Modern biotechnology has the tools to edit genes and design drugs, but thousands of rare diseases remain untreated. The missing ingredient for many years, according to Insilico Medicine and GenEditBio executives, was finding enough smart people to continue the research. AI is becoming a power multiplier that allows scientists to tackle problems that industry has long neglected, they say.
Speaking at Web Summit Qatar this week, Alex Aliper, president of Insilico, explained his company’s goal to develop “pharmaceutical superintelligence.” Insilico recently launched the MMAI Gym, which aims to train general-purpose large-scale language models such as ChatGPT and Gemini to perform as well as expert models.
The goal is to build multimodal, multitasking models that can solve many different drug discovery tasks simultaneously with superhuman precision, Alipar said.
“We really need this technology to improve productivity in the pharmaceutical industry and address the workforce and talent shortages in this field, because there are still thousands of diseases for which there are no treatments or treatment options, and there are thousands of rare diseases that are being ignored,” Alipar said in an interview with TechCrunch. “So we need more intelligent systems to tackle this problem.”
Insilico’s platform incorporates biological, chemical, and clinical data to generate hypotheses about disease targets and candidate molecules. Insilico says that by automating steps that once required large numbers of chemists and biologists, researchers can explore vast design spaces, design high-quality therapeutic candidates, and even reuse existing drugs, all at dramatically reduced cost and time.
For example, the company recently used its AI model to identify whether existing drugs can be repurposed to treat ALS, a rare neurological disease.
However, labor bottlenecks extend beyond drug discovery. Even if AI can identify promising targets and treatments, many diseases require intervention at a more fundamental biological level.
tech crunch event
boston, massachusetts
|
June 23, 2026
GenEditBio is part of the “second wave” of CRISPR gene editing, which moves the process from editing cells outside the body (ex vivo) to precise delivery inside the body (in vivo). The company’s goal is for gene editing to be accomplished by simply injecting the gene directly into the affected tissue.
“We have developed a unique ePDV, or artificial protein delivery vehicle, which is a virus-like particle,” Tian Zhu, co-founder and CEO of GenEditBio, told TechCrunch. “We learn from nature and use AI machine learning techniques to mine natural resources to find out which types of viruses have an affinity for certain types of tissues.”
The “natural resource” Zhu refers to is GenEditBio’s vast library of thousands of unique non-viral, non-lipid polymeric nanoparticles, essentially delivery vehicles designed to safely transport gene-editing tools into specific cells.
The company says its NanoGalaxy platform uses AI to analyze data and identify how chemical structures correlate with specific tissue targets, such as the eyes, liver, and nervous system. The AI then predicts what adjustments to the chemistry of the delivery vehicle can be made to deliver the payload without triggering an immune response.
GenEditBio tests ePDV in vivo in a wet lab, and the results are fed back into the AI to improve prediction accuracy for the next round.
Efficient tissue-specific delivery is a prerequisite for in vivo gene editing, Zhu says. She claims her company’s approach lowers the cost of goods and standardizes processes that have historically been difficult to scale.
“It’s like getting an off-the-shelf drug.” [that works] “It can be used in multiple patients, making this drug more affordable and available to patients around the world,” Zhu said.
Her company recently received FDA approval to begin a clinical trial of CRISPR therapy for corneal dystrophy.
Dealing with persistent data issues
As with many AI-driven systems, advances in biotechnology will eventually run into a data problem. Modeling edge cases in human biology requires much higher quality data than is currently available to researchers.
“We still need more real data from patients,” Alipar said. “The corpus of data is heavily skewed towards the Western world where it is generated. I think we need to do more work on the ground to have a more balanced set of original data, ground truth data, so that our models can also handle it better.”
Aliper said Insilico’s automated lab generates multi-layered biological data from disease samples at scale and feeds it into an AI-driven discovery platform without human intervention.
Zhu says the data needed by AI already exists within humans, having been formed through thousands of years of evolution. Only a small portion of DNA directly “codes” for proteins, and the rest acts like instructions for how genes work. That information has historically been difficult for humans to interpret, but it is becoming increasingly accessible to AI models, including recent efforts like Google DeepMind’s AlphaGenome.
GenEditBio applies a similar approach in the lab, testing thousands of delivery nanoparticles in parallel instead of one at a time. The resulting dataset, which Zhu calls “the gold of AI systems,” is used to train models and even support collaborations with external partners.
One of Aliper’s next big efforts will be building digital twins of humans to run virtual clinical trials, but he says this process is “still in its early stages.”
“The number of drugs the FDA approves each year has plateaued at about 50, and needs to increase,” Alipar said. “As the world’s population ages, chronic diseases are on the rise. In 10 to 20 years, we hope to see more tailored treatment options for patients.”
