Biohybrids Threaten Neuralink and Others

In the winner-takes-all market of BCI, Science Corp's new “biohybrid” could disrupt industry giants like Neuralink, Paradromics, Synchron, and others, who have collectively raised over $1 billion.

Science Corporation (founded by former Neuralink president Max Hodak) is pioneering a “biohybrid” brain-computer interface (BCI) that uses living neurons embedded in a microelectronic device to communicate with the brain​. Instead of inserting metal wires or silicon probes into brain tissue, Science Corp’s approach integrates genetically engineered neurons into a chip, essentially adding a new layer of brain cells that bridge the device and the brain​ (see initial patent). These neurons, derived from stem cells and modified with optogenetic genes, are grown on the device in vitro and then engrafted onto the cortex, where they extend their axons and dendrites into the host brain. I’ll first break down the molecular biology behind the engineered cells, the materials and design of the chip, and the engineering that enables these “living electrodes” to both read neural signals and write to the brain. Then, I’ll share my thoughts on Science Corp’s biohybrid approach to other BCI technologies (Neuralink, Paradromics, Blackrock Neurotech, Precision Neuroscience, Synchron, INBRAIN, and the BrainGate program), highlighting how biohybrids promise orders-of-magnitude more channels (100,000+ vs ~1,000) without the tissue damage of traditional implants. Finally, I’d love to share some stories about Max Hodak’s journey, from co-founding Neuralink to starting Science Corp, and his vision for the company’s evolution: one that I share, and which inspired me to start my own company Synaptrix Labs two years ago.

It’s clear that Hodak’s decision to found Science Corp was driven by a desire to pursue this different approach that he likely couldn’t fully execute at Neuralink (which was more focused on traditional electrodes and near-term trials). He saw an opportunity to combine gene therapy, optogenetics, and cutting-edge fabrication to leap ahead in BCI capability. In doing so, he left a high-profile role at Neuralink and quietly built Science Corp, emerging once there was something tangible to show. As he put it in 2022, “Neuralink’s vision was great, but our approach diverges in that we integrate biology directly. We want to redraw the borders around the brain by adding new neurons, rather than just listening to what’s there.” (This paraphrases sentiments from his blog and interviews).

In the world of BCI, performance and safety are our north stars, inherently linked. This industry is undoubtedly a winner-takes-all market, and if the current trajectory holds, I believe Science Corp will ultimately emerge as the clear leader.

Creating a Biohybrid: Engineered Neurons on a Microfabricated Chip

To understand Science Corp requires a more nuanced understanding on how the core technology behind their biohybrids works. Please bear with me!

Biological Component – Stem Cell-Derived Neurons: Science Corp grows hundreds of thousands of neurons in the lab from stem cells, engineering them to suit the implant’s needs. The neurons are typically derived from induced pluripotent stem cells or embryonic stem cells that are differentiated into cortical neuron types​. They are “highly engineered”. For example, they are genetically modified to express light-sensitive ion channels (opsins) so that they can be stimulated with light (the basis of optogenetics)​. This allows the implanted neurons to fire in response to specific light pulses from the device. The cells may also be edited to reduce immune rejection (a major challenge for any graft). In fact, Science Corp has discussed creating “hypoimmunogenic” neuron cell lines: cells cloaked from the immune system so they won’t be attacked as foreign​. Engineering such universal donor neurons involves gene edits (for instance, removing or altering MHC markers), and it’s an active area of development. The engineered neurons are kept alive and immature in vitro until implantation, at which point they will continue maturing in the brain. Notably, because these graft neurons are derived in the lab, their cell bodies can be held within the device (where they receive nutrients and support), and only their long processes will enter the brain tissue​. This “keep the cell bodies with us” design is crucial: it means the delicate soma (cell body) stay anchored in a controlled environment, while the axons and dendrites do the outreach into the host brain​.

Materials and Chip Design – A “Living Electrode” Scaffold: The physical interface is a microfabricated chip that serves as a scaffold and communication platform for the neurons. Science Corp’s device is described as a honeycomb-like silicon structure with over 100,000 tiny cavities or microwells. Each microwell is a cylindrical hole about ~15 micrometers deep; just large enough to host a single neuron’s cell body​. Neurons are seeded into these wells so that each well contains a neuron that is effectively “trapped” in the device (the well walls hold the soma in place)​. The chip is thin and is surgically placed on the surface of the cortex (in the subdural space, just above the brain’s layer 1) – creating what the inventors call an “artificial cortical layer 0” on top of the brain​. The underside of the chip faces the brain, with openings that allow neural processes to grow out. The topside of the chip contains dense arrays of electronics: microelectrodes for recording and microLEDs for stimulation, arranged around or beneath the microwells​. In essence, each neuron-in-a-well becomes a biological electrode that can pick up and send signals (Initial patent). Figure 1 below illustrates this concept: a neuron sits in a tiny well between a light emitter and an electrical sensor on the device.

Figure 1: A conceptual rendering of an engineered neuron (cell body in white) interfaced with a microLED (glowing square) that activates it optically, and surrounded by electrode tiles (gray) for recording its electrical activity​.

Science Corp’s biohybrid chip incorporates optical stimulators and electrical sensors around each embedded neuron. The neurons are genetically modified with light-sensitive proteins (opsins), so a microLED on the device can trigger the cell to fire when illuminated​. Nearby recording electrodes detect the neuron’s activity, including any incoming signals from the brain via synapses onto that neuron​. In this way, each neuron acts as a bidirectional relay. It can be stimulated by the device (via light) and can relay impulses from the brain back to the device (via the electrode sensing when the neuron fires in response to brain inputs). This design achieves single-cell resolution communication: the device talks to and listens to the grafted neurons, which in turn talk to the patient’s own neurons via natural synapses. The use of light for stimulation confers cell-type specificity (only opsin-expressing graft cells react to the microLED flashes​ and avoids the electrical interference issues that come with traditional electrode stimulation.

Manufacturing the Neural Scaffold: The chip itself is produced with advanced microfabrication techniques (Science Corp has an in-house Science Foundry for MEMS fabrication). The microwell array is likely made by etching silicon wafers to create a high-density grid of tiny wells. Each well is surrounded by insulating material and embedded wiring that connects to on-chip electronics for that site. Science Corp has developed custom ASICs (dubbed “Nixel” and “Pixel” chips) to handle the massive channel count – Nixel 512 interfaces with 512 electrical channels for recording, while Pixel 2K/16K drive up to 2,000 or 16,000 optical channels (microLEDs) per chip. These ASICs are bonded to the microwell array, effectively turning the whole ensemble into an “active probe” with integrated stimulation and recording at each site​. The material choices are critical: silicon provides a rigid support and can endure microfabrication; biocompatible coatings (e.g. polyimide or parylene) are used on surfaces that contact tissue to encourage neural growth and prevent corrosion. The wells’ interior might be coated with extracellular matrix proteins or poly-D-lysine to encourage the neurons to adhere and survive (without direct knowledge, I can only best infer). The device likely also has porous features or open channels to allow nutrients and oxygen from cerebrospinal fluid (and maybe ingrowing capillaries) to reach the cells. This is important because keeping neurons alive on a chip requires that they get sufficient nourishment once implanted. In the first mouse prototypes, about 50% of the implanted neurons survived 3 weeks post-implant, indicating there is room to optimize the environment for better cell viability (through improved materials, coatings, or perhaps co-culturing supportive glial cells). Nonetheless, that half of the cells survived and functionally integrated is a strong proof-of-concept.

Developmental Integration: Once the biohybrid device is implanted on the brain, the grafted neurons don’t just immediately start working, they undergo a maturation process. Notably, the engrafted layer of neurons (sometimes called “Layer 0”) shows synchronous waves of activity in the days after implantation, akin to developing neural tissue​. Over time, this activity settles down to resemble normal cortical firing patterns. This suggests the graft is maturing and integrating into the brain’s existing circuits. During this period, the neurons are forming connections: their dendrites and axons sprout out of the microwells into the host brain, seeking synaptic partners. This is guided by the natural propensity of neurons to form networks. The axons (the long output fibers) grow into the cortical tissue, extending vertically and laterally, and the dendrites (receiving branches) reach into nearby neural layers. Crucially, this outgrowth is non-destructive: the processes snake through the brain’s extracellular space, which though tight, can accommodate these microscopic fibers without displacing or crushing resident cells​. Over weeks, the graft axons and dendrites form synapses with the host neurons, meaning the new neurons become functionally connected into the brain’s signaling networks​.

Figure 2: Histological image of a mouse cortex with an engrafted biohybrid layer on top (green). The green filaments are axon and dendrite projections from the implanted neurons, infiltrating the host brain tissue below. Cell nuclei in the host cortex are stained blue, showing the dense existing neural cells. Importantly, the graft fibers integrate between the host cells without displacing them (no large voids or scars are apparent). This demonstrates nondestructive integration: the living electrodes (graft neurons) weave into the native neural fabric and establish synaptic connections. Because these are true biological synapses, the communication between the device and brain is mediated by natural neurotransmitters and action potentials, not just electrical currents. This allows bidirectional communication: the graft neurons can receive signals from the brain (neurotransmitters released onto their dendrites will make them fire, which the device can record) and send signals into the brain (when the device optically stimulates the graft neuron to fire, its axon will release neurotransmitters onto target neurons in the brain). In the mouse studies, three weeks after implantation the graft neurons had indeed functionally connected. Shining light on the implant (to activate those neurons) caused the mice to behave as if they perceived that stimulation, proving the signal went into the brain.

Bidirectional Communication: How Biohybrid Neurons Read and Send Signals

The ultimate goal of a BCI is bidirectional information flow with the brain – i.e. the ability to record brain activity (read signals) and stimulate brain circuits (write signals). Science Corp’s biohybrid achieves this by using the grafted neurons as intermediaries between silicon and brain tissue. Each neuron is effectively a biological transducer: converting optical/electrical stimuli from the device into biochemical nerve signals in the brain, and converting neural activity from the brain into electrical signals the device can detect.

Writing to the Brain (Stimulation): Traditional BCIs stimulate neurons by passing electrical current through an electrode, but the biohybrid uses the more natural route of making a neuron fire via its native mechanisms. As described, the graft neurons are made responsive to light. Tiny microLEDs embedded in the device (on the side or bottom of each microwell) deliver pinpoint light pulses. The neurons express an ion channel (e.g. Channelrhodopsin-2 or similar) that opens when specific wavelengths of light hit it, causing the neuron to depolarize and fire an action potential. This technique is known as optogenetics and is widely used in neuroscience because of its precision. In the Science Corp implant, there are hundreds of thousands of microLEDs distributed across the array​, allowing many individual neurons to be stimulated in complex patterns. When a graft neuron fires (triggered by a light pulse), it will release neurotransmitters at all the synapses it has made onto the brain’s neurons – just as an ordinary neuron would. This means we can drive specific neural circuits in the brain by selectively activating particular graft neurons (for example, a graft neuron that connected into a motor cortex circuit could initiate movement, or one in visual cortex could trigger a percept). Because the graft neurons could be of various types (excitatory, inhibitory, modulatory), there is a possibility of very nuanced control – e.g. using excitatory neurons to turn things “on” or inhibitory neurons to suppress activity, or even dopaminergic neurons to deliver localized dopamine signals. This chemical signaling is something electrode-based stimulators cannot do (they stimulate everything nearby electrically), so it’s a unique advantage of the biohybrid approach. Again, the key is that stimulation is achieved by shining light on specific neurons in the implant, causing them to fire and convey signals to the brain in a physiologically natural way​.

Reading from the Brain (Recording): To record neural activity, the device doesn’t listen to the brain’s neurons directly, but rather listens to the grafted neurons as proxies. When the brain’s neurons fire and communicate, some of that activity will synapse onto the graft neurons’ dendrites. If a graft neuron is sufficiently stimulated by its inputs from the brain, it will fire an action potential. The device can detect this because each microwell has a tiny recording electrode that monitors the graft neuron’s electrical spiking. Essentially, the electrode is like an ECG for that single neuron – it picks up the voltage change when the neuron fires, converting it to a digital signal. Science Corp’s chip includes on-board amplification and multiplexing (via the Nixel ASIC) to handle these signals from potentially 100,000+ neurons simultaneously. By recording which graft neurons are firing (and when), the BCI gains a window into the activity of the host brain. Each graft neuron, through its synapses, is listening to a subset of the brain’s circuitry; if it fires, it indicates those connected circuits were active. Early results showed the graft cells exhibit spontaneous firing patterns that correlate with the host brain’s activity. In principle, if you can monitor a million neurons (via the million graft cells) you could capture a huge amount of information about the brain’s state, which is far more than is possible with current electrode arrays. Science Corp implant’s design ensures the recording is stable: since the electrode is not directly in the brain tissue (it’s in the well with the neuron), issues like electrode drift or scar encapsulation are minimized. The signal-to-noise ratio is high because the electrode is measuring a single, closely apposed neuron’s spike rather than distant cells​. Additionally, because the neurons are part of the device, their positions relative to electrodes don’t change over time (whereas in traditional implants, electrodes can move or cells can migrate away). Overall, the recording side of the biohybrid functions like a massive multi-neuron amplifier: it “listens” to the brain through the ears of the grafted neurons.

Closed-Loop Capability: By combining these read/write functions, Science Corp’s biohybrid interface is inherently closed-loop. It can detect when the brain (via a graft neuron) is doing something and respond by stimulating other neurons, all in real-time. For example, one could program the system to recognize a particular neural firing pattern (say, indicative of an upcoming epileptic seizure or a motor intent) and then trigger a pattern of optical stimulation to counteract or augment that activity. Traditional BCIs also aim for closed-loop control, but achieving it at single-cell precision over large brain areas has been out of reach. Biohybrids could change that by offering an army of “local observers” (the graft cells) distributed across the cortex.

Several cutting-edge technologies converge in Science Corp’s device to make this possible:

1. High-Density Microfabrication: The manufacturing of 100,000+ microwell/electrode/LED arrays on a chip is possible thanks to modern semiconductor fabrication adapted for biocompatible devices. Techniques like deep silicon etching, thin-film metallization, and wafer bonding are used to create structures at micron scales. Science Corp’s Foundry (in North Carolina) has experience dating back to the 1990s in producing MEMS and neural interfaces, which it leverages to build these complex chips.

2. Custom Electronics: Off-the-shelf electronics cannot handle tens of thousands of channels in parallel, so Science Corp developed custom integrated circuits (ASICs). The Nixel chip is an amplifier and digitizer array specifically for neural recording (512 channels per chip). The Pixel chip is designed to drive microLEDs; notably, a Pixel 16K can control 16,000 LED elements. By tiling multiple Nixels and Pixels, the system scales to the desired channel counts. These ASICs also manage data throughput and power: one key challenge is reading from 100k channels without overheating or generating massive data bottlenecks. Likely, on-chip processing is used to filter and compress signals, and optical or high-bandwidth wired links send data out.

3. Optogenetic gene therapy: The concept of using light to stimulate neurons required the discovery and refinement of opsins. The neurons used in the biohybrid are encoded with a light-gated ion channel (commonly channelrhodopsin-2 or an improved variant) such that blue light pulses trigger them. This capability is the result of molecular biology advances from the 2000s that Science Corp can now exploit in a clinical context.

4. Cell Engineering and Culture Techniques: Producing a consistent supply of human neurons and getting them to integrate is non-trivial. Advances in stem cell culture (feeder-free iPSC growth, differentiation protocols) and gene editing (CRISPR/Cas9 to create hypoimmunogenic cell lines, viral vectors or electroporation to introduce opsins) are key. The Science team, led by co-founder Dr. Alan Mardinly (Director of Biology), has been refining methods to grow neurons that remain healthy on-device and form the needed connections. For instance, controlling the maturity of neurons is important. They likely implant the neurons at a somewhat immature stage so they can adapt and grow in the host brain, much like a developing neuron would.

5. Surgical and Biointegration Techniques: Implanting the device requires neurosurgical precision to ensure the chip sits snugly on the brain surface without damaging underlying cortex. The device is designed to conform or at least cover the curvature of the brain (perhaps flexible ribbons or smaller tiles could cover larger areas modularly). Moreover, the surgery must keep the cells alive. Possibly the implant is kept bathed in solution until placement. The biohybrid approach also benefits from decades of research in neural tissue grafts (for example, fetal tissue transplants in brains) which showed that neurons can survive and wire in if placed correctly. The team likely uses immunosuppressive measures in animal tests (and will in humans) until immune-evading cell lines are perfected.

To put things bluntly, Science Corp’s biohybrid is an incredibly beautiful marriage of molecular biology (designer neurons), materials science (biocompatible silicon scaffolds), and neural engineering (optical/electrical interfaces) to create a BCI that behaves like an extension of the brain’s own network. Next, I’ll share my notes, comparing this paradigm with the approaches of other leading BCI companies, and why Science Corp’s technology could represent a disruptive leap forward in bandwidth and safety.

Biohybrid vs Traditional BCI Implants: How Science Corp Stands Apart

Most current BCI efforts rely on electrodes made of metal or silicon to interface with neurons. These range from invasive needle-like arrays inserted into the brain, to electrode grids laid on the brain’s surface, to electrodes delivered via blood vessels. Below I outline the approaches of key players who’ve raised nearly a combined $1 billion for their BCI technology: Neuralink, Paradromics, Blackrock Neurotech, Precision Neuroscience, Synchron, and the academic BrainGate consortium, and contrast them with Science Corp’s biohybrid strategy.

Invasive Implant Approaches (Neuralink, Paradromics, Blackrock, BrainGate)

Neuralink (Invasive Microthreads): Elon Musk’s Neuralink has developed a surgical robot that implants an array of ultra-fine polymer threads, each studded with electrodes, into the cortex. Each N1 Neuralink device contains 1,024 electrode channels, an order of magnitude more than older Utah arrays, and is implanted via a cranium-mounted module. These threads penetrate a few millimeters into brain tissue. Because they are very thin (40–50 microns wide), they cause less damage than rigid spikes, but insertion still kills some cells along each thread’s path. Neuralink’s design is fully implanted (including a wireless transmitter in the skull). It achieves high-quality recordings and has shown flashy demos through its first human patient, Nolan Arbaugh. However, the limitation is scaling: inserting thousands of threads would mean thousands of tiny needle insertions. As Hodak noted, “destroying 10,000 cells to record from 1,000 might be perfectly justified… but it really hurts as a scaling characteristic”. In other words, Neuralink’s ~1k channels might be okay, but trying for 100k channels with that approach would cause extensive tissue damage and inflammation. Over time, the foreign body response (gliosis) can encapsulate electrodes, potentially degrading signal quality. Neuralink is pursuing human trials for medical applications (e.g. paralysis), and for now, its channel count is state-of-the-art for implanted electrodes. Science Corp’s biohybrid aims to far exceed Neuralink’s channel count (100,000 vs. 1,000) by using densely packed neurons instead of threads – dramatically changing the scaling laws of how many neurons you can interface with versus how much damage you do.

Paradromics (Microwire Arrays): Paradromics is another startup focusing on high-channel-count implants. They are developing the Connexus interface, which uses bundles of rigid microwire electrodes (around 7,500 channels in latest public demos) bonded to a ceramic module. Their approach essentially is like a larger Utah array: lots of parallel needles inserted into cortex to achieve very high bandwidth. Paradromics has reported successful recordings in animals with thousands of channels. However, this approach faces similar biological challenges. Inserting a dense bed of ~7,000 wires inevitably causes significant acute injury to tissue and can provoke immune reactions or scar walls around each electrode. The trade-off between channel count and damage is a central issue here, as with Neuralink. Paradromics targets applications like restoring speech (requiring tapping many neurons in speech motor cortex). If Science Corp’s tech matures, a biohybrid implant could interface with just as many or more neurons without piercing the brain at all (neurons from the device simply grow in). Thus, Paradromics could eventually be outpaced by a biohybrid approach that achieves similar channel counts more safely. In the near term, Paradromics is closer to human trials (they have received FDA Breakthrough Device designation), while Science Corp’s BCI work is still preclinical. Paradromics might respond by emphasizing validated performance and perhaps investigating ways to mitigate damage (e.g. smaller electrodes, or combining electrodes with drug delivery to reduce scarring).

Blackrock Neurotech (Utah Array, BrainGate): Blackrock produces the Utah Array, a bed-of-nails style silicon chip with 96 hard electrodes (1 mm length) that are punched into the cortex. This is a well-established device used in the BrainGate academic trials since the early 2000s. Up to two Utah arrays (192 channels) have been implanted in human volunteers, enabling point-and-click computer control or robotic arm movement by paralyzed patients. The Utah Array provides fairly good signal quality initially, but it does cause damage: each of the 96 spikes kills tissue where it’s inserted and over months to years the signal often degrades due to gliosis (scar tissue) encapsulating the electrodes. Patients in BrainGate had to have open-brain surgery to implant the arrays, and infection risk from percutaneous connectors was an issue. Newer versions aim for wireless telemetry, but the fundamental limitations remain: you cannot pack hundreds of Utah arrays in the brain (that would be devastating to tissue), so channel count is limited to a few hundred per implantation. By contrast, a Science Corp biohybrid device could host thousands of neurons per square millimeter, potentially scaling to millions of channels over a larger area​. A million-device-embedded neurons might make a billion synapses in the brain, a connectivity level utterly unachievable with discrete electrodes. This puts the biohybrid in a different league of potential bandwidth. Blackrock’s strategy to remain relevant (among the many recent controversies) might be to move to less invasive electrodes (they are exploring flexible arrays and surface electrodes) or to provide specialized applications (e.g. high-durability implants for specific small targets) until biohybrid tech catches up. Blackrock is also part of BrainGate’s ongoing research.

BrainGate Consortium: BrainGate isn’t a company but a long-running academic collaboration (Brown University, Stanford, etc.) that has pushed the BCI field forward using implants like the Utah array. The BrainGate trials proved that even ~100 channels of brain input can allow a person to control a computer cursor or prosthetic limb. However, they also highlighted the challenges of stability and safety. BrainGate participants often had diminishing performance as months went on, and the system required careful calibration. Science Corp’s technology directly addresses some issues BrainGate faced: stability (the biological integration could yield more stable long-term recordings as the neurons become incorporated into neural circuits) and safety (no shrapnel of electrodes in the brain). On the flip side, BrainGate’s experience underscores the importance of reliable hardware and regulatory approval: areas where more mature electrode tech has a head start. It’s possible that BrainGate researchers or similar academic groups will begin testing biohybrid interfaces in animal models, given the encouraging early results. If Science Corp’s approach continues to advance, it could eventually render electrode-based systems obsolete for high-channel-count applications, effectively threatening the core tech used in efforts like BrainGate.

Minimally Invasive Approaches (Precision Neuroscience, Synchron, INBRAIN)

Not every BCI effort involves penetrating brain tissue. Some companies aim to get useful signals with less risky methods, albeit usually with fewer channels or lower fidelity.

Precision Neuroscience (Subdurally “Peelable” Array): Precision, co-founded by Neuralink alum Ben Rapoport, is developing a thin, flexible electrode grid that can be slid under the skull and rest on the surface of the cortex (above or below the dura). They dub their device a “Layer 7” interface (since it sits on the surface, which they jokingly call the 7th cortical layer). This approach is akin to an advanced ECoG (electrocorticography) grid but inserted through a tiny slit in the skull rather than a large craniotomy. It’s minimally invasive because it doesn’t poke into the brain, it just lies on top. The trade-off is that it records superficial field potentials rather than single-neuron spikes, so the information is less detailed. Precision’s first device may have on the order of 100–1,000 electrodes (significantly less than Neuralink’s penetration but more than noninvasive EEG). They have reportedly tested in at least one human patient during a neurosurgery, showing it can detect neural signals. Compared to Science Corp’s biohybrid, Precision’s array is lower risk initially (no foreign cells, no deep penetration), but its bandwidth is limited by the physics of ECoG. It cannot approach the single-cell, high-density readout that an integrated layer of neurons can. Moreover, even a flexible sheet can cause a foreign-body reaction on the brain’s surface over time (though likely milder than penetrating probes). Science Corp’s neurons actually integrate and possibly even get vascularized, potentially resulting in no chronic immune response in the ideal case. If the biohybrid achieves stable long-term integration, it would outperform surface arrays in both resolution and longevity (because surface electrodes might move or get encapsulated by a thin scar layer eventually). Precision’s competitive strategy might be to pitch their tech as a bridge or complement – for instance, use a less invasive grid for moderate channel needs (like seizure monitoring or basic motor decoding) in the next 5 years, while truly high-bandwidth applications wait for something like Science Corp’s tech to mature.

Synchron (Endovascular Stentrode): Synchron takes a radically different route – going through the blood vessels. Their Stentrode device is an expandable mesh electrode array that can be delivered into a cortical vein via a catheter (similar to how cardiac stents are placed). Once in the vein adjacent to brain tissue, the stent’s electrodes contact the vessel wall and pick up nearby brain signals. The big advantage is no open-brain surgery at all. The procedure is like an angiogram. Synchron has already implanted stentrodes in a few patients, enabling them to do tasks like type using brain signals. However, the stentrode currently has a very limited channel count (on the order of 16 electrodes) and because it’s in a blood vessel, the signals are filtered (they likely capture aggregate activity, not individual neurons). It’s a minimally invasive, low-bandwidth BCI suited for basic communication needs in paralyzed patients. In terms of tissue health, the stentrode avoids direct brain injury, but it lives in a vein which can lead to issues like thrombosis or the vessel’s inner lining reacting. Compared to Science Corp’s biohybrid, Synchron’s approach is at the extreme end of safety/low invasiveness but also the extreme low end of data resolution. Science Corp aims to be as safe as possible while still achieving massive bandwidth. If successful, the biohybrid would have the safety of a cell transplant (which is a procedure neurosurgeons are familiar with from stem-cell trials) combined with the communication capacity of an implanted electrode array. It basically occupies a middle ground: not fully noninvasive, but biologically gentle. From Synchron’s perspective, their niche might remain distinct. They could serve patients who absolutely cannot have brain surgery or don’t need high bandwidth (since a surgery to implant even a biohybrid is more involved than a vein procedure). But in a head-to-head technical capability comparison, a matured Science Corp implant could potentially send and receive far richer information than a stentrode. If Science Corp demonstrates safety in humans, it could negate Synchron’s current advantage (safety/minimally invasiveness) by offering both safety and high performance. Competitors like Synchron might then look into whether their approach can be scaled up (more electrodes in more vessels, or hybrid systems combining their tech with others), but there are physical limits. After all, blood vessels only go certain places.

INBRAIN’s implant also faces from similar obstacles, and there isn’t a need to repeat myself here.

Threat to Existing BCI Companies and Potential Counterstrategies

If Science Corp can translate their biohybrid technology from animal studies to human applications, it poses a significant disruptive threat to companies relying on electrode-based implants. Here’s how:

1. Unparalleled Channel Count

2. Improved Longevity and Stability

3. Lower Tissue Reaction and Safer Long-Term

Achieving on the order of 100,000+ channels (and in the future possibly millions) would allow BCIs to tap into brain signals at a scale needed for high-resolution mind reading or writing. For context, Neuralink’s 1024 channels are enough to control a cursor or maybe type a few words per minute with a lot of decoding software. 100k channels could theoretically capture complex neural patterns corresponding to speech, vision, or thoughts with much greater fidelity. This opens the door to applications like full-immersive virtual reality interfaces, decoding internal speech, or restoring movement with far more degrees of freedom. Companies like Paradromics and Neuralink, even as they up their channel counts, will struggle to reach those numbers with pure electrodes (due to the physical constraints discussed). Thus, Science Corp could outperform in raw capability, making their tech the platform of choice for any high-bandwidth BCI applications in the future.

One of the known issues with current implants is that signal quality tends to degrade over months/years as scar tissue forms or as electrodes corrode or move. A biohybrid that truly becomes “part of the brain” could have signals that improve over time (as connections strengthen) rather than degrade. This would be extremely attractive for chronic implants. It would also reduce the need for reimplantation surgeries or calibration retraining, which competitors might face with their devices. Invasive BCI companies could find their implants at a disadvantage if patients and clinicians see that a biohybrid graft can last, say, 10+ years with stable performance, whereas a conventional array might need replacement or gives diminishing returns.

If the immunogenicity problem is solved (e.g., via hypoimmunogenic cells or immunosuppressants), biohybrid implants might prove to be biologically inert in terms of harm, meaning a patient could have one in their brain with very low risk of inflammation or damage. Traditional implants carry risks of infection (if transcutaneous) or chronic inflammation and even microhemorrhages from electrode movement. A safer profile would make regulatory approval easier for broad use (beyond just the most severe paralysis cases). It also makes a case for elective use BCIs (for enhancement) more plausible in the long run – something electrode implants might never be deemed safe enough for. Competing companies would thus be threatened not just in performance but in the market scope: Science Corp’s tech could address both medical and eventually consumer BCI markets if proven extremely safe, whereas others might remain limited to specific medical indications.

All that said, it’s important to note Science Corp’s biohybrid is still in early stages. The technology readiness level is low (an “emerging” tech)​. It has significant hurdles to overcome: cell survival and immune acceptance (50% survival at 3 weeks in mice needs improvement​), scaling up manufacturing of living tissues, and demonstrating reliable function in complex tasks. Meanwhile, competitors have momentum – Neuralink and Synchron are in or starting human trials now.

So how might these companies respond or adapt? It’s something I’ve been thinking about for years, here are some thoughts:

Competitors will continue to innovate on electrodes – e.g. making electrodes smaller, softer, or coated with biomolecules to reduce the damage and scarring. Neuralink’s threads are a step in this direction (soft and thin), and future iterations might use even softer materials (like ultra-flexible polymers or hydrogels) to make “living-friendly” electrodes. The goal would be to narrow the gap by making electrodes less foreign to tissue. While this might not achieve the non-destructive ideal of biohybrids, it could allow higher channel counts than currently possible (maybe Neuralink can push to 10,000 channels by making threads so small and so close that damage is still minimal). Essentially, competitors can try to emulate the gentleness of the biohybrid by improving their materials. For example, there’s research on carbon fiber electrodes ~5 µm in diameter that cause very little damage. Companies might adopt such tech to stay relevant.

It’s conceivable that other BCI companies could explore incorporating biological elements themselves. While Neuralink as of now is purely electronic, nothing stops a competitor from researching, say, neurotrophic electrodes (electrodes that encourage neurons to grow onto them, a concept proven in the 1990s) or even partnering with cell therapy companies to test neuron-seeded electrodes. If Science Corp’s results continue to impress, we might see a convergence where even “traditional” devices attempt to use the brain’s biology. For instance, coating electrodes with stem cells or growth factors to draw host neurons closer (reducing the distance and improving coupling). However, doing what Science does (pre-growing neurons on a device) requires expertise that these companies currently lack, so a partnership or hiring spree in that area might occur. In fact, some key researchers in Science Corp’s orbit (like Dr. D.K. Cullen’s work on “living electrodes”) have been known in academia. Competitors could tap that existing research as well.

Understandably, competitors might downplay the need for ultra-high bandwidth in the short term and dominate the market for simpler BCI applications. For example, Synchron’s 16 channels are enough to restore basic communication for a locked-in patient today. Science Corp’s solution for that use-case might be 5-10 years away. Neuralink might achieve keyboard/pointing control or spinal cord stimulation for paraplegics much sooner than Science Corp can. By addressing immediate patient needs and gaining regulatory approvals, they can establish a foothold and revenues. This would give them resources to possibly invest in next-gen tech (maybe even license or acquire aspects of Science Corp’s tech if it proves itself). Essentially, their strategy could be: win the present, and buy/absorb the future. If Science Corp’s tech starts to outshine them, these companies might attempt to merge technologies or even acquisitions (though Science Corp, being well-funded and patent-protected, would be a tough acquisition).

Also, each company could highlight what their method can do that Science’s cannot (or will not for a while). For instance, Precision’s surface array might not give single-unit detail, but it can cover a broad area of cortex easily, perhaps useful for monitoring overall brain states (e.g. for epilepsy) with minimal invasiveness. They could position it not as a competitor to Science Corp’s high-density implant but as a complementary tool for different use-cases. Similarly, Synchron can claim the least invasive procedure, for patients who can’t undergo brain surgery, their stentrode is the only option. Paradromics might focus on periphery or biofeedback where they interface not just brain but nerves (Science Corp’s current work is brain-focused). By carving out specific domains or emphasizing certain reliability aspects (maybe electrode implants can be turned off or removed more easily if needed, whereas a neuron graft might be harder to explant), they can maintain relevance.

It’s likely all the players are keeping an eye on each other’s intellectual property. Science Corp’s approach, while novel, was presaged by prior art (their patients cite works from 1990s neurotrophic electrodes and 2010s biohybrid experiments). Neuralink actually filed a patent on a similar “layer 0” concept when Hodak was there, meaning Neuralink Corp (as the assignee) might have some claims in this space. Competitors could strengthen their position by filing improvements or alternative approaches to the biohybrid concept. For example, a patent on using a patient’s own iPSC-derived neurons for an autologous biohybrid could be pursued by another company, which would hedge against Science Corp’s allogeneic approach. By building a patent portfolio in bio-integration tech, they prepare to either implement it later or negotiate cross-licenses. (I’m not a IP/patent lawyer, so please take this with a grain of salt!)

Ultimately, Science Corp’s biohybrid neural interface represents a bold reimagining of the BCI: embedding living cells to merge with the brain, rather than pushing conductive probes into neural tissue. Its potential to provide massive channel counts (100,000+ vs. the ~1,000 of the best current devices) and do so with minimal brain damage is a clear advantage. This could enable truly high-bandwidth brain-machine communication, unlocking new possibilities that electrode-based systems would struggle with. Naturally, this poses a serious competitive threat to existing BCI companies who have invested in more conventional approaches. Those firms will likely respond through a mix of improving their own technology and, eventually, exploring biohybrid concepts themselves. In the near term, they retain an edge in readiness. Science Corp’s device has yet to be proven in humans, whereas Neuralink, Synchron, and others are already moving through clinical trials. It will be telling to watch the next few years: if Science Corp can demonstrate a functioning high-channel-count biohybrid in a primate or human with durable results, it may spark a paradigm shift in the industry, forcing all players to adapt to a new reality where biology and electronics work hand-in-hand in the interface.

In the world of BCI, performance and safety are our north stars. Thus, the industry is certainly a winner-takes-all market. And, in my opinion, if current trajectory continues, in the long-term, Science Corp will be the winner.

Afterword on Max Hodak: From Neuralink Cofounder to Science Corp Visionary

Max Hodak is a central figure in this narrative, having been intimately involved in both Neuralink’s founding and the creation of Science Corp’s biohybrid approach. Hodak’s background is in biomedical engineering and neurotechnology. He earned his B.S. in Biomedical Engineering from Duke University in 2012, where he worked in Dr. Miguel Nicolelis’s lab building BCI systems for monkeys (one of the pioneering labs in the field). An entrepreneur at heart, he had already started and sold a company (MyFit) while in college and later founded Transcriptic, a cloud laboratory platform, before turning his focus to brain interfaces.

In 2016, Max Hodak co-founded Neuralink alongside Elon Musk and a team of engineers and neuroscientists. At Neuralink, Hodak served as president and helped build the company’s initial technology – the surgical robot and flexible thread implant system. He often represented Neuralink in public tech demonstrations and was known as a driving force behind its science. Under his leadership, Neuralink pushed the envelope of electrode miniaturization and implantation techniques. However, by 2021, after five years, Hodak parted ways with Neuralink. He announced his departure in early 2021, amidst reports of internal tension and slow progress toward human trials. Hodak has refrained from speaking ill of Neuralink (“I have nothing bad to say about Neuralink” he told Futurism), but his departure hinted at differing visions. Indeed, Hodak later hinted that the technical approach he wanted to pursue was “extremely different” from what Neuralink was doing.

Shortly after leaving Neuralink, Max Hodak founded Science Corporation in 2021​. The company initially operated in stealth mode, but by late 2022 it emerged with a bold agenda. Science Corp’s first public reveal was the “Science Eye”, a prosthetic vision restoration device. In a December 2022 interview, Hodak described the Science Eye as combining “a simple gene therapy to add a protein to the optic nerve cells which makes them light sensitive, and a thin-film microLED display laid over the retina”. In other words, rather than putting electrodes in the brain, Science Corp’s first product uses photonics and genetic modification to interface with neurons; an approach philosophically aligned with minimizing invasiveness. The gene therapy delivers an opsin to retinal ganglion cells (or optic nerve fibers), and the implanted microLED film stimulates them, effectively replacing lost photoreceptors in diseases like macular degeneration. This strategy, using light to talk to neurons, is a recurring theme in Science Corp’s portfolio (it foreshadows the optogenetic stimulation used in the cortical biohybrid). By early 2023, Science Corp had raised around $160 million in funding, making it one of the best-funded neurotech startups (second only to Neuralink at the time)​. Investors were likely drawn by Hodak’s track record and the ambitious melding of biotech and hardware.

Hodak’s vision for Science Corp extends well beyond treating blindness. He views the “engineering of the brain” as a grand, “transcendent goal”. In interviews, he’s expressed that intermediate medical devices (like prosthetic vision or assistive BCIs for paralysis) are important milestones, but the ultimate promise of BCIs lies in high-bandwidth cognitive interfaces​. This aligns with Science Corp’s dual-track efforts: a near-term focus on the retina (which is relatively accessible and has a clear patient need) and a longer-term research program on cortical biohybrid implants. Hodak deliberately set Science Corp on a course distinct from Neuralink’s. “The technical approach we’re developing is extremely different from what Neuralink was doing,” he told Futurism, noting that their vision goes beyond just the first clinical application. Instead of placing electrodes in the brain, Science Corp explores “using the process of photonics rather than physically implanted chips” where possible, again emphasizing optical, minimally invasive strategies.

Under Hodak’s leadership, Science Corp assembled a team of top scientists and engineers. Notably, two of his co-founders are Dr. Alan Mardinly (leading the biology team) and Dr. Yifan Kong (leading microfabrication/devices)​, both mentioned in Science Corp’s blog as pivotal to the biohybrid project. The company had a clear culture of fast-paced tech development and deep research. Hodak has spoken about “letting ambition grow with success”, meaning Science Corp is careful to validate each step (like the Science Eye in patients) before making wild claims about the future. This is a contrast to Neuralink’s very public lofty goals; Hodak’s company has been a bit more tight-lipped until results are in hand (with the notable exception of sharing detailed blog posts and an open-access philosophy for their science).

It’s clear that Hodak’s decision to found Science Corp was driven by a desire to pursue this different approach that he likely couldn’t fully execute at Neuralink (which was more focused on traditional electrodes and near-term trials). He saw an opportunity to combine gene therapy, optogenetics, and cutting-edge fabrication to leap ahead in BCI capability. In doing so, he left a high-profile role at Neuralink and quietly built Science Corp, emerging once there was something tangible to show. As he put it in 2022, “Neuralink’s vision was great, but our approach diverges in that we integrate biology directly. We want to redraw the borders around the brain by adding new neurons, rather than just listening to what’s there.” (This paraphrases sentiments from his blog and interviews).

Now in 2025, Max Hodak remains CEO of Science Corp, steering the company through clinical trials for vision restoration and pioneering research for brain interfaces. He is an outspoken advocate for the long-term goal of BCIs unlocking human potential, but also a realist about the steps needed to get there. If Science Corp’s biohybrid neural interfaces prove successful, it will in many ways validate Hodak’s bold move to break away from Neuralink and blaze his own trail. In interviews, when asked if he’s been in touch with Elon Musk since leaving, Hodak diplomatically said “No comment”. With a mix of humility and ambition, Hodak and his team are gradually revealing a path toward BCIs that could transcend the limitations of current technology. I certainly wouldn’t dare bet against him.

As he and co-authors wrote in Science Corp’s blog, “there is an immense amount of science to do here”.

---

Top Relevant Sources & Further Reading (few of many):

• Science Corp’s “Layer 0” Patent – PATENTS.JUSTIA.COM
  Details the concept of using cell-based biohybrids as a cortical interface, providing foundational IP for Science Corp’s approach. (Referenced in the technology breakdown of Science Corp’s biohybrid neural interface.)

• 2024 BioRxiv Paper on Optogenetic Graft Integration – RESEARCHGATE.NET
  Describes experimental results showing how Science Corp’s engineered neurons integrate with the host brain and function as bidirectional relays.

• Science Corp’s Official Blog & News – SCIENCE.XYZ
  Authored by Hodak and the Science Corp team, these posts outline the architecture, motivation, and latest developments behind their biohybrid technology. (Referenced throughout the technology deep dive and competitive analysis.)

• Futurism Interview with Max Hodak (2022) – FUTURISM.COM
  Hodak discusses his departure from Neuralink, his vision for Science Corp, and why their approach differs fundamentally from traditional BCIs. (Cited in the section on Max Hodak’s journey and his reasoning for starting Science Corp.)

• Core Memory Interview (Ashlee Vance, 2025)
  Provides the latest insights from Hodak on Science Corp’s progress, challenges, and future goals in the neurotech industry. (Used in the Max Hodak section to provide updated perspectives.)

• New Scientist Interview with Alan Mardinly via SingularityHub – SINGULARITYHUB.COM
  Alan Mardinly, Science Corp’s Director of Biology, explains the science behind biohybrid neurons and how they form natural synapses with the brain. (Referenced in the technology breakdown of biohybrid neurons and their bidirectional capabilities.)

• Science Corp’s Vision for a Neuron-Based Brain Interface – SCIENCE.XYZ
  Discusses the long-term implications of using neurons instead of electrodes to create a more scalable and biocompatible brain interface. (Referenced in the conclusion, highlighting how this approach could redefine the BCI landscape.)