Alchemist Accelerator is showcasing a new array of AI-driven companies in their demo today, if you’re interested in tuning in. Additionally, the accelerator is expanding its footprint internationally into Tokyo and Doha. Continue reading for our curated selections from this latest cohort.
In a pre-demo day conversation with Alchemist’s CEO and founder, Ravi Belani, it became apparent that ambitions for AI startups have become more focused, which could actually be beneficial.
It’s increasingly unrealistic for early-stage startups to aspire to become the next OpenAI or Anthropic given their significant head start in foundational large language models.
“The costs for developing a basic LLM are astronomically high, reaching into the hundreds of millions just to get it off the ground. The question for startups is, how do you compete?” Belani explained. “VCs aren’t interested in companies that merely serve as wrappers around LLMs. We’re looking for ventures with a vertical strategy that own the end user, create network effects, and ensure long-term lock-in.”
This viewpoint aligns with my own observation that the startups chosen for this group are highly specialized in their applications. They leverage AI to address specific issues within distinct domains.
Take healthcare, for example, where AI models have begun assisting in diagnoses, care planning, and other areas. While these innovations are being tested cautiously due to concerns about liability and bias in this heavily regulated industry, they also present opportunities to replace outdated processes with significant benefits.
One notable example is Equality AI, which isn’t aiming to revolutionize cancer care but rather to ensure that AI models comply with non-discrimination regulations. This is crucial because any diagnosis or care model found to be biased against a protected class—such as assigning higher risk to someone based on religion or sexual orientation—could result in product failure and legal ramifications.
Would you trust the model maker or vendor with such significant responsibility? Or would you rather rely on a neutral specialist who understands the policies thoroughly and knows how to evaluate a model rigorously?
Maia Hightower, CEO and founder, emphasized the necessity for trustworthy AI in the healthcare sector when speaking to TechCrunch. “Ensuring that the AI driving medical decisions is both safe and effective is a fundamental right,” she said. The healthcare industry faces significant challenges in keeping pace with an intricate regulatory framework and swiftly evolving AI technologies. In the coming years, the pressure for AI compliance and the potential for legal ramifications will intensify, propelling the adoption of responsible AI practices in healthcare. The threat of substantial penalties, including the loss of certification, underscores the urgency of our solution.”
Cerevox is similarly focused on mitigating AI errors, such as hallucinations, within current large language models (LLMs). Their approach involves collaborating with companies to refine data pipelines and structures, targeting the reduction of such AI model inaccuracies. This effort goes beyond merely correcting something like ChatGPT’s fictitious accounts of historical figures; it’s about ensuring a risk evaluation engine doesn’t misinterpret incomplete data.
The company initially partners with fintech and insuretech firms—a practical but perhaps not glamorous first step. Belani noted, “It may not be the most alluring application, but it’s a viable path to developing a marketable product,” highlighting the importance of attracting paying customers to sustain a business.
Quickr Bio leverages the advancements in biotech, particularly those driven by Crispr-Cas9 gene editing, which present both new opportunities and risks. Accurate verification of genetic edits is crucial due to regulatory and liability concerns. Quickr Bio asserts that their method to measure and understand genetic modifications is up to 100 times faster than current techniques. Their focus isn’t on creating a new paradigm but rather on being the most effective solution within the existing framework. If they can substantiate their efficacy claims, they could become indispensable in various labs.
Additional details and the broader scope of this innovation can be explored further, with demos starting at 10:30 a.m. Pacific.
The program has garnered significant engagement, especially for its operations in Tokyo and Doha. “Japan represents a pivotal point, promising rich sources of innovation,” Belani stated. Recent tax policy changes are poised to unlock early-stage capital, and as investment shifts away from China, Japan, particularly Tokyo, is positioned to emerge as a renewed tech hub. This view is bolstered by OpenAI’s satellite expansion in the area.
Mitsubishi and the Japan External Trade Organization are notable investors in this initiative. The resurgence of Japan’s startup ecosystem is worth monitoring.
The Alchemist program in Doha is receiving a $13 million injection from the government, coupled with a unique focus. “The goal is to support founders from emerging markets, addressing the 90% of the global population often overlooked by technological advancements,” Belani explained. Highlighting the tendency for exceptional U.S. companies to have international roots, he underscored the value of diverse perspectives. The program aims to offer larger investments, ranging from $200,000 to $1 million, potentially attracting a different caliber of participants.
