High-layer-count AI server boards Market Outlook 2026–2034: Growth Drivers, Key Companies, Strategic Developments & Adoption Forecast — A Meticulous Next™ Foresight Brief

This Meticulous Next™ brief examines the printed circuit boards and package substrates that carry AI accelerators and servers, meaning high-layer-count boards, ABF-based IC substrates and the materials behind them. It tests how far supply of these parts can keep pace with accelerator volumes, and which products and companies gain as package sizes and layer counts rise by 2034.

Each accelerator package needs more substrate than the one before. Ajinomoto describes an AI substrate with about 3.5 times the board area, 18 layers against 6 and roughly 10 times the ABF consumption of a conventional design, and Ibiden plans substrate sizes of 90 mm by 90 mm in 2026, 110 mm by 110 mm in 2028 and 130 mm by 130 mm or larger from 2030. The conclusion is that substrate demand grows faster than accelerator unit volumes, because each unit uses more area and more layers.

Supply is concentrated and tight. Ajinomoto is reported to hold 95 percent or more of the ABF film market and to have run its film capacity at full load in the second quarter of 2026, and it told substrate makers in May 2026 that it would raise ABF prices by about 30 percent from the third quarter. The conclusion is that a single material supplier sets the pace for the whole substrate industry, and substrate makers must pay to secure film.

Substrate makers are responding with capital. Ibiden approved about JPY 500 billion for its electronics business for fiscal 2026 to 2028 on 27 February 2026, and Samsung Electro-Mechanics is reported to be spending USD 1.2 billion on ABF substrates with volume production expected by the third quarter of 2027. The conclusion is that new capacity arrives from 2027, so the tightest period for supply falls before then.

The brief treats AI server boards and substrates as an accelerating signal with an eight year horizon to 2034. It identifies which products and suppliers move first and what each means for substrate makers, material suppliers and buyers. It is a focused 107-page decision brief that presents an indicative trajectory rather than a segmented market model.

An AI accelerator package sits on an IC substrate, a thin laminated board built up from layers of resin film and copper wiring, and that substrate sits on a server board that links accelerators, memory and networking. Both parts need more layers, finer wiring and larger areas than boards for ordinary servers, since accelerators draw more power and move more data.

The material base is narrow. Ajinomoto supplies the build-up film used on most ABF substrates, and Unimicron, Ibiden and Shinko together are reported to hold roughly three quarters of the substrate market. The conclusion is that a small number of companies decide how fast capacity can grow, and a delay at any one of them affects every accelerator maker.

Layer counts and sizes keep rising. Ibiden plans build-up structures of 10-X-10 in 2026, 12-X-12 in 2028 and 14-X-14 from 2030, and Nan Ya Printed Circuit Board plans boards of 24 layers in 2026 and more than 24 layers in the first half of 2027. The conclusion is that yield at high layer counts becomes the main competitive advantage, since each added layer raises the chance of scrap.

Organic substrates lose flatness at large sizes, which opens the way for glass. Intel showed a package that pairs embedded bridges with glass at NEPCON Japan in January 2026 and places commercial glass deployment in the latter half of the decade, and Ibiden puts glass core on its roadmap around 2030. The conclusion is that glass-core substrates enter after the current shortage and extend the growth path beyond 2030.

The limits are shortages of build-up film and glass cloth, yield loss at high layer counts, the capital needed for new lines and the time to qualify with accelerator makers. The brief tests each limit against the cost trajectory to 2034.

The AI server board and substrate market, covering high-layer-count server boards, accelerator baseboards and ABF-based IC substrates for AI accelerators and AI server processors, and excluding boards and substrates for consumer devices, automotive and general-purpose servers, laminate and film materials sold separately, and assembly services, is valued at USD 20.0–28.0 billion in 2026 and is projected to reach USD 61.18–85.65 billion by 2034, at a CAGR of 15%.

The 2026 base rests on high-layer-count boards and ABF substrates sold for AI servers, mainly from Taiwan, Japan, China, South Korea and Austria. Growth then follows accelerator volumes, package size and layer count, and the arrival of glass-core substrates.

The forecast is sensitive to accelerator shipments, the pace at which film and glass cloth capacity expands and the yield of new high-layer-count lines.

The base case assumes new substrate and film capacity arrives from 2027 and larger packages reach volume by 2028, bringing the inflection to about 2028 and a 2034 value near the middle of the range. An accelerated case assumes glass-core substrates and panel formats scale from 2027, bringing the inflection to 2027 and lifting the 2034 value to the top of the range. A delayed case assumes film and glass cloth shortages persist and yields at high layer counts stay low, moving the inflection to 2030 and holding growth at the bottom of the range.

Growth drivers

• AI accelerator volumes.
• Board and substrate content per server.
• Package size growth.
• Higher speed interconnects.

Enablers

• Long-term supply agreements.
• Finer line and space processes.
• Public incentives.
• Capacity expansion.

Restraints and barriers

• Film and glass cloth shortages.
• Yield loss at high layer counts.
• Capital intensity.
• Skills.

The direction of AI server boards and substrates is set by accelerator shipments, package size growth, layer count growth, limits on materials and the pace of capital spending on substrates.

The key participants in AI server boards and substrates include Ibiden Co., Ltd., Unimicron Technology Corp., Shinko Electric Industries Co., Ltd., AT&S Austria Technologie & Systemtechnik Aktiengesellschaft, Nan Ya Printed Circuit Board Corporation, Kinsus Interconnect Technology Corp., Samsung Electro-Mechanics Co., Ltd., Ajinomoto Co., Inc., Sekisui Chemical Co., Ltd., Nitto Boseki Co., Ltd., Victory Giant Technology (Huizhou) Co., Ltd., Shennan Circuits Co., Ltd., TTM Technologies, Inc., Zhen Ding Technology Holding Limited, Compeq Manufacturing Co., Ltd., Gold Circuit Electronics Ltd., Elite Material Co., Ltd., Tripod Technology Corporation, SKC Co., Ltd., Panasonic Holdings Corporation, NVIDIA Corporation, Advanced Micro Devices, Inc., Broadcom Inc., Marvell Technology, Inc., Intel Corporation, Taiwan Semiconductor Manufacturing Company Limited, Samsung Electronics Co., Ltd., SK hynix Inc., Micron Technology, Inc., ASE Technology Holding Co., Ltd., Amkor Technology, Inc., Hon Hai Precision Industry Co., Ltd., Quanta Computer Inc., Wistron Corporation, Inventec Corporation, Dell Technologies Inc., Super Micro Computer, Inc., Hewlett Packard Enterprise Company, Lenovo Group Limited, Amazon.com, Inc., Alphabet Inc., Microsoft Corporation, Meta Platforms, Inc., Corning Incorporated, Taiwan Printed Circuit Association, IPC and SEMI.

The brief profiles representative participants in each archetype and assesses which are positioned to lead as packages grow larger and layer counts rise.

Ibiden, Unimicron, Shinko Electric Industries, AT&S and Nan Ya Printed Circuit Board lead among substrate makers, Victory Giant, Shennan Circuits, TTM Technologies and Zhen Ding among server board makers, and Ajinomoto, Sekisui Chemical and Nitto Boseki among material suppliers. Competitive intensity is moderate in substrates, where few suppliers qualify, and rising in boards.

Where value migrates.

In 2026 value sits with substrate makers and the film supplier, because film and substrate capacity decide how many accelerators can ship. By 2028 it moves toward server board makers and larger substrate formats as new capacity arrives and layer counts rise. By 2034 it settles with glass-core and panel substrates for the largest packages, where material and process knowledge are scarce. Margins hold for suppliers with committed capacity and high-yield lines, and narrow for suppliers of standard boards.

The signals converge on a market where supply, not demand, sets the pace. A 30 percent film price rise, full capacity at the main film supplier and approved investments by Ibiden and others support growth, while film and glass cloth shortages, yield loss at high layer counts and capital intensity hold back output.

Structural limits show where progress could slow: the capacity of the film supplier, the yield of lines at 18 layers and above, the pace of glass-core qualification and the time new plants take to reach volume. The brief therefore tracks four indicators: substrate capacity additions, build-up film prices, layer counts in new accelerator packages and glass-core pilot orders.

Tight ABF substrate supply first, then larger substrates and new material capacity, then glass-core adoption is the expected order of change.