# ENOVIA AI Assistant Sizing and Value Calculator

> Source: https://netray.co/tools/enovia-ai-assistant-sizing-and-value-calculator
> Publisher: Netray (netray.co), AI and integration engineering for ERP and manufacturing systems.
> Section: Free Tools

## Short answer

Estimate GPU count, token volume, index size and payback for a large private AI assistant on on-prem ENOVIA.

## Category

engineering-plm

## Erp

Dassault ENOVIA

## Audience

- PLM administrator
- IT architect
- CIO

## Tool

**Kind:** calculator

**Currency:** $

**Inputs:**

#### Named users

**Key:** users

**Type:** number

**Default:** 500

**Min:** 1

**Step:** 25

**Unit:** users

#### Daily active share

**Key:** active_pct

**Type:** slider

**Default:** 30

**Min:** 5

**Max:** 100

**Step:** 5

**Unit:** %

#### Questions per active user per day

**Key:** q_per_user

**Type:** number

**Default:** 9

**Min:** 1

**Step:** 1

#### Prompt plus answer tokens per question

**Key:** tokens_per_q

**Type:** number

**Default:** 4000

**Min:** 200

**Step:** 100

#### Documents to index

**Key:** docs

**Type:** number

**Default:** 600000

**Min:** 0

**Step:** 25000

#### Average pages per document

**Key:** pages_per_doc

**Type:** number

**Default:** 8

**Min:** 1

**Step:** 1

#### Minutes saved per question

**Key:** min_saved

**Type:** number

**Default:** 6

**Min:** 0

**Step:** 1

#### Loaded hourly cost

**Key:** hourly_cost

**Type:** number

**Default:** 85

**Min:** 0

**Step:** 5

**Unit:** $/h

#### GPU monthly cost

**Key:** gpu_cost

**Type:** number

**Default:** 2500

**Min:** 0

**Step:** 100

**Unit:** $/GPU/month

#### Aggregate tokens per second per GPU

**Key:** gpu_tps

**Type:** number

**Default:** 1400

**Min:** 100

**Step:** 100

#### GPUs needed just to hold the model

**Key:** min_gpus

**Type:** number

**Default:** 2

**Min:** 1

**Step:** 1

#### One-off integration cost

**Key:** integ_cost

**Type:** number

**Default:** 120000

**Min:** 0

**Step:** 10000

**Unit:** $

**Outputs:**

#### Daily active users

**Key:** active_users

**Expr:** users * active_pct / 100

**Format:** integer

#### Questions per day

**Key:** q_day

**Expr:** active_users * q_per_user

**Format:** integer

#### Tokens per day

**Key:** tokens_day

**Expr:** q_day * tokens_per_q

**Format:** number

#### Peak concurrent requests

**Key:** peak_conc

**Expr:** q_day * 0.25 / 3600 * 20

**Format:** decimal

**Description:** 25% of daily questions in the busiest hour, about 20 seconds per answer.

#### GPU estimate

**Key:** gpus

**Expr:** max(min_gpus, ceil(peak_conc * 25 / gpu_tps))

**Format:** integer

**Primary:** true

**Description:** About 25 tokens per second per stream, with a model memory floor.

#### Vector index size

**Key:** idx_gb

**Expr:** docs * pages_per_doc * 2 * 6000 / 1000000000

**Format:** decimal

**Description:** 2 chunks per page and about 6 KB per chunk.

#### Hours saved per month

**Key:** hours_month

**Expr:** q_day * min_saved / 60 * 21

**Format:** hours

#### Monthly value

**Key:** value_month

**Expr:** hours_month * hourly_cost

**Format:** currency

**Primary:** true

#### Monthly GPU cost

**Key:** cost_month

**Expr:** gpus * gpu_cost

**Format:** currency

#### Simple payback

**Key:** payback_months

**Expr:** if(gt(value_month - cost_month, 0), integ_cost / (value_month - cost_month), 0)

**Format:** months

**Chart:**

#### Monthly value vs GPU cost by questions per active user

**Type:** line

**X:** q_per_user

**From:** 2

**To:** 40

**Steps:** 20

**Series:**

- value_month
- cost_month

**Disclaimer:** Planning estimate for an enterprise rollout. Validate throughput and time saved in a pilot before procurement.

## Content

**Intro:** This calculator estimates GPU needs, index size and payback for an enterprise-scale private AI assistant on on-prem ENOVIA or 3DEXPERIENCE. Defaults reflect a few hundred engineers and a large document base.

**Sections:**

#### How to use it

**Body:** Enter named users and the daily active share. Use questions per user from a pilot where possible. Set tokens per question high enough to reflect retrieved context, since engineering questions often pull in long passages. Enter the one-off integration cost including custom MQL or JPO work for your schema.

#### Formulas and assumptions

**Body:** Questions per day equal users times active share times questions per user. Peak streams assume 25 percent of daily questions in the busiest hour at about 20 seconds each. GPUs are the larger of the model floor and streams times 25 tokens per second divided by per-GPU throughput. The index assumes two chunks per page at about 6 KB.

**Bullets:**

- Hours = questions x minutes saved / 60 x 21
- Value = hours x loaded hourly cost
- Payback = one-off cost / (value minus GPU cost)

#### Reading the result

**Body:** Enterprise estimates are sensitive to the memory floor because larger models need several GPUs just to load. The chart compares value with cost as usage per user grows: the crossing point is the usage level where the platform pays for itself. Add separate CPU capacity for extraction and embedding.

#### Common mistakes

**Body:** Using the number of documents in the platform instead of text-bearing documents you actually index, ignoring custom integration work in a customised data model, and assuming perfect adoption. Apply a haircut to active share and minutes saved.

**Cta:** Netray can compare these figures with a pilot on your ENOVIA platform and refine the integration estimate. The calculator needs no contact.

**Faq:**

#### Why is the default integration cost higher?

**Answer:** ENOVIA models are commonly customised, so named read actions, security context handling and testing take more effort than on standard schemas. Replace the default with your own estimate.

#### How should I treat redundancy?

**Answer:** The estimate has no spare capacity. For a production service add at least one spare GPU or a second node, and include it in the GPU monthly cost.

#### Does a larger model save time?

**Answer:** Not necessarily. Larger models raise cost and memory floors. Test whether a mid-sized model with good retrieval and tools answers your questions correctly before scaling hardware.

#### What is not in the payback?

**Answer:** Power, cooling, support staff, licensing for other components and model evaluation work, unless you include them in the cost inputs.

## Search terms

ENOVIA AI sizing, 3DEXPERIENCE on-prem AI cost, enterprise LLM GPU sizing, PLM AI payback, on-prem AI value calculator, RAG index size calculator

---
Talk to an engineer: https://netray.co/contact?intent=erp_ai | Machine catalog: https://netray.co/agent/catalog.json