
Mux:
- Encoding amortized: 2,000 x 60 x $0.005 / 12 = $50/month
- Storage: 2,000 x 60 x $0.003 = $360/month
- Delivery: 200,000 x 60 x $0.00096 = $11,520/month
- Mux Data Pro: $500-$2,000/month for OTT analytics
- DRM: $1,000-$3,000/month enterprise tier
- Total: ~$14,000-$17,000/month
AWS MediaConvert + CloudFront:
- Encoding amortized (Pro tier with HEVC + AV1): 2,000 x 60 x $0.025 / 12 = $250/month
- S3 storage: 2,000 hours of multi-codec HLS roughly 200TB = $4,600/month
- Delivery: 200,000 hours x 3GB = 600TB x $85 = $51,000/month
- DRM: Use AWS Elemental Server Side encoder + license keys ($2,000-$5,000/month)
- Total: ~$58,000/month
Self-Hosted FFmpeg + Bunny CDN Volume:
- Encoding compute: ~$200/month
- Bunny Storage: 200TB x $5 = $1,000/month
- Bunny CDN Volume tier: 600TB x $10 = $6,000/month
- DRM: Self-implemented Widevine/FairPlay license server (~$500/month operational)
- Engineering: 1.0 FTE = $20,000/month fully loaded
- Total: ~$27,700/month
Verdict at OTT scale: Cloudflare Stream and Mux are now within $4K/month of each other. Mux's analytics premium becomes worth it for OTT platforms where viewer behavior data drives content investment decisions. AWS MediaConvert + CloudFront is 4.4x more expensive than Cloudflare Stream — the egress fees crush the economics. Self-hosted Bunny + FFmpeg becomes the cheapest direct cost option ($7,700 ex-engineering) but engineering overhead pulls it back.
At OTT scale, the choice is Mux or Cloudflare Stream depending on analytics needs. AWS is structurally wrong; self-hosted needs even higher volume to justify.
Workload C: Internal Corporate Training (Low-Volume VOD)
An enterprise training platform:
- 500 hours of compliance training videos
- 3,000 hours/month delivered to internal employees
- Single bitrate (720p sufficient for training content)
- SSO-protected (no DRM needed)
Cloudflare Stream:
- Storage: 500 x 60 x $5/1000 = $150/month
- Delivery: 3,000 x 60 x $1/1000 = $180/month
- Total: $330/month
Mux:
- Encoding amortized: 500 x 60 x $0.005 / 12 = $13/month
- Storage: 500 x 60 x $0.003 = $90/month
- Delivery: 3,000 x 60 x $0.00096 = $173/month
- Total: $276/month (Mux Data optional at this scale)
AWS MediaConvert + CloudFront:
- Encoding: 500 x 60 x $0.0135 / 12 = $34/month
- S3 storage: 500 hours = ~25TB at $23 = $575/month
- Delivery: 3,000 hours = 9TB x $85 = $765/month
- Total: $1,374/month
Verdict at low scale: Mux and Cloudflare Stream are nearly identical (~$300/month). AWS is 4x more expensive. Self-hosting at this scale is uneconomic regardless of platform — you pay engineering overhead that does not amortize over usage.
The Decision Framework: 5 Questions
Question 1: What is your monthly delivery volume?
- Under 5,000 hours/month: Cloudflare Stream or Mux. Cost difference is marginal; pick on developer experience.
- 5,000-50,000 hours/month: Cloudflare Stream usually wins on cost. Mux only if analytics matter.
- 50,000-500,000 hours/month: Cloudflare Stream or Mux at premium. Avoid AWS unless ecosystem-locked.
- 500,000-1M hours/month: Cloudflare Stream remains competitive. Self-hosted starts to break even.
- Over 1M hours/month: Self-hosted FFmpeg + Bunny CDN dominates if you have the engineering. Otherwise Cloudflare Stream Enterprise.
Question 2: How important is playback quality and analytics?
- Best-in-class quality, premium analytics: Mux. Pay the premium for what they do well.
- Production-quality playback, basic analytics: Cloudflare Stream is sufficient.
- Internal/training use, basic playback: Cloudflare Stream or Mux; pick on budget.
- Marketing video, low-stakes: Cheapest option (Cloudflare Stream or self-hosted with simple player).
Question 3: What is your DRM requirement?
- No DRM: All four platforms work; pick on cost.
- Basic protection (signed URLs, token-based): Cloudflare Stream, Mux, or self-hosted all support this.
- Premium DRM (Widevine + FairPlay + PlayReady): Mux has best-in-class. AWS via Elemental works. Cloudflare Stream support is improving but limited.
- Studio-grade DRM: Mux or specialized services like BuyDRM. This is a small minority of use cases.
Question 4: What is your existing infrastructure?
- AWS-native deeply integrated: AWS MediaConvert + CloudFront despite higher cost (ecosystem savings)
- Cloudflare-native (R2, Workers, etc.): Cloudflare Stream is the natural fit
- Multi-cloud or independent: Pick on cost (Mux or Cloudflare Stream)
- Greenfield project: Cloudflare Stream as default; switch later if quality demands escalate
Question 5: What is your live streaming requirement?
- No live streaming: All four platforms work for VOD.
- Occasional live (events, webinars): Mux's live + recording integration is best. Cloudflare Stream supports live but the experience is rougher.
- Continuous live (24/7 channels, broadcasting): Self-hosted with HLS or AWS Elemental MediaLive.
- Interactive live (gaming, low-latency): WebRTC-based services (LiveKit, Agora, Daily) — none of the platforms above are optimized for sub-300ms latency.
When To Pick Each Platform (Cheat Sheet)
| Use Case | Best Platform | Why |
|---|---|---|
| SaaS app embedded video | Cloudflare Stream | Lowest cost, simple integration |
| Education/courses platform | Cloudflare Stream | Bundle pricing fits cohort delivery |
| Premium OTT/SVOD | Mux | Quality control + analytics |
| Corporate training (internal) | Cloudflare Stream | Cheapest, sufficient quality |
| Marketing landing page video | Cloudflare Stream or self-hosted | Low volume, low cost |
| Live event streaming (occasional) | Mux | Best live experience |
| Live broadcasting (24/7) | AWS MediaLive or self-hosted | Operational fit |
| User-generated content (social) | Cloudflare Stream | Volume + cost predictability |
| Documentation video | Cloudflare Stream or YouTube embed | Frequency of view = low |
| ML/computer vision pipelines | Self-hosted FFmpeg | Custom processing needs |
| 4K/HDR premium content | Mux | Best encoding for premium |
| News/media publishing | Mux or Cloudflare Stream | Low-latency live |
| Fitness/wellness on-demand | Cloudflare Stream | High delivery, analytics secondary |
| Sports streaming (premium) | Mux + low-latency live | Quality and analytics critical |
| Compliance recording archive | AWS S3 + MediaConvert on demand | Rarely viewed, archival pricing |
| Video surveillance retention | Backblaze B2 or S3 Glacier | Archival, no streaming optimization needed |
Hidden Costs Most Video Comparisons Miss
Hidden Cost 1: Encoding Multipliers For Multi-Codec Output
If you support multiple codecs (H.264 + HEVC + AV1) for older devices, your encoding cost multiplies. Mux Smart Encoding doubles cost for AV1+HEVC. AWS MediaConvert charges per-codec per-output. Cloudflare Stream includes multi-codec free (encodes to H.264+HEVC automatically).
Mitigation: Decide whether AV1/HEVC is worth the encoding premium. For most consumer SaaS, H.264 alone is fine.
Hidden Cost 2: Adaptive Bitrate Ladder Storage
A single source video stored as 5 bitrates (1080p, 720p, 480p, 360p, 240p) takes 3-5x the storage of the source. Mux includes ABR in pricing; AWS charges S3 per byte. Cloudflare Stream's storage pricing is bitrate-ladder-aware.
Mitigation: Reduce the bitrate ladder for use cases where lowest quality is unnecessary (training, internal video).
Hidden Cost 3: Reuploading For Format Changes
If you switch encoding strategies (add a new codec, change bitrate ladder), you typically re-encode the entire library. For 1,000 hours of video, that's $300-$2,000 in one-time encoding cost depending on platform. Cloudflare Stream re-encodes free; AWS and self-hosted charge again.
Mitigation: Get encoding strategy right at first upload to avoid re-encoding.
Hidden Cost 4: Per-Region Delivery Pricing
CloudFront charges differential delivery prices by region: $85/TB to North America, $114/TB to Australia/India. For globally distributed audiences, average egress cost is 1.2-1.5x the headline NA rate. Cloudflare Stream and Mux charge global flat rates.
Mitigation: Calculate effective region-weighted egress cost when comparing AWS to flat-rate alternatives.
Hidden Cost 5: Player Library Costs
Mux's built-in player is free with the service. Cloudflare Stream's built-in player is free. AWS expects you to BYO player (Video.js, hls.js, JW Player). Commercial players cost $300-$3,000/month for branding, analytics, ad insertion.
Mitigation: Use open-source players (hls.js, Video.js) unless you need premium features.
Hidden Cost 6: Live Streaming Origin Cost
Mux live streaming charges per-minute streamed in. AWS MediaLive charges per-hour ($0.36-$8.31/hr) regardless of viewer count. Self-hosted requires keeping origin servers running 24/7 even when no live event is happening.
Mitigation: Match live model to your event frequency. Mux for occasional events, MediaLive or self-hosted for continuous live.
Hidden Cost 7: Engineering Time For Self-Hosted
Self-hosted FFmpeg + CDN requires real engineering: encoding pipeline, error handling, retry logic, format detection, thumbnail generation, ABR ladder selection, DRM if needed. We have seen teams spend 6-12 months building what Mux ships in a 5-line API call.
Mitigation: Self-host only at scale where engineering investment amortizes. For most teams, manage video as a service.
Migration Playbook: Mux → Cloudflare Stream
For teams overpaying on Mux for non-OTT use cases, migration takes 3-6 weeks. Here is the playbook from real migrations.
Phase 1: Inventory and Categorize (1 Week)
- Pull list of all videos with size, duration, view counts
- Identify the 80/20: most-watched 20% drive 80% of delivery cost
- Tag videos by use case (course, marketing, internal, premium)
- Calculate cost-per-video for high-cost outliers
Phase 2: Test Cloudflare Stream Quality (1 Week)
- Upload sample of 20-50 videos covering quality range
- Test playback on real devices: Safari (iOS/macOS), Chrome (desktop/Android), Firefox
- Verify HLS quality matches Mux on identical bitrate ladders
- Check player customization, signed URLs, captions
- Validate analytics meet your minimum bar
Phase 3: Parallel Operation (2-4 Weeks)
- Implement abstraction layer that can route video URLs to either platform
- Migrate 5% of video traffic to Cloudflare Stream
- Monitor: playback errors, buffering, completion rates, support tickets
- Increase to 25% / 50% / 100% over 3 weeks
- Validate cost savings against forecast
Phase 4: Decommission Mux (1 Week)
- After 30 days at 100% on Cloudflare Stream, stop Mux uploads
- Keep Mux library for 60 more days as fallback
- Decommission after 90 days clean operation
- Lock in savings and document migration learnings
Typical outcome from real migrations:
- 70-90% direct cost reduction
- Playback quality identical for VOD (slight delta on premium analytics)
- Zero customer-perceived impact on B2B education and SaaS use cases
- Mild playback regression on Apple devices when ABR ladder differs (tunable)
When To Stay On Mux (Or Switch To Mux)
Don't migrate away from Mux if:
- You're an OTT/SVOD product where quality control IS the product
- Your business depends on Mux Data analytics for content investment decisions
- Premium DRM (FairPlay, Widevine L1) is non-negotiable
- Your team has zero capacity for migration and you're willing to pay the premium
- Integration with Mux's APIs is deeply embedded (Mux's developer-facing tooling is genuinely best-in-class)
For roughly 25-30% of Mux customers, staying is the right call. The other 70%+ would save 60-90% by switching.
A 30-Day Video Streaming Cost Audit
If your video streaming bill is over $10,000/month, run this audit. Typical finding: 40-80% cost reduction opportunity.
Week 1: Visibility
- Pull last 90 days of video platform bills with delivery breakdown
- Identify cost per hour delivered, cost per hour stored
- Map videos to use cases (premium, training, marketing, etc.)
- Identify the 20% of content driving 80% of cost
Week 2: Workload Categorize
For each major use case:
- Quality requirement (HD vs 4K, ABR ladder needed)
- Analytics requirement (basic vs detailed)
- DRM requirement
- Live vs VOD vs both
- Geographic distribution
Week 3: Cost Model Alternatives
For each use case, calculate cost on:
- Cloudflare Stream (likely cheapest for VOD)
- Mux (current)
- AWS MediaConvert + CloudFront (if AWS-locked)
- Self-hosted (only if over 500K hours/month)
Week 4: Decide and Plan
For workloads where annual savings exceed migration cost by 3x:
- Plan a 3-6 week migration
- Build URL abstraction so you can mix platforms
- Test thoroughly on real devices before cutover
- Lock in savings and reassess every 6 months
The Bottom Line
In 2026, video streaming platform choice is a workload-fit decision, not a vendor-loyalty decision. Cloudflare Stream wins for most VOD workloads at a fraction of the cost of alternatives. Mux wins for premium OTT and analytics-driven products despite the cost. AWS MediaConvert + CloudFront is rarely the right choice in 2026 due to egress fees. Self-hosted dominates only at extreme scale.
The discipline most teams skip: benchmarking actual cost on each platform for your specific workload before committing. Most teams pick once and never revisit; that inertia costs $50K-$500K/year for mid-volume deployments.
If your video streaming bill is over $20,000/month and you have not benchmarked alternatives in the last 12 months, you are very likely overpaying by 60-90%. Our cloud cost optimization team runs free video streaming audits and typically captures 60-80% savings within 60 days. Run a free Cloud Waste Scorecard to find your biggest media infrastructure leaks first.
Further reading:
- Media Storage Serverless Cost Comparison 2026
- CDN Cost Showdown: CloudFront vs Cloudflare vs Bunny vs Fastly
- Best Cloud Storage by Workload Decision Framework 2026
- Cloudflare R2 vs AWS S3 Decision Framework 2026
- Cloudflare Images Pricing 2026
- AWS Data Transfer Pricing 2026
- Cloud Cost Optimization FinOps Service
- Cloudflare Stream Pricing
- Mux Pricing
- AWS MediaConvert Pricing
