ABCDEFGHIJKLMNOPQRSTUVWXYZ
1
AI Cost Volatility: Build vs Buy Analysis
2
3
IMPORTANT: This is an educational model to illustrate AI cost volatility concepts.
4
Defaults are illustrative only. Customize all blue cells with your actual data.
5
6
Purpose:
7
Show how AI inference costs make traditional build-vs-buy TCO analysis unreliable.
8
9
The Four Options:
10
 1. Traditional SaaS - Predictable per-seat costs
<-- The traditional options are just here as reference points.
11
 2. Traditional Build - High upfront, low ongoing
12
 3. SaaS + AI - Per-seat PLUS variable inference
13
 4. Build + AI - High upfront PLUS variable inference
14
15
How to Use:
16
 1. Edit BLUE cells in Assumptions sheet
17
2. Focus on AI parameters (monthly interactions, tokens, pricing)
18
3. Try usage multiplier at 1x, 5x, 10x to see cost volatility
19
 4. View Model sheet for 3-year totals
20
 5. View Charts for visual comparison
21
22
About The Baseline Assumptions:
23
This model uses mid-market defaults (1,000 users). At this scale:
24
Most companies choose SaaS for lower risk and faster deployment
25
Traditional costs are somewhat predictable (grow with seats).
26
AI costs are volatile (can spike with usage patterns).
27
I've just scaled additional costs for things based on seats, which might not be strictly true, but this is for just a rough cut.
28
29
Main Points: Whatever You Do With the Numbers
30
Four different general rules of thumb…
31
32
Traditional SaaS: Costs scale with seats
33
Traditional Build: Mostly Fixed costs, not much scaling
34
SaaS + AI: Split between seat costs and AI variable costs
35
Build + AI: Fixed base + variable AI costs
36
37
The result:
38
39
Traditional options: Usually Predictable
40
AI options: Still Volatile
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100