8 MIT Startups to watch in 2026

The MIT Startup Exchange featured eight ventures in a recent Virtual Demo Day that span the health care, e-commerce, energy, and artificial intelligence industries. Each was founded or is led by an MIT alumnus or faculty member. See what each of these startups is doing to solve some of the world’s most challenging problems.

Blue Sarah

Market Pain Point: The convenience of online purchasing across various websites leads to both cluttered email inboxes and the risk of encountering counterfeit goods. While both consumers and brands are looking for alternatives to Amazon, the one-stop-shop model retains strong appeal.

Solution: A mobile app that simplifies the shopping process by enabling consumers to consolidate multiple carts and complete a single checkout for all items. 

How It Works: Blue Sarah uses AI to deliver a tailored search experience that ensures the exclusion of counterfeit products. Advertisements are placed into a dedicated area in its app, allowing users to visit for inspiration or product discovery on their own terms. Brands receive 100% of the sale and are listed as the merchant of record. Blue Sarah charges customers $5 to check out from a single cart.

Evoloh

Market Pain Point: Electrolysis for hydrogen production faces grid constraints that limit the large electricity supply needed, and the high cost and poor efficiency of electrolyzer stacks account for a third or more of an electrolyzer plant’s total cost. Consequently, many hydrogen projects are canceled before a final investment decision is reached. 

Solution: Building low-cost electrolyzers to create more cost-effective electrochemicals. 

How It Works: Evoloh’s direct connection to solar power eliminates grid connection queues and fees, reducing capital costs and making small-scale projects more economical. Building electrolyzers using common, inexpensive materials, like stainless steel, aluminum, and plastic for the casing and stainless steel/hydrocarbon membranes for the cells, lowers the company’s material costs.

Manolin

Market Pain Point: Aquaculture — such as fish farming — is the world’s fastest-growing food production system, but the industry suffers from significant shortcomings in farming system technologies, genetics research, and biotech development. Essential infrastructure like commodity markets, insurance, and real-time risk analytics are also lacking.

Solution: A data platform that integrates fish health and oceanographic data to accelerate the feedback loop between farmers and suppliers.

How It Works: Manolin uses data from satellites, ocean sensors, cameras, and farm records to deliver insights that are packaged into two integrated platforms — Watershed for farmers and Harpoon for suppliers. The company’s data engine restructures raw farm data into detailed health and performance timelines to help farmers understand what works and why. 

For example, Manolin helped Hofseth and Cargill improve performance by combating sea lice, a major salmon-farming challenge. Using Manolin’s predictive models, seafood company Hofseth anticipated sea lice pressure weeks in advance, enabling it to administer better-timed treatments. This has resulted in over $5 million in direct revenue gains by allowing Hofseth to bring more fish to market, according to Tony Chen, co-founder and CEO of Manolin.

Provocative

Market Pain Point: Provocative addresses pain points in two different markets simultaneously: 

• AI data centers: Cloud and data center providers face a dual financial squeeze when building AI infrastructure. First, they’re pouring cash into construction and equipment purchases, which forced them to borrow a record $182 billion in 2025. Second, smaller players struggle with investor skepticism, which means they’re paying 9% –12% interest rates while their bonds lose value in secondary markets. AI chips generate heat that requires expensive cooling systems, adding to operational costs while wasting energy.
• Beverage-grade carbon dioxide: The supply of beverage-grade CO₂ is under considerable strain.

Solution: A carbon capture cooling system that uses the AI chips’ waste heat to extract beverage-grade CO₂. 

How It Works: Provocative uses a cooling system with modular, lightweight vacuum-chamber carbon capture filters. Its temperature-pressure swing process uses heat from the chips for CO₂ absorption. The filters’ thin steel shells make them light enough for manual handling. In dense, liquid-cooled setups, the hot computer is surrounded by heat-evacuating containers. The carbon capture system integrates into the cooling loop, using the water heated by the chips to warm the filters and release CO₂. Provocative can harvest about 40% of the chips’ waste heat to extract beverage-grade CO₂. The company projects that a 5-megawatt facility could generate significant revenue from CO₂ sales, with beverage-grade CO₂ selling for $600 – $1,000 per metric ton amid severe supply shortages.

Satellite Bio

Market Pain Point: Severe liver diseases damage the primary cells in the liver, called hepatocytes, shifting them from a healthy to an unhealthy state. This causes toxins, including ammonia, to build up in the blood. These toxins affect the brain, leading to symptoms resembling those of dementia due to compromised brain function. Over a third of cirrhosis patients develop hepatic encephalopathy, which has a 50% one-year mortality rate. 

Solution: A therapy that delivers an active hepatocyte to patients who need a boost in their liver function, without the need for a transplant. 

How It Works: Satellite Bio has two methods for cell-based liver support. SB-101 is a direct liver infusion, where cells engraft and boost liver function. For cirrhotic livers that are too scarred for a direct infusion, Satellite Bio uses a 100-micron tissue seed to create satellite liver tissue outside the liver. The cryopreserved drugs are centrally manufactured, globally distributed for trials and customers, and offer a low cost of goods (single-digit thousands of dollars) compared with typical cell therapies (hundreds of thousands of dollars or higher).

Entrepreneurship Development Program

In person at MIT Sloan

Register Now

Tessel Biosciences

Market Pain Point: Despite advances in biopharma, conditions like chronic obstructive pulmonary disease, severe asthma, and Crohn’s disease remain largely treatment-resistant. The best discovery models — patient-derived cells grown as organoids that mimic real tissue behavior — are too complex and expensive to use early in the drug discovery process. Currently, such cells are used only to generate regulatory data in later stages of research rather than where they’d have the most impact: finding and validating new therapeutic targets.

Solution: An AI-powered active learning platform that simulates tissue behavior in silico and then iteratively tests only the most promising targets in vitro using methods like CRISPR — a substantially more efficient approach than traditional brute-force screening. 

How It Works: Scalable, complex, patient-derived tissue models are created in the lab, and AI is used to predict the most promising therapeutic interventions. This reduces the experimental burden and cost. The technology has already led to the development of a drug candidate targeting goblet cell hyperplasia, which causes the excess mucus production that severely affects breathing in patients with controlled COPD, asthma, and bronchiectasis.

Undermind

Market Pain Point: Determining whether a new research idea is novel or interesting and how it compares to existing literature requires significant time and effort. Research texts are often complex and dense, making the sifting process challenging for scientists.

Solution: An AI platform that conducts a careful, optimized search of academic research and compresses the findings into a digestible report. 

How It Works: Unlike many of the deep research tools that summarize a few search results using a large language model, Undermind’s system uses curve diagrams to illustrate a saturation effect. This means that its AI agents continue searching until statistically modeled progress confirms that the search is complete. The AI platform also rescans databases to identify new literature, to keep the researcher updated on relevant information.

Veir

Market Pain Point: AI growth is driving massive electricity demand that conventional power distribution systems can’t support. Copper cables and busbars cannot deliver the power densities needed for emerging megawatt-class AI racks, making current data center designs archaic. The industry needs a way to deliver much higher power quickly, compactly, and efficiently.

Solution: Superconducting power cables that carry 10 times more power than copper, with near-zero electrical losses, enabling higher power densities than conventional conductors. Their compact, modular design simplifies data center construction, reducing build costs by 40% and accelerating timelines by three to six months, according to Sasha Ishmael, one of Veir’s principal engineers.

How It Works: Liquid nitrogen-cooled superconducting cables replace traditional copper distribution in a data center campus. One cooling system supports multiple cables, reducing the infrastructure footprint, eliminating heat dissipation, and minimizing cooling requirements. Veir’s STAR — Superconducting Technology for AI Racks — system delivers 3 megawatts of power through a single low-voltage cable.

MIT Startup Exchange actively promotes collaboration and partnerships between MIT-connected startups and industry, exclusively for members of MIT’s Industrial Liaison Program. MIT Startup Exchange and the ILP are integrated programs of MIT Corporate Relations.

“MIT-connected” startups are based on licensed MIT technology or are founded by MIT faculty, staff, or alumni. Currently, over 800 startups are registered with MIT Startup Exchange, and monthly additions are helping to shape and define an innovative and entrepreneurial community.