Dr Long Sheng Fan, CEO, discusses Iridium Medical Technology’s development of a 4,000-pixel bionic retina. By leveraging Taiwan’s semiconductor expertise and AI convolution, the firm addresses retinal degeneration. Currently in clinical trials, the venture targets high-value medical markets, aiming to restore functional vision through sophisticated neural electronic interfaces and global collaboration.
The founding of Iridium Medical Technology occurred at a time when the field of artificial vision was notably constrained. What provided the conviction that such a sophisticated technology was achievable, and how did you identify its commercial potential?
Our journey began with fundamental research in the laboratory. During my time at the University of California, Berkeley, we collaborated closely with its neuroscience institute to conduct basic research. We studied retinal neuron spiking and stimulation. We continue the study ever since after I am back to Taiwan. By recording the correlation between that stimulation and the subsequent firing, we established an artificial intelligence model to simulate what those signals represent by the neural population codes – essentially decoding what they mean in terms of human vision.
Coming from a semiconductor background, I noticed a stark disparity. Consumer electronics, such as mobile phones, already possessed multimillion-pixel sensors. At that time, the only company with US FDA approval was producing an artificial vision device with only 60 pixels, which barely met the needs of the patients. Their technology were limited to the off-the-shelf electronic technology and components at that time; because only those were available to the primary surgeons driving the field.
Taiwan is a global powerhouse in semiconductor technology. We realized our strength lay in leveraging this technological infrastructure to create something truly transformative for blind patients. We knew we could not settle for a rudimentary solution; we wanted to bring a sophisticated, high-resolution product to the market.
Your technology reportedly provides a 4,000-pixel bionic retina, representing a significant leap forward. Which specific breakthroughs in semiconductor materials or processes enabled this?
In the past, the industry standard involved placing passive micro-electrode arrays in the eye to communicate with resonant neurons. These systems required an external camera mounted on eyeglasses to transmit information to the chip. This was highly inconvenient. Furthermore, bandwidth limitations made it difficult to achieve a resolution useful to the patient. Because the camera was fixed to the glasses, patients had to rotate their entire head to change their view; they could not simply move their eyeballs.
Our approach is fundamentally different. We have integrated our retina and neural interface chips so that we can utilize the natural optics of the eye. Light is projected directly onto our implanted chips, which then perform AI convolution to translate those images into neural signals. By communicating directly with the surviving neurons to send signals to the brain, the patient experiences vision in a much more natural and effective manner.
What does “functional vision” mean in the context of your product? If successful, what specific capabilities will it restore to the patient?
A 60-pixel resolution is comparable to looking at an old-fashioned alarm clock. We have targeted 4,000 pixels because we analysed the requirements of human sight. Central vision is primarily concentrated within the central 10-degree field of view, which has a resolution 10 times higher than the best peripheral vision. This is the area you use to read, watch a film, or recognise human faces.
Our objective – provided the underlying neural condition allows for it – is to enable a blind patient to read a book printed in 14-point font. This would be a monumental victory for older patients. By providing a 10-degree field of view at a 4,000-pixel resolution, we establish the necessary conditions for a patient to regain the ability to read.
In terms of your target market, how are you segmenting the patient population?
We are primarily focused on patients with retinal degeneration, which falls into two major categories. The first is retinitis pigmentosa (RP). Approximately one in 3,500 to 4,000 new-borns carries one of roughly 200 gene mutations that lead to this degeneration. It is considered a rare disease.
The second, much larger category is age-related macular degeneration (AMD), which affects roughly seven million people. Within that group, the majority suffer from “dry” AMD. As the condition progresses to its late stage, known as geographic atrophy (GA), patients develop dark, obstructed areas in their central vision. While they may still sense movement in their periphery, they cannot identify objects because they have lost that vital central focus. In the US alone, there are 160,000 new GA patients every year. This represents a massive population in need of a solution.
Given the progress in gene editing and medicine, do you view those players as competitors or potential partners?
I view gene therapy as a complementary partnership. There is a company that recently received FDA approval for an RP treatment, which costs approximately 400,000 USD per treatment. It is very expensive. Furthermore, gene therapy is largely effective only in the early stages of the disease; if you have already lost your vision, it cannot currently be reversed.
Given that there are 200 different gene mutations, one would need to spend billions of USD to address just one of them. To address the entirety of RP through gene therapy would require hundred times more than that. Currently, that is a difficult proposition for investors. The existing treatment addresses roughly one percent of RP patients. If a patient is in the early stages and has the specific mutation, they should pursue gene therapy. However, for those who are already legally blind, a retinal prosthesis currently offers the only solution of seeing again.
The field of artificial vision has seen many attempts but limited commercial success. What differentiates the eventual winners in this space?
One must stand on the shoulders of giants. I have a high degree of respect for the pioneers at Second Sight; their product was a proof of concept that vision could indeed be restored. However, while they succeeded in moving from zero to one, “one” was not adequate.
The difference now is the integration of engineering and physiology. The earlier devices were driven by surgeons who used off-the-shelf parts. Armed with capabilities of semiconductor technology development and integrated circuit, we have advantage in providing innovative technical solutions. We can perform processes that even TSMC cannot – such as creating paper-thin, curved packaging for chips that must be biocompatible as an ocular implant. We leverage what the large foundries can do and extend it into the specialized realm of biomedical implants.
You are currently in the clinical trial phase. What can you share regarding your progress?
We are conducting our first-in-human clinical study. The primary endpoints are to verify safety, the surgical procedure, the durability of the device, and biocompatibility. The Taiwan FDA does not allow us to comment on specific results until the trial is concluded, but we are very optimistic.
Initially, we planned to conduct these trials in the US. We had face-to-face meetings with the US FDA and had arranged for surgeons from US hospitals to come to Taiwan to practice the procedure. Then the pandemic occurred. What we thought would be a six-month delay lasted three years. During that time, we were maintaining a clean room and expensive assets, so we shifted our strategy to begin in Taiwan first. We received approval last year, commenced this year, and we intend to begin European trials next year.
Over the next two years, we intend to prove the safety of the device across different ethnicities, moving from Taiwan into Europe. We want to verify the effectiveness of the system to a degree that allows us to transition into a larger-scale pivotal trial.
Neuro-implants are exceptionally complex. What kind of partnerships are you seeking to support this growth?
We are situated in the heart of Taiwan’s semiconductor industry. While we have our own clean room for specialized materials, we leverage TSMC and others for the basic semiconductor wafers manufacturing. We then perform the post-processing to make them biocompatible and effective as neural interfaces.
On the clinical side, we have collaborated with National Taiwan University Hospital and others for the past eight years. Their surgeons helped develop the surgical procedures. Since the pandemic subsided, we have trained retinal surgeons from the US and Europe who will act as our “seed” surgeons for international trials.
Taiwan is synonymous with semiconductors, but you are a prime example of leveraging that expertise for biomedicine. What potential does Taiwan hold for next-generation medical technologies?
In the past, Taiwan was largely driven by the manufacturing needs of companies like Apple and Nvidia. We perfect the semiconductor technology to produce the chips for applications others would create. Now, we are trying to drive the development of the medical devices themselves. This is a “blue ocean” area.
It is also highly meaningful from a commercial perspective. In traditional manufacturing, one might make components with relative low margins. Here, we are developing a high-value device that could command a price of 100,000 USD or more, similar to gene therapy or high-end Medicare reimbursements. We want to showcase that Taiwan can lead in high-value, internationally competitive medical products.
Do you see an appetite from the “big players” like TSMC to invest directly in healthcare?
The business models are different. TSMC’s principle is generally that they do not want to compete with their customers by making end products. They focus on foundation technology. For our growth, we rely more on specialist biotechnology venture capital in Taiwan and the US.
Regarding capital intensity, developing a drug can cost more than one billion USD. Device development is typically lower, perhaps 100 million USD for the high-end devices. However, the return can be significant. If we achieve just five percent market penetration, we are looking at a revenue of more than one billion USD. It is very time-consuming – the regulatory and clinical process for a Class III active implant is comparable to drug development timelines – but it is highly cost-effective in the long run.
How prepared are regulators for these kinds of “active” intelligent implants?
It is a learning process for everyone. The US FDA published guidelines based on earlier, simpler devices, but this is new territory. In our meetings with the US FDA, they have been open to dialogue, essentially telling us that as the experts in the technology, we must help define the consensus on safety and efficacy. It is a peer-reviewed journey toward finding what needs to be in place.
Is an IPO a formal part of your current roadmap, and which milestones are essential to reach before a market debut?
Our business plan, supported by venture capital investment, naturally accounts for an eventual exit. At this juncture, we remain quite flexible and have not committed to a singular, fixed path. Our primary focus remains on hitting the necessary clinical markers to prove the efficacy of the technology. In the medical device sector, we typically operate within a two-phase structure, which differs from the three-phase process common in drug development. We begin with a feasibility study involving a small cohort of patients, followed by a larger pivotal study. We are currently in the feasibility study stage. Progressing to the pivotal trial is a significant milestone; it demonstrates a level of risk control that provides the market with much greater confidence in our eventual commercial success.
How are you attracting top-tier talent, and are you drawing primarily from Taiwan or leveraging your international connections?
We maintain several international collaborations, but for the current development phase, we rely heavily on Taiwan’s Semiconductor Science Park and the nation’s premier academic institutions – specifically National Taiwan University, National Tsing Hua University, and National Yang Ming Chiao Tung University.
In terms of integrated circuit (IC) design and semiconductor technology, we are competing for the same talent pool as giants like TSMC and Nvidia. It is a fierce environment for recruitment. We have had managers lost to them in the past. To retain our people, we offer a vision that goes beyond short-term financial incentives or annual bonuses. People are looking for meaning. They want to know they are not just designing chips for Bitcoin mining etc., but creating something with a profound impact on society.
If you were to look further down the line, how do you see this technology evolving? What is the long-term potential for these neural interfaces?
While our current device addresses a fraction of normal vision, one can certainly envision “super vision” capabilities. Our chips are technically capable of seeing into the infrared spectrum, though we intentionally filter that out for now.
More broadly, we position ourselves as a neural electronic interface company. Our foundational technology is not limited to the retina; it is applicable to various forms of neural atrophy and the many neurological disorders that cause immense suffering. This project serves as a rigorous training ground. Developing a device that can be immersed in warm salt water for decades, communicate wirelessly, and perform complex AI convolution inside the body is one of the most significant engineering challenges one can face. Once we master this, the applications are limited only by one’s imagination.
In this development journey, do you seek partnerships with major industry or medical device players, or do you intend to remain independent?
Collaboration is essential, though it requires a nuanced approach. While we are willing to collaborate with foundries such as TSMC etc. – which would negate the need to maintain our own clean room – the scale of our operations is vastly different. They produce hundreds of millions of wafers each week. In contrast, our production of 20,000 chips per year, which may help 20,000 individuals, is a monumental achievement for us but is not necessarily a primary priority for a high-volume foundry. Consequently, we maintain our own research laboratory within the Science Park to conduct neural recording and develop our proprietary code. We also collaborate closely with hospitals and intend to seek deeper partnerships with neuroscience research labs in the future. Given the immense complexity of the human neural system, there is always substantial room to improve our interface and understanding.
As a scientist who has transitioned into the role of Chief Executive, what is your philosophy on successfully marrying business with science?
Success is driven by identifying a genuine need. We evaluate what is technologically feasible and then undertake the long process of verifying that feasibility within the human body. That remains the primary bottleneck.
On the commercial side, venture capitalists are exceptionally savvy at validating the business case. They may spend three months or more scrutinising the market before investing. I have also personally invested in my own education, returning to university for executive training in business strategy & market finance etc. to ensure I can navigate both worlds effectively.
What is your final message regarding the potential of Taiwan within the global biomedical field?
Taiwan has the opportunity to leverage its existing industrial strengths to contribute to the global community. This should not be viewed through the lens of competition, but rather as a win-win collaboration. By merging our semiconductor heritage with biomedical innovation, we can make significant business and humanitarian progress happen more efficiently.
