Build vs Buy: Auth, Database, Hosting at $100/Hour

"Should I build or buy infrastructure?" is the most-asked early-founder question, and the answer is almost always buy. This article runs the decision engine on three real components (auth, database, hosting) with realistic founder economics, breaks down each verdict, and names the rare cases where building is actually the right call.

1. Three components priced literally

Inputs to the decision engine: hourly value $100/hour (loaded rate for a solo founder with mid-career engineering experience), three components enabled (auth, database, hosting), industry-typical build effort and maintenance hours, managed-service costs from real vendor pricing.

• Auth: 40 hours to build, 4 hours/month maintenance, $25/month managed (Clerk Pro^[1] or Supabase Auth^[2] at typical solo scale)
• Database: 20 hours to build (self-hosted Postgres on a VPS), 2 hours/month maintenance, $25/month managed (Supabase Pro tier with included database, backups, and read replicas)
• Hosting: 16 hours to build (self-managed reverse proxy + CI/CD on a VPS), 3 hours/month maintenance, $20/month managed (Vercel Pro^[3])

# build-vs-buy-decision-engine (computed live from /engines/build-vs-buy-decision-engine.js)
Engine input
  hourly_value          = 100
  components[0].name    = auth
  components[0].enabled = true
  components[0].time_to_build_hours= 40
  components[0].maintenance_hours_per_month= 4
  components[0].managed_service_cost= 25
  components[1].name    = database
  components[1].enabled = true
  components[1].time_to_build_hours= 20
  components[1].maintenance_hours_per_month= 2
  components[1].managed_service_cost= 25
  components[2].name    = hosting
  components[2].enabled = true
  components[2].time_to_build_hours= 16
  components[2].maintenance_hours_per_month= 3
  components[2].managed_service_cost= 20

Engine output
  components[0].name    = Auth
  components[0].buildCostOneTime= 4000
  components[0].annualMaintenanceCost= 4800
  components[0].annualBuildTotalCost= 8800
  components[0].annualBuyCost= 300
  components[0].breakEvenMonths= null
  components[0].verdict = BUY
  components[0].reasoning= Buying Auth at $25/mo costs $300/year but saves 40h upfront + 4h/mo of your time ($8800/year equivalent).
  components[1].name    = Database
  components[1].buildCostOneTime= 2000
  components[1].annualMaintenanceCost= 2400
  components[1].annualBuildTotalCost= 4400
  components[1].annualBuyCost= 300
  components[1].breakEvenMonths= null
  components[1].verdict = BUY
  components[1].reasoning= Buying Database at $25/mo costs $300/year but saves 20h upfront + 2h/mo of your time ($4400/year equivalent).
  components[2].name    = Hosting
  components[2].buildCostOneTime= 1600
  components[2].annualMaintenanceCost= 3600
  components[2].annualBuildTotalCost= 5200
  components[2].annualBuyCost= 240
  components[2].breakEvenMonths= null
  components[2].verdict = BUY
  components[2].reasoning= Buying Hosting at $20/mo costs $240/year but saves 16h upfront + 3h/mo of your time ($5200/year equivalent).
  totalBuildHours       = 76
  totalAnnualBuildCost  = 18400
  totalAnnualBuyCost    = 840
  annualSavings         = 17560
  overallVerdict        = BUY
  insight               = Building everything yourself costs $18400/year in your time. Buying saves 9 hours/month for $70/month. At $100/hour, buying frees up time worth more than the subscription cost.

The engine returns the breakdown per component plus the aggregate. Total build hours: 76. Total annual build cost: $18,400 ($7,600 upfront at $100/hr + $10,800/year of maintenance). Total annual buy cost: $840 ($70/month × 12). Annual savings if you buy: $17,560.

2. The $17,560/year annual savings line

The $17,560/year savings number is the single most useful output of the engine. It is the founder time recovered by buying instead of building, expressed in dollars at the founder's hourly value. At 9 hours per month of maintenance work avoided, that is 108 hours per year — roughly 14 working days of focused engineering time, freed up to spend on the product surface that actually moves revenue.

The non-obvious part of this number: it does not include the opportunity cost of the original 76-hour build. The 40+20+16=76 hours of upfront engineering, at $100/hour, is $7,600 of founder time spent on infrastructure that does not differentiate the product. The vendor delivers the same capability at $70/month, with no upfront cost and a fixed ongoing line item.

The 14-working-days-per-year freed by buying is also self-compounding. Those days, redirected to product or distribution work, become the revenue that funds the next quarter's buy decisions. The build vs buy solo founder framework covers the broader pattern; this article runs the literal numbers on the three components most common in solo SaaS.

3. Auth: the easiest buy decision in the stack

The engine returns: build cost $4,000 (40h × $100/hr), annual maintenance $4,800 (4h/mo × $100/hr × 12), annual build total $8,800, annual buy cost $300 (Clerk Pro at $25/month). The build verdict math: $8,800 against $300, no break-even ever arrives. Verdict: BUY.

The hidden cost the engine does not capture: regulated and trust-critical surfaces have edge cases the vendor has already shipped. Password reset flows, MFA, session security, social provider integration, OAuth2 compliance, abuse mitigation, SOC 2 audit trails — every one of these is a feature Clerk or Auth0 ships in the $25/month plan and that a custom build would have to implement and maintain. The 40-hour build estimate covers the happy path. The unhappy paths are where founders spend the next two years.

The U.S. Bureau of Labor Statistics^[4] reports software developer mean wages that translate to roughly $88/hour unloaded, $115/hour loaded. The $100/hour input in the worked scenario is slightly conservative. Use $150/hour for a senior solo founder with strong engineering background, which only widens the buy case.

4. Database: cheaper to build than auth, still a buy

The engine returns: build cost $2,000 (20h), annual maintenance $2,400 (2h/mo), annual build total $4,400, annual buy cost $300 (Supabase Pro database tier). Build verdict math: $4,400 against $300, BUY by 14x.

The database build case is the cleanest of the three to argue. Postgres on a $20/month DigitalOcean droplet is a well-trodden path, takes 20 hours to set up properly (including backups, monitoring, point-in-time recovery), and runs reliably for years. The reason buy still wins: the 2 hours/month maintenance is real, and the failure modes (drive full, replication broken, certificate expired) compound when the founder is on vacation or focused elsewhere.

Supabase Pro at $25/month bundles the database, automated backups, point-in-time recovery, connection pooling, and the auth surface on top. For solo founders, the trade-off is rarely about cost — it is about operational surface. One fewer thing to monitor is one fewer thing that can break at 2am. The embeddings DB cost article covers the related vector-storage decision.

5. Hosting: where the build case is occasionally real

The engine returns: build cost $1,600 (16h), annual maintenance $3,600 (3h/mo), annual build total $5,200, annual buy cost $240 (Vercel Pro at $20/month). Build verdict math: $5,200 against $240, BUY by 22x.

Hosting is the component where the build case has historically been most defensible. A $5/month VPS running nginx + a Node app handles 10,000 users on a small SaaS, and the engineering effort to set it up is real but bounded. The maintenance, however, is the cost the engine prices honestly: 3 hours/month is realistic for a self-managed VPS (security updates, certificate renewals, log rotation, deploy debugging, occasional restart).

The buy verdict on hosting is also a verdict on developer experience. Vercel's deploy-on-git-push, preview-per-branch, and rollback-in-seconds features are not just convenience — they are velocity. Solo founders ship 2x to 3x more features on a Vercel-class platform than on a self-managed VPS, not because the deployment is meaningfully different in capability, but because the friction is. The 2026 AI solopreneur stack covers the broader hosting picture.

6. How hourly value swings the answer

Re-run the engine at $40/hour (the lower end of credible solo-founder loaded rates). Auth maintenance drops to $1,920/year, hosting maintenance to $1,440/year, database to $960/year. Total annual build cost drops from $18,400 to $7,360. Annual buy cost is unchanged at $840. Savings drop from $17,560 to $6,520. Verdict still BUY decisively.

Re-run at $150/hour. Total annual build cost rises to $27,600. Annual buy still $840. Savings climb to $26,760. The verdict is BUY at any rate above $20/hour. The only scenario where build wins is hourly value under $10/hour, which represents zero founders and a small number of student-built side projects.

The implication: do not over-think the hourly value input. Pick a number between $50 and $150 that maps to your actual market rate, the verdict will not change for the auth/database/hosting trio. The verdicts get more interesting in the next tier of components — analytics, monitoring, search relevance — where the build/buy math can flip at specific scale points.

7. When the build verdict is correct

Three named cases where the engine's BUY verdict should be overridden.

1. You already have the component built. The 40-hour auth build is sunk cost if you already wrote it for a previous project. The relevant comparison is then 4h/mo maintenance ($4,800/year) against $300/year vendor. Buy still wins on maintenance alone if you value your time, but the gap is smaller and the sunk-cost component matters for migration planning.
2. The vendor's pricing scales faster than your revenue. Per-seat or per-user pricing models bite at 12-month-out scale. A $25/month Clerk plan is $25/month at 1,000 users and $1,200/month at 50,000 users. If revenue does not scale linearly with users, the vendor cost can exceed the build maintenance at 12-month-out scale. Run the math at projected scale, not current.
3. The vendor is a single point of failure for your core differentiation. If your product's value depends on a specific vendor capability and the vendor could deprecate it, change pricing punitively, or disappear, the build investment becomes a moat investment. Rare, but real for some AI-tooling products dependent on a specific frontier model.

The methodology behind the engine's break-even calculation and verdict logic is documented at the Build vs Buy Decision Engine methodology page^[5].

8. FAQ

Should I build or buy auth? Buy. The engine shows $8,800/year build vs $300/year buy at $100/hr. The regulatory tail makes the buy case even stronger than the raw math.

What hourly value should I use? Loaded opportunity cost. $75 to $150 for most solo founders. The verdict does not change between $40 and $200.

When does building actually beat buying? When you already have it built, when vendor pricing scales faster than your revenue, or when the vendor is a single point of failure for your core differentiation.

Why is break-even null on every component? Because monthly maintenance cost (at your hourly rate) exceeds the monthly vendor cost. The build never pays back at solo-founder hourly values.