To The Moon

[你懂的]   $Arteris (AIP): The Company Building the “Highways” Inside AI Chips

Let’s start with the simplest explanation:

Arteris doesn’t manufacture chips. It provides the IP that helps different parts of a chip communicate with each other.

Modern AI chips can contain CPUs, GPUs/NPUs, memory controllers, caches, accelerators, security blocks, and I/O interfaces — all of which need to move massive amounts of data.

As chips become more complex, especially with the rise of Chiplets and multi-die architectures, moving data efficiently inside the chip becomes a major engineering challenge.

That is where Arteris comes in.

Its Network-on-Chip (NoC) technology can essentially be viewed as the highway system inside a chip, helping different IP blocks communicate efficiently while balancing performance, power consumption, security, and reliability.

1️⃣ How does AIP actually make money?

This is one of the most interesting parts of the story.

Arteris is not a traditional hardware company. It doesn't need to build fabs or spend billions manufacturing chips.

Its business model is mainly built around:

① IP Licensing

Customers pay to use Arteris' semiconductor IP.

② Support & Maintenance

Customers can continue paying for technical support and maintenance, creating a recurring component of revenue.

③ Variable Royalties

This is the part I pay the most attention to.

When a customer designs a chip using Arteris' IP and eventually puts that chip into production, Arteris can receive royalties based on chip shipments.

The potential flywheel looks like this:

Design win → Chip production → Chip shipments → Royalties

That creates an attractive operating model.

In 2025, Arteris' variable royalties reached $6.6 million, up 50% year over year. By Q1 2026, trailing-twelve-month variable royalties had increased to approximately $7.9 million, up 67%.

So I don't think investors should only look at AIP's current revenue.

The bigger question is:

«Can today's design wins become tomorrow's chip shipments and royalty revenue?»

2️⃣ Is AIP actually profitable?

Here is where I want to be very clear:

Not yet — at least not on a sustainable GAAP basis.

For 2025:

• Revenue: $70.6M

• Net Loss: $34.7M

• GAAP Operating Loss: $33.1M

• Non-GAAP Operating Loss: $12.5M

• Non-GAAP Free Cash Flow: +$5.3M

So if you ask:

«“Is AIP already a highly profitable company?”»

The answer is no.

But there is an important trend.

Revenue growth is accelerating while losses are narrowing.

2025 revenue grew 22%.

Then in Q1 2026:

Revenue reached $22.9M, up 39% YoY.

Non-GAAP operating loss improved from $3.2M to $2.5M, while non-GAAP net loss narrowed from $3.59M to just $1.18M.

That creates an interesting potential path:

Revenue ↑

Gross Profit ↑

Operating Loss ↓

→ Potential profitability

That is where the potential valuation re-rating could come from.

3️⃣ Why am I watching RPO?

RPO stands for Remaining Performance Obligations.

In simple terms, it represents contracted business that has not yet been fully recognized as revenue.

At the end of 2025, AIP's RPO reached approximately $116.8M, up 32% YoY.

By Q1 2026, RPO had increased further to approximately $118.3M, up 33% YoY.

This is important because revenue tells us what has already happened.

RPO gives us visibility into part of what could become future revenue.

4️⃣ Another number I want to watch: ACV + Royalties

Arteris also tracks Annual Contract Value plus royalties.

In 2025:

$83.6M, up 28% YoY.

In Q1 2026:

$92.8M, up 39% YoY.

If that growth can remain around the 30–40% range while royalties continue growing rapidly, the economics of the business could become increasingly attractive.

5️⃣ Why could AI be a major catalyst?

This is where the story gets interesting.

The evolution is roughly:

CPU → SoC

Then:

CPU + GPU/NPU + Memory + Accelerator + I/O

And increasingly:

Chiplet + Multi-Die + AI Accelerators

As chips become more complicated, data movement becomes more complicated.

That is Arteris' opportunity.

In September, Arteris launched FlexGen Multi-Die, targeting data movement challenges in multi-die and Chiplet architectures.

The company is positioning itself around the industry's transition from increasingly complex monolithic SoCs toward Chiplet-based architectures, driven in part by growing AI compute requirements.

Arteris has also been working with IC-Link, part of imec, on next-generation AI/HPC Chiplet and ASIC development.

So this isn't simply:

«“AI is popular, therefore AIP should benefit.”»

The more interesting thesis is:

AI chips are becoming more complex → Chiplet adoption increases → Data movement becomes harder → Demand for advanced interconnect IP potentially increases.

6️⃣ And AIP isn't only an AI story

This is another part I like.

Arteris' technology can potentially serve multiple markets:

🚗 Automotive SoCs

🤖 AI / Robotics

🖥️ Data Centers

📡 Communications

🏭 Industrial applications

🛡️ Aerospace & Defense

📱 Consumer electronics

For example, SiEngine selected Arteris FlexNoC for next-generation automotive SoCs covering applications such as intelligent cockpits and ADAS.

So the thesis isn't simply:

«“AI slows down → AIP is finished.”»

Instead:

«AI may be one of the biggest growth catalysts, while the underlying technology can serve a much broader semiconductor market.»

7️⃣ Why could AIP have significant upside?

I don't think the thesis should simply be:

“AI = stock goes up.”

The more interesting potential catalyst is:

Revenue growth + Royalty growth + Operating leverage + Potential profitability + Valuation re-rating

Imagine a future scenario where:

Revenue:

$90M → $120M → $160M

Royalties:

$8M → $15M → $25M

And operating results move from:

-$10M → -$2M → +$15M

At that point, the market could potentially start viewing AIP very differently.

Instead of:

«“A small semiconductor IP company that is losing money”»

It could become:

«A high-margin semiconductor infrastructure company with recurring royalties, AI/Chiplet exposure, and expanding operating leverage.»

That is the potential fundamental re-rating I am watching.

Not a prediction that AIP will definitely explode.

Rather, the company has a potential pathway where improving fundamentals could eventually justify a very different valuation.

8️⃣ But what are the biggest risks?

This is absolutely not a stock where I would only discuss the upside.

❌ 1. It is still losing money

GAAP profitability has not been established.

❌ 2. The company is still relatively small

2025 revenue was only around $70.6M, nowhere near the scale of semiconductor IP giants such as Synopsys or Cadence.

❌ 3. Customer concentration

If a major customer delays a project, revenue recognition and future royalties could be affected.

❌ 4. Design wins take time

A design start does not mean revenue arrives tomorrow.

Some semiconductor designs can take years to reach mass production.

So:

Design Start ↑ ≠ Revenue immediately ↑

The chain I want to see is:

Design Starts → RPO → Revenue → Royalties

The key question is whether that entire chain continues to work.

9️⃣ One more interesting development

In January 2026, Arteris acquired Cycuity, a semiconductor security verification company.

The total consideration was up to approximately $45M, including cash, stock, and performance-based consideration.

Why does this matter?

As Chiplet architectures become more complicated, the challenge isn't only:

«“How do we move data faster?”»

It increasingly becomes:

«“How do we move data securely?”»

That gives Arteris another potential expansion path:

Data Movement + Integration + Security

If the acquisition is successfully integrated, it could expand Arteris' addressable market over time.

---

So, how do I view AIP?

Company quality:

Still a growth-stage company, not a proven profit machine.

Business momentum:

Interesting. Growth accelerated significantly in 2026.

Business model:

One of the strongest parts of the thesis — Licensing + Recurring Revenue + Royalties.

AI exposure:

Not another GPU story. It is an infrastructure play on the increasing complexity of AI chips.

Chiplet exposure:

Potentially one of the most interesting long-term growth drivers.

Risk:

High. Losses, small scale, customer concentration, valuation, execution, and the long time between design wins and mass production all matter.

So I would describe AIP as:

«A high-risk semiconductor growth company that is showing improving fundamentals, but has not yet fully proven itself.»

Not “the next NVDA.”

What I really want to see over the next few quarters is:

Revenue growth remaining strong

Royalties continuing to accelerate

RPO continuing to grow

Non-GAAP losses shrinking rapidly

Eventually, GAAP profitability

If those five things start moving in the same direction, the AIP story becomes much more interesting.

I'm not watching for the stock to simply “go up.”

I'm watching for the business fundamentals to catch up with the story.

That, to me, is where the real opportunity — and the real risk — lies.[思考]