The Next AI Chip Race Is About How Chips Connect

Applied Materials will lift its stake in Dutch equipment maker BESI from about 9% to 20%. The bet is less about control than about a new way to link chips.

Why would a chip equipment maker buy a stake in another one?

Applied Materials (AMAT), the US semiconductor equipment maker, announced on Oct. 26 at 4:30 p.m. ET an agreement to raise its stake in the Netherlands' BE Semiconductor Industries (BESI) to about 20%.

The two companies are already close. They have co-developed hybrid bonding technology since 2020, and AMAT first bought a 9% stake in BESI in April 2025. The new plan adds roughly 11 percentage points.

AMAT said it has no intention of seeking more than 20% or a seat on BESI's board. The announcement is therefore better read as an investment to deepen a long-term partnership with a key technology partner than as a step toward a full takeover.

Why shrinking chips is not enough

Process shrinks have long been the center of chip performance gains. Smaller transistors and higher density improve compute performance and power efficiency.

For AI workloads, though, moving data matters as much as computing speed inside the chip. However fast a GPU calculates, overall performance is capped if memory cannot feed it data quickly enough.

That is why advanced packaging has become important. It combines several semiconductor dies into one system and shortens the connections between them to move data more efficiently.

High-bandwidth memory (HBM) stacks multiple memory dies vertically to deliver high data bandwidth. Hybrid bonding is one technology that can be used for such dense chip connections. Not every HBM product is made with hybrid bonding today, however.

How does hybrid bonding work?

Some existing stacking methods use bumps, tiny metal protrusions, to make electrical connections between chips.

Hybrid bonding differs. It joins the insulating layer and the copper contacts on the chip surfaces directly.

Chip-to-chip connection methods compared

  • Bump connection: connection through metal bumps
  • Hybrid bonding: direct bonding of insulating layers and copper contacts

The original diagram is a concept illustration. It does not show actual contact size or spacing to scale.

The process is precise. The surfaces to be bonded are flattened and cleaned, and the copper contacts are aligned exactly. The insulating surfaces are then joined, and heat treatment completes the copper bond.

The method could allow tighter contact spacing and shorter wiring. That could improve connection density, data transfer efficiency and power efficiency.

The tradeoff is that tiny particles on the surface, flatness errors and misalignment can hurt yield. Equipment precision and process control therefore matter more.

What are AMAT and BESI developing together?

Their cooperation now goes beyond hybrid bonding alone.

Development areaWhat it means
Hybrid bondingDense, direct connection between chips
3D ICVertical integration of multiple chips
Logic-memory integrationBetter connection between compute chips and memory
Thermocompression bonding (TCB)Advancing the existing method of bonding with heat and pressure
Next-generation packagingNew manufacturing methods that attach dies to wafers, other dies and panels

AMAT is strong in materials processes such as deposition, polishing, cleaning and inspection. BESI specializes in precision chip placement and bonding equipment.

Combined, these strengths could let customers adopt a manufacturing process whose pieces are optimized for each other, rather than buying individual tools. The core of the partnership is turning technology development into volume production.

What matters for chip equipment investors?

It is hard to say AI chip competition is moving entirely from process shrinks to packaging. The change looks more like both becoming important at the same time.

The deal suggests that companies like AMAT, strong in conventional wafer fabrication equipment, could also find business opportunities in advanced packaging. For BESI, wider adoption of hybrid bonding is a growth opportunity.

Some things remain to be confirmed. Even if a customer adopts the technology in a next-generation chip design, actual equipment orders, yield stabilization and volume ramp-up can take time. US antitrust review has also not been completed.

The equipment industry is also sensitive to customers' capital spending cycles. Even with a good long-term technology outlook, near-term orders and earnings can swing.

Applied Materials' move to about 20% of BESI is best read as a long-term partnership in hybrid bonding, not a takeover.

Insight Times Editorial Desk

More by topic

Starting with this story's topics, then what readers are following now

Investor psychology, decision-making and books on investing

HBM

This story's topic65 storiesSee all →

Memory stacked next to the GPU.

Has Memory Peaked? Prices Are Cooling, but the Shortage Looks Set to Run to 2028

Price growth has slowed and consumer demand is already breaking. Server and HBM supply remains short, and Micron says more than 75% of its 2027 volume is already committed to customers.

Tech

AI Chip Peak Check: Demand Still Rising, Risk Building in 2027-2029 Supply and Funding

DRAM and NAND contract prices are still climbing and HBM suppliers hold the edge. Power shortages, AI financing costs and simultaneous memory expansion are the risks for later years.

Investing

US AI Data Centers Are Headed for 90GW. Gigawatts Now Matter More Than Dollars

Goldman Sachs expects US data center capacity to grow from 64GW at the end of 2026 to 90GW at the end of 2027. The AI race is moving from buying GPUs to building power plants and transmission lines alongside them.

Tech

How Far Can China Push DRAM Without EUV

CXMT says it has shrunk DRAM features to 11.95nm using quadruple patterning instead of EUV lithography. The real question is not whether China caught up, but how much more it now costs China to get there.

Tech

Memory Chip Supercycle: Not at Peak Yet, But 5 of 6 Warning Signs Are Flashing

Memory prices are still climbing, but the pace is slowing and PC and smartphone demand is starting to crack under the weight of higher chip costs. This looks less like a price peak and more like a profit plateau building toward the next supply cycle.

Investing

The Real ASML Number Isn't the 3% Pop, It's the 30% Capacity Hike

ASML's rally isn't just an AI trade. Customers are already locking in 2027 tool slots, and ASML is raising Low-NA EUV capacity by roughly 30% to keep up.

Investing

2027 Isn't the End: Citi's Case for a Memory Supercycle Running to 2031

Citi says the memory shortage timeline is stretching again. If AI shifts from "train once" models to continual learning, HBM, server DRAM and enterprise SSDs could all fall short at the same time.

Tech

GPU

This story's topic53 storiesSee all →

A chip built to draw graphics, whose ability to run thousands of identical cal…

Why Nvidia Is Happy to Help Make AI Models Nearly Free

Falling model prices can squeeze model makers. For Nvidia, the number that matters is the total volume of tokens generated worldwide and the compute that produces them.

Tech

The US-China Summit Didn't Erase Taiwan Risk for Chip Stocks

Washington's Taiwan policy held, but roughly $14 billion in arms sales is still on hold. For AI chip investors, the summit reads less like risk removed and more like a fragile balance extended.

Macro

Amazon Goes After Nvidia's Strongest Moat, and It Is Software, Not GPUs

Amazon's open-source TorchNeuron aims to let existing AI code run on its Trainium chips. The contest comes down to the total cost of AI compute and the return on Amazon's $220 billion in annual capital spending.

Tech

Nvidia Put $2 Billion Into Lumentum. Optics May Be the Next Bottleneck, but the Stock Has Already Run

Optical networking is emerging as the next constraint on AI data centers. Investors need to separate the industry's growth potential from the price they are already paying for it.

Tech

Nvidia's Next Target After GPUs: Controlling What AI Agents Are Allowed to Do

OpenShell and Sentry are not about policing what an AI says. They are about controlling, at the server infrastructure level, which files and APIs an AI can touch, what it can execute, and who can actually pull the plug when something goes wrong.

Tech

Nvidia's $500 Billion Experiment: Wall Street Asks Who Backs the GPU, Not Just the Chip

The real issue in Nvidia's financing platform is not chip sales. It is how far Wall Street will trust GPU resale prices and rental income three to four years out.

Tech

AMD Crossed $1 Trillion, and the Market Started Looking at CPUs Again

AMD's stock topped a $1 trillion market cap for the first time. The rally isn't just about being an Nvidia alternative. It reflects a reassessment of AMD as a full-stack AI infrastructure vendor, one that bundles CPUs, GPUs and networking, as AI shifts from chatbots to agents and reignites data center demand for CPUs.</deck> <parameter name="oneline">As AI moves from chatbots to agents, the market has started counting the CPU sitting next to the GPU again.

Tech

Semiconductors

This story's topic50 storiesSee all →

The whole business of designing, fabricating and testing chips.

Musk's Real AI Strategy Isn't the Model, It's a Chips-to-Space Physical Stack

On the All-In Summit stage, Elon Musk and Gwynne Shotwell laid out AI peer review, Terafab, Starship, and orbital data centers as if they were separate topics. Strung together, they reveal Musk betting the AI race is decided less by who builds the smartest model and more by who can deploy intelligence at the lowest cost and largest scale.

Tech

Lam Research Falls 12%: The Problem Is the Timing of AI Growth, Not Its End

Revenue up 54%, an operating margin above 40% and $3.2 billion in free cash flow were not enough. The selloff likely reflects worry that revenue arrives later than hoped.

Investing

Stocks Bounce, but the AI Selloff May Not Be Over: A Two-Week Playbook

The Dow rose more than 1% while the Nasdaq barely moved and semiconductors kept falling. A W-shaped retest looks more likely than a V, and the task now is surviving the next drop.

Investing

ASML Is Discussing 2030 Orders. How Long Can the AI Chip Supercycle Last?

Despite bubble worries, ASML sees equipment shortages through 2028. Investors should watch future production capacity, not just today's orders.

Tech

Samsung Posts 107 Trillion Won Profit, TSMC Sales Jump 50%. Why Did the Shares Fall?

The AI chip supercycle has not ended. But the market's question has changed: not how fast earnings grow, but how long today's enormous margins can last.

Tech

S&P 500 Slips 0.22%, but 73% of the Market Fell

Long-term yields above 5% weighed on small caps and semiconductors on Oct. 7. A strong 10-year Treasury auction and a few large tech names limited the index's losses.

Tech

White House Swaps AI for "SI." The Real Investment Signal Is Infrastructure, Not the Name

An executive order renames artificial intelligence "Super Intelligence" and six frontier labs sign a voluntary safety accord. What investors should watch is the data center, power, nuclear and chip policy now gathered under one national-competitiveness frame.

Tech

A building that houses servers and networking gear and supplies them with powe…

Tesla Energy's 13.7GWh Paradox: Expectations Ran Faster Than Deployments

Q2 deployments recovered but margins collapsed. Q3 set a record for deployments and still missed estimates. The issue is not the size of AI power demand but how fast it converts into Tesla revenue and profit.

Tech

Why the White House Swapped "AI" for "Super Intelligence"

Executive Order 14434 renames AI across the federal government without defining a new technology. The real story is about roughly $700 billion in capex, free cash flow and electricity.

Tech

AI Needs to Find $1 Trillion a Year From Somewhere to Pay Back Its Capex

Saying AI capex is big is no longer the point. The harder question is what end users have to pay for the buildout to pencil out, and Goldman Sachs puts that number at roughly $1 trillion a year.

Tech

The AI Bubble Question Has Changed: By Late 2026, It's ROIC Over GPUs

There is no sign yet that AI investment is slowing. If anything, hyperscaler spending forecasts keep climbing. But as the scale of that capital has grown, the market's question has shifted from "is AI growing" to "when, and how much, does this capital come back as cash."

Tech

2027 Investing Needs Alignment, Not Prediction

A new investing book argues the hard part was never picking the next winner. It's staying aligned with where value and bottlenecks are moving while keeping old human weaknesses in check.

Investor Mind

If GPUs Are the Muscle, the Network Is the Nervous System

AI data centers do not get faster just by buying more GPUs. When thousands of accelerators cannot exchange data on time, the most expensive asset in the building sits idle. That is the whole investment case for Arista.

Investing

What Jensen Huang sold at the G20 was not GPUs. It was national infrastructure

Nvidia's message is that AI should be laid down like electricity, roads and the internet. After the hyperscalers, the buyers are governments, telcos, regional clouds and industrial data centers.

Tech