Skip to main content
TOMPW Service logo
TOMPW
← Back to Knowledge Hub

IP Fit Planning Before Tape-Out: A Checklist for ASIC Teams

A practical IP planning checklist for ASIC teams: interface, memory, analog and transceiver IP categories, MPW vs. full-mask license terms, IP hardening and royalty vs. upfront trade-offs.

August 31, 2026Updated August 31, 20268 min readBy Henry
IPIP CoresASICTape-OutMPW

Most IP discussions start with the wrong question. Teams ask which IP to license, when the more useful question is which IP must be locked down before the foundry is chosen.

Get those four or five decisions right and the rest of the program — verification, physical design, packaging, test — gets measurably easier.

Why IP Planning Comes Before Process Planning

Process selection and IP selection are not independent. A node may look attractive on paper, but if the interface IP you need is not foundry-qualified for it, schedule slips quickly.

Treat IP as the constraint and the process as the variable. It usually gives a more defensible plan.

The Four IP Categories to Lock Down First

Across most ASIC programs the same four categories drive schedule and cost more than anything else.

  • Interface IP — USB, PCIe, Ethernet, MIPI, DDR/LPDDR controllers and PHYs
  • Memory IP — SRAM/ROM compilers, eFlash/eEEPROM controllers
  • Analog IP — PLL, ADC/DAC, voltage regulators, SerDes building blocks
  • Transceiver IP — radio, mmWave, high-speed SerDes

MPW vs. Full-Mask License Terms

Several IP vendors treat MPW use differently from full-mask production. Some require a separate evaluation agreement, some charge a higher per-die royalty for prototype runs, and some restrict MPW use entirely.

Confirm license terms with the IP vendor before booking a shuttle. This is one of the cheapest due-diligence steps and one of the most often skipped.

Hardening: One Node, One Library, One Sign-Off

A hard macro is committed to a single PDK and library version. That is the whole point of hardening — it buys you timing and power that are characterised rather than estimated — and it also means moving to another foundry or node is re-hardening, not a file conversion.

Before hardening starts, two questions need written answers: does the licence permit hardening at all, and does it permit the resulting macro to be used in production? Some vendors gate both behind a separate addendum. For MPW tape-outs, foundry-qualified IP is usually required, so hardening time and cost belong in the schedule from the start.

Royalty vs. Upfront: A Commercial Trade-Off

Upfront fees buy access. Royalty fees pay for production use. The right mix depends on expected volume and the MPW-to-mask transition plan.

Some vendors offer a royalty cap or a one-time conversion to a fully paid-up license. Negotiate this early; it rarely improves later in the program.

Where TOMPW Helps

TOMPW does not sell IP. The value here is helping teams think through IP fit, license constraints and the transition from MPW to production before the design is committed to a process.

Where IP procurement is already in motion, we help flag schedule risk and coordinate verification and physical design around the IP choices that have already been made.

Article FAQ

Quick answers tied to this topic

Pick IP first. IP availability (especially for hard IP and PHY) is the most common reason a node choice has to be revised later.

Yes — but only after re-verification. Soft IP portability is real, but the verification cost is often underestimated.

Usually yes. Most foundries require IP qualification before it can be included in a shared mask set. Confirm with the foundry's MPW rules before locking the IP list.

Related Tools

Plan this with a free calculator

Related Reading

Where this topic leads next

One deeper read on the same stage of the program, and one on the next decision you will hit after it.