Taiwan Takes Center Stage at TechCrunch Disrupt 2026 as AI Moves Into the Physical World

Taiwan has made its mark on the TechCrunch Disrupt 2026 stage, putting the island in the spotlight as artificial intelligence ventures into the physical realm. In an era where Physical AI is all the rage in Silicon Valley and the next big draw for investors, Taiwan was there to make the case that the coming AI revolution will be about much more than software.

San Francisco | October 13, 2026 | TechCrunch Disrupt 2026

Silicon Valley has long been the place to dream up and fund what is to come in technology; Taiwan has been where those dreams are put to work. Be it semiconductors, AI servers or the smartphones in your pocket, the island has been instrumental in turning a breakthrough into a product for the world. Yet with AI turning over a new leaf, the old dynamic is shifting.

No longer just the manufacturing arm of Silicon Valley's latest ideas, Taiwan is fast becoming a magnet for venture capital and a home for deep tech startups, a strategic partner in the rise of the next cohort of AI firms.

The proof of this was on display in San Francisco on Oct. 13 at TechCrunch Disrupt, the world's most influential early stage startup conference. At 1:30 PM on the Breakout Stage at Moscone West, a contingent from Taiwan's ITRI, Phison Electronics and two of the leading Taiwanese startups took part in a 50 minute stage session, "Taiwan Powering the Next AI Revolution: Connect, Invest & Scale." Sparknify, a leading Silicon Valley based PR firm was behind the scenes coordinating and putting these technology leaders not only in front of the TechCrunch audience, but more importantly, on the official TechCrunch agenda this year.

With its international startup programs and a corporate venture capital network on the grow, ITRI has shown how it forges links between Silicon Valley’s investors and founders and Taiwan’s industrial heavyweights, strategic capital and technology ecosystem. In so doing the organisation is opening up new ways to put money behind, work with and scale innovations on a worldwide basis. ITRI also put the spotlight on Taiwan’s unique capabilities in semiconductors and advanced manufacturing as well as its world class supply chains to make the case for the island as an ideal place for global investment in Physical AI, intelligent hardware and deep tech startups of the future.

Phison has taken the chance to put its AI platform and plans for strategic investment on display, in an effort to link the ingenuity of Silicon Valley with Taiwan’s manufacturing prowess. In doing so the company has also made its mark on one of the most consequential early stage startup venues in the world. Phison is using the occasion to put its growing corporate venture capital arm before a new breed of founders, making it known that it will be investing in and fostering startups with potential even as it pushes for wider uptake of its technologies in the Physical AI ecosystem that is taking shape.

TechCrunch Disrupt 2026 has broken with tradition to put forward a Real World AI stage for the first time ever. The move is a nod to the fast rising significance of robotics, autonomous systems, intelligent manufacturing and the point where artificial intelligence and physical hardware meet. It is a landmark change that speaks to a reordering of what Silicon Valley deems worthy of its innovation and investment. With the coming round of AI advances set to go beyond data centres and software into the machines and factories that run the physical world, the shift is well under way.

AI Is Leaving the Cloud. Taiwan Is Ready.

OpenAI, Anthropic and NVIDIA have been the drivers of the recent boom in foundation models and the data centers to back them up. But the conversation is now being had around what occurs when AI has to deal with the real world. Put an algorithm in charge of a robot or an industrial process and the problem set changes entirely. You need more than code; you need the right semiconductors, sensors and a supply chain you can count on. Not to mention dealing with the realities of cost, latency and privacy at scale.

This is the territory of Physical AI, and few are as well placed as Taiwan to handle the transition. The island already makes some of the world's best chips and the bulk of the AI server infrastructure. During the session, figures were put to the table showing that some 90 per cent of global AI server shipments come from Taiwan's supply chain.

But as AI makes its way into vehicles and factories, the industry will have to rely on Taiwan to marry software with physical systems. The next wave of AI will not only be written, it will be built and scaled. That is Taiwan's role to play. BoGaun Chen, a senior leader at ITRI International and the institute’s Deputy International Affairs Officer, set the tone for the main presentation with an overview of Taiwan’s technological evolution and what it means for international investors.

In his talk, Aggregated Intelligence in Taiwan, Chen put forward the case for why Taiwan has become one of the premier technology ecosystems on the planet. He pointed to ITRI as the linchpin of that story. Since its founding in 1973, the Industrial Technology Research Institute has been instrumental in turning a manufacturing economy into a global tech power. ITRI was there at the beginning of the semiconductor industry, helping to put in place the foundations for firms like TSMC and UMC. Over the years the institute has done more than just facilitate research; it has been the link between science and commercialization, between venture creation and the private sector.

The numbers tell part of the tale: by the first quarter of 2026 ITRI had 174 startups and new business units under its belt, alongside 224 incubated companies. And while Chen made sure to underscore Taiwan’s industrial might in areas from AI infrastructure to advanced packaging and electronic components, his point was not to dwell on the past. He wanted to show where the next round of investment is heading.

ITRI has singled out three avenues for the next wave of AI innovation in Taiwan. There is the matter of AI Infrastructure, the chips and silicon photonics that will underpin more complex computing. Then you have Edge and Intelligent Hardware, which puts on-device computing and embedded systems right where the data is being made. And finally, Physical and Industrial AI, where robotics and digital twins are changing how operations run in the real world.

This should change the way investors see the country. It is no longer enough to view Taiwan only in terms of its established semiconductor names; there are early stage businesses at the crossroads of hardware and AI that international VCs can find here. The advantage is hard to copy. In one highly concentrated ecosystem you have the researchers, the engineers, the manufacturers and the supply chain all in one place. For a startup making the jump from prototype to production, that is invaluable. For an investor, it is a chance to spot something that may be beyond the reach of the traditional Silicon Valley network.

Erica Lin, Director of ITRI International, took the stage to discuss a side of Taiwan that is less well known to outsiders: the emerging startup investment scene. She presented ITRI International as the entry point for overseas founders and capital.

Taiwan has been working to tie its entrepreneurs into the American community for some time. Lin put forward figures showing that from 2017 to 2026, over 500 local startups have gone through three month immersion programs in Silicon Valley via TIEC, TTA and TREE. Some 54.7 per cent of them have been able to secure follow on funding. Cumulatively, more than 1,500 Taiwanese startups have been in attendance at events in the U.S. like CES or TechCrunch since 2017.

But ITRI is now looking to do more by linking up with the corporate venture capital side of things. The Taiwan CVC network is designed to open doors for international investors to make co-investments and get in front of the relevant government programs. A corporate VC is not the same as a financial one; they can put engineering talent, distribution and customers in front of a startup, which is of particular use for those in the hardware and Physical AI space.

Lin also put out an invitation for a Taiwan Innovation Visit in 2027 for anyone wanting to see the island for themselves. The intent is unambiguous: Taiwan is not just asking the world to be a customer of its technology, but to come and help build the companies of tomorrow.

Taiwan’s Next Generation of Deep Tech Founders

From the industrial prowess Taiwan has long been known for, the session made a shift to focus on the entrepreneurs of tomorrow. Morale.ai and JmemTek, two homegrown startups, were on hand to put on display the full scope of the country’s innovation ecosystem and show how a fresh crop of founders is taking Taiwan’s technological edge into markets yet to be explored. It was a defining occasion for them, as well as a great honor, to share the TechCrunch Disrupt stage with some of the visionaries and legendary names from Silicon Valley. In doing so, they found themselves in the global eye and sent a clear message: when it comes to competing and leading the world, Taiwan’s next wave of startups is more than ready.

Morale.ai: Intelligent Agents on the Factory Floor

Phill Kao, the CEO and cofounder of Morale.ai, was on hand to present his company’s manufacturing AI platform. Where some AI firms are content with general purpose language models for business productivity, Morale.ai is set on tackling the particularities of the manufacturing sector. The goal is to infuse industrial operations with autonomous agents and a form of domain specific intelligence.

A factory is awash in operational data, yet much of it is siloed in machine readouts, engineering files or the heads of seasoned workers. Morale.ai wants to put all that under one roof with an intelligence platform of its own making. By way of MEMO, an industrial memory system, and other technologies, the company can support workflows in semiconductors, machinery, textiles, metalworking and even footwear.

What makes their contextual intelligence layer so vital is the preservation of an information’s history and source for the AI to reference. In an industrial setting the stakes of a wrong call are nothing like a chatbot error. It was a case study in how the future of AI may be less about the size of the model and more about its capacity to grasp complex industrial processes.

JmemTek: Hardware Security for the Quantum Age

Then there was JmemTek, with CEO John Chiang to talk about the frontiers of cybersecurity in a post quantum world. With quantum computing on the rise, there is mounting pressure to harden the security around critical infrastructure and connected devices. JmemTek is putting hardware based solutions to work against those risks.

Chiang laid out a business model that covers everything from IP licensing to pacq kaged integrated circuits for customer systems on chips, as well as products with PCIe and USB interfaces. It is a good example of the deep tech side of Taiwan: taking semiconductor level intellectual property and making it a deployable product.

Taken together, the duo made a statement. Taiwan’s startup scene is not just about the next round of semiconductors; it is producing the intelligent software and integrated hardware that sit on top of them.

Phison: A Storage Giant with New Ambitions

Perhaps the most telling presentation of the day was from Michelle Chan of Phison Electronics. To many in the room Phison is a known quantity as a top tier supplier of NAND flash controllers and firmware, if not by name on the end product. Since 1998 the company has been a force in storage technology; Chan put it at one in five NAND flash devices globally.

But Phison had bigger things on its mind at TechCrunch Disrupt than talking shop about storage. Chan wanted to show how the company is using its expertise to underpin enterprise AI and the Physical AI ecosystem. She began with an anecdote about a partnership that saw Phison storage tech make the trip to the Moon in a hardware data center – a testament to where their technology can go.

Still, the focus was on the future rather than the last 26 years of accomplishments. “The biggest challenge in AI is no longer building the model,” Chan contended. Plenty of organizations have dabbled with it, but very few have made it part of their core business. That is where the real work lies. Various surveys were put forward in the presentation to show the chasm between first forays into AI and true operational integration, with the caveat that the definitions and methods behind them are not uniform.

But the issue is really quite simple. An AI application can be a marvel in a controlled demo, yet once it is rolled out across a business, one encounters a host of new challenges. There are infrastructure costs to be had, matters of data privacy and system integration to sort out, and the question of how to scale while dealing with the vagaries of model or API changes. All of this takes on added significance when the AI in question is underpinning vital operations. A factory is not going to halt production over an unavailable cloud service; hospitals and financial firms have to keep a tight rein on sensitive data. Industrial setups demand low latency and a predictable relationship with the hardware at hand. In such cases, the economics and demands of the AI infrastructure are every bit as consequential as what the models can do.

The task ahead is to make intelligence something that is practical, secure and affordable to deploy, not just more clever. Phison sees an opening there.

Take aiDAPTIV, for instance. It is the technology platform at the heart of Phison’s approach to giving storage a greater part in artificial intelligence. Chan made the case by way of NVIDIA’s CUDA ecosystem: where graphics processing units were once for graphics, CUDA opened the door to general purpose computing and a vast array of new applications. Phison is looking to do the same for NAND flash. Historically a means of storing information, through aiDAPTIV the company wants to see it take on a wider remit in AI systems for things like memory management and larger models. To that end they are constructing an integrated AI Data Platform, or AIDP, to help enterprises and edge environments get past the prototype phase and deal with the infrastructure hurdles that stand in the way.

The strategy is not confined to Phison’s own wares. The company has mapped out an ecosystem of OEMs, software vendors, system integrators and others. The intent is to be an integrated provider of AI technology and not just another semiconductor supplier.

Phison offered some numbers to back up its claims. More than 72 per cent of staff use AI on a daily basis, amounting to some 8,400 hours in productivity savings each week. Deployments have been made in over 30 hospitals, 10 banks and a dozen or so large enterprises. These are the kinds of figures that suggest the platform is already at work in the real world.

Then there is the matter of corporate venture capital. Perhaps the most telling part of the presentation was the focus on investment and working with startups. For an established name from Taiwan, this is something of an evolution. Phison is positioning itself as a collaborator and investor in the companies of tomorrow. Startups are being invited to come to the table for joint development and strategic investment, or to put their wares on display at the Silicon Valley AI Demo Site and meet with potential customers.

For an early stage founder with a software innovation that needs to become a physical reality, having a partner with Phison’s manufacturing ties and global reach can be the difference between a good demo and a viable business. And for Phison, it is a way to find new commercial avenues for its platform. While no specific fund sizes or portfolio details were given, the message from the TechCrunch Disrupt stage was clear: the company is done merely providing the infrastructure for innovation. It intends to have a hand in shaping it.

“Silicon Valley invents. Taiwan builds.” It is a simple statement, yet one of the most powerful put forward at Phison’s session. In many ways it sums up the technological alliance that has been the backbone of the global electronics industry for some time.

The division of labour has always made sense. Silicon Valley has an edge in venture capital and the creation of startups, not to mention its prowess in software and AI research. Taiwan has the industrial might to put sophisticated products on the market at scale, backed by advanced semiconductor and systems engineering. These have long been complementary strengths. But with the advent of Physical AI, the need to work together is more pronounced.

Take an AI application that runs without a hitch on a laptop; getting it to function within an industrial robot is another matter entirely. An autonomous machine is only as good as its integration of software intelligence with sensors, power and mechanical systems, and the ability to make decisions in real time. To move from a promising prototype to something commercially viable and reliable requires the kind of expertise Taiwan has in abundance. This is the strategic advantage ITRI and Phison were keen to underline in their efforts to forge better ties with the entrepreneurs and investors of Silicon Valley.

The aim is not merely to be the manufacturer for American inventions. The real prospect lies in deeper collaboration across the Pacific between the software innovators, the hardware firms and the venture capital community to build new businesses. By connecting with TechCrunch Disrupt, Sparknify helped put ITRI and Phison in front of an audience on the conference’s official Breakout Stage, introducing the speakers and steering the program.

It was a timely move. Taiwan is no stranger to the international technology circuit and the major semiconductor shows. But TechCrunch Disrupt is a different proposition. You will find the early stage founders, the angel and corporate VCs, the tech executives – the ones who decide which companies get funded and which technologies are worth a commercial bet. For Taiwan, this is a chance to go beyond old industrial relationships. It puts Taiwanese founders in contact with investors who may be more accustomed to software plays, and it gives established names like Phison a platform to talk strategy with their Silicon Valley counterparts.

In the end, it was about more than putting Taiwan’s technology on display. It was a way to have a say in the investment story for the next wave of AI.

For years the world has known Taiwan as the indispensable manufacturer, the place where the supply chains for the largest tech companies are supported and the semiconductors for every smartphone are made. Physical AI is opening up a wider field. Taiwan is becoming a destination where global investors can find innovative startups and where international founders can secure the partners they need to bring intelligent systems to market.

The presentations from ITRI and Phison made that clear. One showed the maturation of the local venture ecosystem and a new breed of entrepreneur in fields like digital twins and cybersecurity. The other showed how a veteran of the industry is moving into AI infrastructure and startup investment. They offered a view of Taiwan that is more than just the world’s semiconductor factory.

There is logic to the timing. As AI leaves the cloud to run vehicles and factories, the centre of gravity moves to the capabilities needed to make those physical systems work. Silicon Valley will keep coming up with extraordinary new ideas. But the companies that will prevail are the ones that can turn that intelligence into something practical. In that regard, Taiwan will not just be building the next AI revolution. It will have a hand in deciding who does.

The discussion took the form of “Taiwan Powering the Next AI Revolution: Connect, Invest & Scale” on October 13 at Moscone West in San Francisco as part of TechCrunch Disrupt 2026. With representatives from ITRI International, Phison Electronics, Morale.ai and JmemTek (and a nod to MetAI in the startup showcase), the session was all about forging those connections.

This report draws on the event agenda and materials from the organizations involved, as well as the published program from TechCrunch. Any statistics on company or ecosystem performance are as presented by the respective parties.