Machine learning has been used from everything to learning how to play Go to building better chat bots to improving search engines. Now a small, new biotech company aims to use machine learning to save lives. BioXcel Therapeutics, a subsidiary of BioXcel Corp., has filed for a $69 million IPO. In its SEC report, BioXcel says that it is “focused on drug development that utilizes novel artificial intelligence, or AI, to identify the next wave of medicines across neuroscience and immuno-oncology.”
But while BioXcel wants to portray itself as a bold innovator using AI and machine learning, its actual advertised products are a peculiar combination of conservative yet risky. BioXcel has a great idea, but execution matters more than ideas and there is little reason to think it can pull off such an ambitious dream.
BioXcel and BioXcel
The first problem with BioXcel is what it actually is. As noted above, the company which is going public is called BioXcel Therapeutics. BioXcel Therapeutics was split off from BioXcel Corp. last year, and the parent company owns 93 percent of the subsidiary. BioXcel Corp. will continue to own the majority of the company after this IPO, though the exact percentage is unknown as of this time.
BioXcel Therapeutics talks about leveraging a research and development engine called EvolverAI, but this engine is owned by BioXcel Corp. Rather than doing AI development itself, BioXcel Therapeutics exists to develop the drugs found by EvolverAI and BioXcel Corp.
It is not like BioXcel Corp. is the only biotech company interested in leveraging AI and machine learning. TechEmergence points out that medical titans like Pfizer (NYSE:PFE), Roche, and GlaxoSmithKline (NYSE:GSK) are using artificial intelligence in novel ways to discover new combinations of drugs, find new compounds, and link diseases with treatments. EvolverAI by contrast works by using millions of data points to discover new potential therapeutic uses for drugs which already exist, at which point BioXcel Therapeutics reformulates them.
So what are the drugs which EvolverAI discovered and BioXcel Therapeutics is working on? The two main ones are called BXCL501 and BXCL701, also known as dexmedetomidine and talabostat respectively. Dexmedetomidine is used today under the brand name Precedex as a sedative. Talabostat was tested as a potential treatment for lung cancer by the now defunct Point Therapeutics, but failed Phase 3 testing over a decade ago. By reformulating them, BioXcel believes that dexmedetomidine can fight Alzheimer’s and schizophrenia, and talabostat can fight pancreatic cancer.
Unfortunately, these drugs are only in the latter stages of Phase 1 testing, with plans to start Phase 2 testing later this year. BioXcel hopes that the fact that the drugs have already passed prior levels of testing in the past will help them get past the drug approval process quicker, but the reality is that investors will be waiting even by biotech companies longer than normal for these drugs to hit the market.
Long-Term Viability
Like most biotech IPOs, BioXcel’s revenue is nonexistent, and it in fact reports no revenue whatsoever in 2016 and 2017. Over those two years, it decided to procure to pay its debts and earned a net loss of $2.1 and $4.5 million respectively. BioXcel also reported $1.3 million in assets and $2.3 million in liabilities as of December 31, 2017.
Given these losses, raising $69 million may seem to be adequate to keep BioXcel running until its drugs are developed, but net losses will exponentially increase as it moves further in the development process. Unfortunately, BioXcel does not state how long it will be able to keep itself running with the raised IPO proceed, though they do concede that “we expect that we will need to obtain substantial additional funding in order to complete our clinical trials.”
But if BioXcel will need substantial additional funding, why did they choose now to go public? Plenty of other companies, biotech or otherwise, are staying private because they can continue to secure rounds of funding from private investors. BioXcel’s decision to eschew said funding and go public very quickly is strange and may have happened because said investors lack confidence in their product.
Some Good, Plenty Concerning
BioXcel Therapeutics has some things going for it compared to other biotech IPOs. Its close relationship with BioXcel Corp. means that it is not nearly as dependent on one or two drugs getting past FDA testing, especially as BioXcel Corp. will likely continue to have it test whatever drugs it discover with EvolverAI. A connection to machine learning, even it is indirect, is also a useful tool to attract investors. But it will take longer than normal for its drugs to get through the development pipeline and its disinterest or inability to secure private funding is seriously concerning. There are much better pipelines to take advantage of the growing interest in machine learning in the biotech industry. Investors may want to pay attention to what BioXcel Therapeutics has to say on the roadshow, but should not show any interest in this IPO as it currently stands.