3 Take-Aways from Huawei’s Innovation and IP Forum (Tech Strategy)

Posted on by

This week Huawei held its sixth Innovation and Intellectual Property (IP) Forum. This is Huawei’s big event on IP protection, patents, standards, published papers and other activities that are critical for open source and collaborative innovation.

Why is this so important?

Because this is how innovation goes from thousands of people at one company to hundreds of thousands across lots of companies, universities and institutions. Working collaboratively.

Openness and IP protection are what enable super-charged, industry-wide innovation.

Here are my three take-aways from the forum.

Take-Away 1: Huawei Is Really Good at Innovation at Scale. And They Just Doubled their R&D Spending.

In 2024:

  • Huawei published 37,000 patents.
  • They published over 1,000 academic papers.
  • They submitted more than 10,000 technical contributions to standards organizations.
  • Their source code for OpenHarmony (their device operating system) grew by 10 million lines. And had submissions from over 8,100 community contributors.
  • openEuler (their server operating system) exceeded 10 million installations.

The engine of all this is Huawei’s 104,000 employees working in R&D That’s out of 204,000 employees in total. The R&D (and innovation) activities at Huawei are very large.

And many of them work at Huawei’s Europe Town R&D facility in Dongguan. Which is pretty great.

But it wasn’t always this way.

Compare all that to Huawei’s original office in Shenzhen in 1987.

In these early years, Huawei was mostly selling simple PBX switches to hotels and other smaller Chinese companies.

But even then, Huawei was putting about 10% of their small revenue into R&D. That is how they started building their own products and technologies. And that tradition has continued for +35 years.

Founder Ren Zhengfei’s reason for was very logical. Telecommunication equipment and ICT products all become obsolete. They have a short lifecycle. So new products have to be continually created and re-created. And therefore, the only asset Huawei really has is its people. And how you organize them to innovate at scale.

“Resources can be exhausted; only culture endures. Huawei does not have any natural resources to depend upon. What we do have is the brainpower of our employees. This is our oil, our forest, and our coal. Human ingenuity is the creator of all wealth.” – Ren Zhengfei

So, when Huawei talks about its annual R&D spending, it is really talking about investing in its people.

That was the story of R&D at Huawei for much of its existence. But in the past five years, their approach has changed.

  • First, the R&D spending has doubled. For the past four years, R&D at Huawei has been +20% of revenue. In 2024, total R&D spending reached $25 billion.
  • Second, Huawei has moved from mostly a technology follower to mostly technology leader. They followed in 4G. But they were a leader in 5G. They are now pioneering entirely new technologies. And this is a different type of research. It is about breakthroughs in science and technology. And it required a cultural change from Huawei’s historically hard-charging sales-driven culture.

So, hold on to your hat. Huawei is going to be a powerhouse in tech innovation over the next 10 years. Think 100,000 engineers and $250B in cumulative R&D spending.

Take-Away 2: IP, Open Standards and Open Source Are the Engine of Global Innovation

There’s a problem with everything I just said.

A big limitation.

Innovation by a single large company can never match collaborative innovation by lots of parties – such as big companies, small companies, developers, businesses, universities, and government bureaus.

Innovation at truly global scale only happens through collaboration. And that is tricky to make happen in practice.

And it’s not just about the scale and pace of innovation. Collaboration is especially necessary for new products and technologies where lots of new components have to be built at the same time. And they all have to work together. The launch of EVs was only possible because the automakers, battery makers, magnet companies, software companies, government officials and other parties all coordinated.

So how does such collaboration happen in practice?

It happens by:

  • Filing patents. Which are then licensed to others.
  • Respecting and protecting intellectual property (IP).
  • Agreeing on and making contributions to common standards.
  • Publishing and contributing to academic papers, which are circulated.
  • Supporting open-source software (and hardware), which everyone can build upon.

Basically, this type of innovation depends on all the things Huawei Innovation and IP Forum is about.

Here are some factoids and announcements from the event.

For patents:

  • In 2024, Huawei published +37,000 patents, a new record for the company.
  • Huawei now has more than +150,000 active patents worldwide.
    • This includes the largest number of granted patents in China (over 50,000 valid patents).
    • This includes has +29,000 granted patents in the United States and +19,000 granted patents in Europe.
  • In the past three years, Huawei has signed or renewed cross-licensing agreements with Nokia, Ericsson, Amazon, Samsung, Sharp and others.
  • In 2024, Huawei had 6,600 published Patent Cooperation Treaty (PCT) applications.

For global standards:

  • In 2024, Huawei submitted +10,000 technical contributions to standard setting organizations (in China and internationally). They have submitted + 140,000 total.

For academic papers:

  • In 2024, Huawei published + 1,000 academic papers.
    For open source:
  • In 2024, openEuler (Huawei’s open-source server operating system) had +10 million installations, 1,956 openEuler community partners and + 21,000 developers.
    • openEuler had +3.85 million downloads worldwide, covering 155 countries and regions and + 2,000 cities.
  • In 2024, OpenHarmony (Huawei’s open-source device operating system) grew by +10 million lines, with submissions from Huawei and +8,100 contributors.
    • Note: OpenHarmony 6.0 was released on September 27, 2025, and + 1 billion devices have used HarmonyOS so far.

Take-Away 3: The Trifold Smartphone is Really Cool. But the AI Supercluster is Huawei’s Big Leap Forward.

We are now seeing an accelerating stream of engineering innovations coming out of Huawei. In the consumer division, we have seen new smartwatches, earbuds, foldable laptops, and even cars. The new Maextro s800 EV is pretty great.

And in September 2024, Huawei announced its Mate XT smartphone, also called the trifold. And it’s fantastic. I’m not even much of a smartphone fan. But it’s really cool.

The three screens can fold out into three configurations. This really makes a difference in video and mapping.

I won’t go through this as I’m not a smartphone analyst. But some of the engineering breakthroughs required for the trifold included the hinges, the thin and light batteries, and the foldable screen.

It’s cool.

But the most important advances (in my opinion) are in Huawei’s new AI infrastructure, which is mostly centered on their AI data centers.

There is currently a global race to create the new infrastructure for Generative AI, intelligence and AI agents. This has resulted in a massive build out of data centers around the world. It’s a crazy level of spending. Except these aren’t really data centers. These are tightly connected data centers that operate as giant integrated computers. They are the digital infrastructure of intelligence.

And this is where Huawei has been making important contributions. Especially with their CloudMatrix 384, which was further unveiled in the last three months. Here is one of them.

Screenshot

The engine of the CloudMatrix is an AI compute system that combines 384 Ascend 910c chips and 192 Kunpeng CPUs. Huawei’s Ascend chip series are for AI training and inference. Their Kunpeng series is for general purpose computing. Plus, there are networking chips.

What the CloudMatrix does it put these together into an all-to-all architecture. They all connect to each other dynamically.

  • The semiconductors are grouped into pods. Which are tightly interconnected.
  • The pods are grouped into clusters. Which are tightly interconnected.
  • Software orchestration happens at the pod and cluster level (not rack level). So, they work dynamically.

Technically, the CloudMatrix is a pod made up of 16 racks. Each rack has 24 Ascend 910C chips, so a pod has 384 in total. Hence CloudMatrix 384.

What is the big innovation?

It’s the super pods and superclusters.

It is a brute-force scaling approach. They are attempting to compensate for less powerful chips with system-level innovations. You use a lot more chips, more networking, and energy required but the performance should be equivalent. It’s a new paradigm for AI infrastructure.

And this is just the beginning. As mentioned, Huawei’s R&D spending is continuing at historic levels.

Finally, keep in mind, Huawei is an end-to-end digital and AI infrastructure provider. That is pretty unique. They provide cloud services, specialize in connectivity, and they have edge devices.

They can offer end-to-end integrated solutions. Currently for digital solutions and increasingly for AI solutions. This can be quite powerful in projects that require end-to-end-control, like automating ports and factories. Where you need everything from the robots that unload ships to the private network to the AI in the cloud.

***

Ok. those are my 3 take-aways from the forum.

It’s fascinating stuff. Collaborative innovation is really powerful. And things like standards, open source and IP protection are the key to it.

Cheers, Jeff

———

Related articles:

From the Concept Library, concepts for this article are:

  • n/a

From the Company Library, companies for this article are:

  • Huawei

——

Leave a Reply