Platform
Solutions
Developers
Company
Pricing
Request a pilot Explore the platform

Pricing

Priced on the value of a single yield point.

On a CoWoS-class or HBM package, one point of yield is worth more than most software budgets. Chipira is priced to land narrow, expand by proof, and — where it matters most — to be paid on the outcome.

See the plans

Plans

Three ways to buy

Start on one tool. Expand across the line. Standardise across the network.

Billed monthly Annual prepay (≈15–20% lower)

Line

$18,000 per tool or bonder line, per month

For packaging houses validating ROI on the wedge workflow.

  • One agent capability on one tool or line
  • Bonding and alignment control, or warpage and void vision, or stacking optimisation
  • Fab-edge runtime and connector layer
  • Review console, alerting and control integration
  • Immutable audit trail
  • Business-hours support

Scope a Line pilot

Fab · Recommended

$120,000 per month

For scaling lines adopting the full closed loop.

  • Whole line: place and bond, stack and TSV, warpage and defect sensing
  • Mold and underfill, robotic handling, test and bin
  • Yield and ramp intelligence
  • Package-and-line twin
  • Fab-edge fleet across the line
  • Priority support and quarterly model review

Scope a Fab deployment

Enterprise

Custom typical landing ACV $800k – $9M

For OSATs and IDMs standardising on Chipira.

  • Multi-fab deployment and governance
  • Custom process, defect and yield models
  • Fab-edge fleet management across sites
  • Named support engineers and contractual SLAs
  • On-premise or air-gapped control plane
  • Outcome-based commercial components

Speak with our team

List pricing shown for transparency; every engagement is scoped against tool count, sensor load, autonomy level and support requirements. Annual prepay figures are indicative of the 15–20% discount band, not a fixed rate card.

The logic

Land low. Expand on proof. Charge for outcome.

The Line plan exists so a fab can find out whether we are real without a committee. One tool, one workflow, one number. If it does not produce the number, you should not buy the second one.

The Fab plan exists because the loop compounds. Perception makes control more accurate; control produces better perception data; the twin makes both defensible; yield intelligence retrains all of them. Buying the loop is worth materially more than buying five point capabilities — but only after the first one has proven itself.

And on the highest-value workflows, list pricing is the wrong instrument entirely. Each point of yield on a CoWoS-class or HBM package is worth so much that an outcome-based component aligns the two organisations better than any subscription tier can.

  • Land: one workflow with unambiguous ROI
  • Expand: tools, modules, lines, fabs, autonomy level
  • Monetise value: usage and outcome components as the agent does more
  • Guardrails: floor pricing protected; discounts traded for term and case studies

What you get

Capability by plan

CapabilityLineFabEnterprise
One agent on one tool or line●●●
Full seven-agent loop—●●
Package-and-line twin—●●
Custom process and defect models——●
Multi-fab fleet management——●
Immutable audit trail●●●
SSO and RBAC●●●
On-premise / air-gappedOn request●●
Named support engineers——●
Outcome-based components—AvailableAvailable

Metering

What you are actually paying for

No surprise line items. The three things that move a bill are stated up front.

Subscription
Tiered platform access by tool, line and fab. This is the predictable base of every agreement.
Usage
Per-package, per-tool or per-action metering above plan limits, agreed at scoping so a good quarter of volume does not produce a bad invoice.
Outcome
On enterprise agreements, a defined share of measured yield, scrap, known-good-die or ramp-time improvement against a jointly agreed baseline.
Add-ons
Dedicated process, defect and yield models, on-premise or fab-edge fleet expansion, premium support and custom integrations.
Not charged separately
The review console, the audit trail, SSO and RBAC, and the connector library for your in-scope tools are part of the platform, not upsells.

Unit economics

How we think about the relationship

  • 80% Target gross margin at scale, as high-volume inference moves to fine-tuned open models.
  • 137% Net revenue retention target, driven entirely by expansion rather than price increases.
  • <13 mo Target CAC payback on enterprise engagements.
  • 15–20% Annual prepay discount band, traded for term length and reference rights.

Internal targets [ASPIRATIONAL], shown so you can see the incentives that shape our commercial behaviour. They are not reported results.

If a point of yield on this package is worth eight figures a year, do not sell me seats. Sell me the point.

Vice president, back-end operationsAI accelerator supply chain

Composite drawn from design-partner and industry conversations. Illustrative, not a customer endorsement.

Getting to a number

How a quote is built

  1. Scope the wedge

    One workflow, one package family, one line. We establish tool count, sensor load and the integration surface required.

  2. Size the edge

    Compute footprint follows sensor count and resolution. This is the main technical driver of cost and is settled before commercials.

  3. Agree the metric

    The pilot success metric, its baseline, its measurement method and who owns the measurement. In writing, before installation.

  4. Structure the term

    Monthly or annual prepay, term length, expansion ramp, and whether an outcome component makes sense for this workflow.

Commercial questions

The awkward ones

  • No. Pilots are paid, and deliberately so. A free pilot has no internal owner, no urgency and no honest measurement. A paid pilot with a signed success metric produces a decision either way, which is what both sides actually need.

  • Then it fails, and you do not convert. We would rather have that outcome documented than have a design partner quietly drift for two quarters. Failed pilots also produce the most useful engineering feedback we get.

  • Against term length and reference rights, within a protected floor. We do not discount to win a comparison against a point tool, because the comparison is wrong.

  • Because a packaging director should be able to work out whether this is a $200k or a $2M conversation before spending a week on it. Hiding the number wastes your time to protect our negotiating position, which is a bad trade.

Start narrow, expand relentlessly

Land one workflow. Own the loop.

A Chipira engagement begins with a single wedge workflow, a shadow-mode baseline and one signed success metric. Everything after that is expansion.

Request a pilot See how we price