Glass Substrate · 02

Lead Frame → PCB → ABF → Glass
60 Years of Substrate Evolution

Every time the substrate material changed, the semiconductor industry was reshuffled. That history tells us exactly what to expect from glass.

Beginner 10 min read ~2,200 words Paradigm Shift Lab

←

Previous · Glass Substrate 01

Why Glass? — The Limits of ABF Substrates and the Rise of Glass Core Technology

There's a pattern hiding inside 60 years of semiconductor history.

Every 15 to 20 years, the material holding the chip in place hits a physical wall. Engineers scramble, debate, and eventually commit to something completely different. And every single time that transition happens, the companies that moved early won, and the ones that held on too long lost their footing.

We are living through that transition right now. Understanding where glass substrates come from — and what the previous transitions looked like — is the clearest possible map of what's about to happen next.

1960s – 70s

Lead Frame

Metal stamping. Simple, cheap, scalable.

1980s – 90s

PCB Substrate

Fiberglass + epoxy. More pins, more complexity.

1990s – now

ABF Substrate

Build-up layers. The current standard — hitting its limits.

2027 – ?

Glass Substrate

Ultra-flat, low-loss. The next paradigm.

Era 1 — The Lead Frame (1960s–70s): When a Chip Just Needed Legs

The first semiconductor packages were almost embarrassingly simple. A silicon die was glued to a flat piece of stamped metal — the lead frame — and thin gold or aluminum wires were bonded from the chip's pads to the metal leads. Encapsulate it in black epoxy, trim the excess metal, and you had a chip ready for a circuit board.

The lead frame worked beautifully for its era. Chips were small, pin counts were low, and the speed demands were modest. The entire industry standardized around it — DIP (Dual Inline Package), TO-92, SOP — these familiar black rectangles with metal legs are all lead frame packages.

💡 Why It Worked

Lead frames were punched from thin metal sheets in high-volume stamping presses — the same technology used for coins and soda cans. Fast, cheap, and scalable to billions of units per year.

But as chips grew more complex through the 1980s — more transistors, more I/O pins — the lead frame hit a fundamental wall. You can only stamp so many thin metal legs around the perimeter of a chip before they're too close together to manufacture reliably. The industry needed a new approach. It found one in a material already sitting on most desks: the printed circuit board.

Fig. 1. The four eras of semiconductor substrate packaging — each transition was driven by a physical wall the previous material couldn't cross.

Era 2 — The PCB Substrate (1980s–90s): Going Grid Instead of Perimeter

The key insight of the PCB substrate era was geometric. Lead frames could only place pins around the perimeter of a chip — a one-dimensional ring of connections. PCB-based packages like the BGA (Ball Grid Array) spread solder balls across the entire bottom surface of the package. Suddenly, pin counts went from dozens to hundreds to thousands.

The substrate itself was a miniaturized version of a printed circuit board — layers of fiberglass-reinforced epoxy with copper traces routed between them. This was a genuine paradigm shift: the package was no longer just a container, it was itself a routing layer, a signal manager, a power distributor.

⚠️ Who got reshuffled: Traditional lead frame manufacturers like Mitsui and Shinko had to pivot fast. Those who adapted became substrate makers. Those who didn't shrank into niche markets. The exact same dynamic is building again today around ABF and glass.

Era 3 — ABF Substrate (1990s–Present): The Build-Up Revolution

The PCB substrate era gave way to something more sophisticated in the late 1990s, driven by a single company's materials innovation. Ajinomoto's build-up film allowed substrate makers to add ultra-thin layers one at a time — each layer laser-drilled, copper-plated, and precisely routed before the next was laminated on top.

ABF substrates unlocked a step change in wiring density. Line widths dropped from 50μm to 10μm to 2μm. Via diameters shrank accordingly. The flip chip BGA package, built on ABF substrates, became the universal format for every high-performance processor, GPU, and networking chip for the next 25 years.

• ✓
  Intel's CPU substrates — entirely ABF-based since the Pentium 4 era
• ✓
  NVIDIA's GPU packages — ABF substrates enabling thousands of HBM connections
• ✓
  Apple's M-series chips — advanced ABF substrates with the tightest commercially available line widths

📊 The Scale of ABF Dominance

The global ABF substrate market was worth over $10 billion in 2024. Ajinomoto still holds a near-monopoly on the raw film. The entire ecosystem — equipment makers, substrate manufacturers, chip designers — is optimized around this single material. That's exactly what makes transitions so difficult and so consequential.

The wall ABF is hitting right now

Fig. 2. Every substrate era ends when performance demand crosses the material's physical ceiling — ABF is reaching that crossing point now.

Era 4 — Glass Substrate: The Pattern Repeats

Here's what makes the historical pattern so useful: the trigger for each transition has always been the same. Performance demand — measured in pin count, bandwidth, power density, or signal integrity — grows exponentially. The incumbent material's physical limits are fixed. When the curves cross, the transition becomes inevitable.

For glass substrates, the crossing is happening in slow motion right now. AI accelerators like NVIDIA's Blackwell and future architectures are demanding substrate-level properties that ABF simply cannot deliver: sub-2μm line widths, CTE values below 5 ppm/°C, and dielectric loss figures an order of magnitude lower than organic materials.

The Pattern Across All Four Eras

01 The incumbent material hits a physical wall — not an economic one. The limits are written in physics, not in cost curves.

02 A new material is technically ready years before it ships — ecosystem lock-in, yield immaturity, and supply chain inertia create the delay.

03 The transition reshuffles the industry — incumbents who move early capture the new value chain; those who wait become commoditized or irrelevant.

04 The window is shorter than it looks — once a lead customer commits to the new material, the ecosystem tips fast.

"History doesn't repeat itself, but it rhymes.
In semiconductor packaging, it rhymes every 15 to 20 years — and right on schedule."

Who Gets Reshuffled This Time?

The lead-frame-to-PCB transition created the substrate industry as we know it. The PCB-to-ABF transition handed near-total dominance to a handful of Japanese and Taiwanese substrate makers — and made a seasoning company one of the most strategically important suppliers in all of semiconductors.

The ABF-to-glass transition will be no different in its consequences — only different in its winners and losers.

• ↑
  Glass material suppliers — Corning, AGC, NEG already control the flat glass technology base. They hold the raw material advantage.
• ↑
  New substrate makers with glass capabilities — SKC, and whoever cracks TGV yield first, will define the supply chain.
• ↓
  Pure-play ABF substrate makers — Those without a glass transition roadmap face the same fate as lead frame makers in the 1980s.
• →
  Ajinomoto — The irony: the company whose material defined an era may see that era end. Whether they adapt or cede ground remains the most interesting subplot.

📌 Summary

Sixty years of substrate history follows one pattern: performance demand grows until the incumbent material can't keep up, a new material takes over, and the industry reshuffles around it. Glass isn't a disruption — it's the next step in a sequence that has repeated three times already. The only question is who moves first.

In the next post, we go deep on the single biggest technical obstacle standing between glass substrates and mass production: the Through-Glass Via (TGV). Drilling thousands of precision holes through glass without cracking it — at scale, at yield — is the engineering challenge that will determine who wins this transition.

Glass Substrate ABF Substrate Semiconductor Packaging Lead Frame Paradigm Shift Deep Tech History Advanced Packaging Intel

Next Post · Glass Substrate 03

TGV — Why Drilling Holes Through Glass
Is the Hardest Problem in Packaging

Through-Glass Vias are the single biggest technical barrier to glass substrate mass production. Here's what makes them so difficult — and who's closest to solving it.