The initiative, known as NJSCI, is designed to establish a shared supercomputing resource for researchers, students, entrepreneurs, businesses and organizations throughout New Jersey. The announcement described access to advanced computing as a potential bottleneck for ideas that require more computational power than ordinary laboratory, classroom or commercial systems can provide. The planned resource is expected to support work including cancer research and flood-risk mapping, as well as other research and training activities that benefit from large-scale calculations.

According to the announcement, NJSCI is expected to double the supercomputing capacity currently available across New Jersey’s research universities. That capacity increase is a planned feature rather than a completed expansion. The initiative is also intended to broaden education and training opportunities, connect university expertise with statewide needs and help prepare students and working professionals for roles in fields where artificial intelligence and data-intensive computation are increasingly important.

The partnership is backed by more than $125 million in public and private investment. The funding announced includes a $75 million personal gift from NVIDIA co-founder Chris Malachowsky and his wife, Melody, a $25 million investment from SHI International and a $25 million state investment made through the New Jersey Economic Development Authority. The announcement did not provide a detailed schedule for procurement, installation or allocation of computing resources, so the timetable and technical specifications of the supercomputer remain to be defined through the initiative’s implementation.

NJIT Interim President John A. Pelesko said the initiative could reduce the barrier between promising ideas and the computing power needed to test them. He described researchers and students as having good ideas that may wait because the necessary machine is inaccessible. NJIT’s role in NJSCI will be to help ensure that shared access is translated into research opportunities and graduates with relevant experience, according to the release. The announcement did not report discoveries, completed research projects or measured outcomes resulting from the partnership.

The institute’s participation will extend beyond its own researchers. Jamie Payton, dean of NJIT’s Ying Wu College of Computing, serves on the NJSCI Governance, Allocation and Access Working Group and the NJSCI Education and Workforce Collaborative. The collaborative is working on curriculum, faculty professional development and workforce-development efforts. Payton said the initiative represents an “AI+X” approach, combining artificial intelligence with subject-matter expertise so that complex challenges can be addressed across disciplines rather than through computation alone.

David Bader, a distinguished professor of data science, associate dean for research in the Ying Wu College of Computing and director of NJIT’s Institute for Data Science, has participated in AI supercomputer working sessions involving the Governor’s Office, NVIDIA, SHI and partner universities. He said the restriction faced by New Jersey’s AI community is often less about the availability of talent than access to computing resources at scale. The initiative is intended to give faculty and students an opportunity to work with serious hardware that is not normally available through laptops or free cloud-service tiers.

Bader also described the partnership as a model for sustained cooperation among state government, technology companies and higher education. NJIT established a Department of Data Science and an Institute for Data Science to develop expertise for a data- and AI-intensive economy, and the supercomputer initiative is expected to extend that training into courses using high-performance resources. The approach would allow students to run models at scales unavailable on personal devices, giving them practical experience that Bader said employers increasingly expect from AI specialists.

Access rules, technical architecture and the distribution of computing time were not specified in the release. The initiative does, however, include governance and allocation structures intended to determine how the shared resource can be used. As a statewide partnership involving multiple universities and agencies, NJSCI may differ from computing projects housed by a single institution, with access and priorities potentially shaped through the working groups and collaborative activities described in the announcement.

Sherrill said the initiative builds on New Jersey’s history of scientific and industrial innovation, citing Thomas Edison, Bell Labs and other examples of research and invention associated with the state. She said the investment is intended to strengthen the use of emerging technologies in scientific discovery, expand students’ access to advanced skills, widen workforce opportunities and improve communities’ capacity to address challenges such as disease and environmental risk. These benefits are stated goals of the launch; the release does not yet document outcomes from the new system.

Malachowsky, whose personal gift forms the largest component of the announced investment, said New Jersey holds special significance because it was where his story began. He said the initiative could unlock talent, transform education and business, support the workforce and bring institutions together around demanding computational problems. With its announced funding, university partnerships and workforce-development activities, NJSCI has been set up as a statewide platform for advanced computing, while its practical effects will depend on the completion and use of the shared supercomputer.

Subject of Research: New Jersey’s statewide initiative to expand shared supercomputing capacity for research, education, workforce development and AI applications.

Article Title: NJIT partners on New Jersey Supercomputer Initiative to drive statewide innovation

Article References: NJIT partners on New Jersey Supercomputer Initiative to drive statewide innovation. Original publication (Bibliographic details unavailable.)

Image Credits: AI Generated

DOI: Not provided

Keywords: New Jersey Supercomputer Initiative, NJIT, high-performance computing, AI research, workforce development, public-private partnership, artificial intelligence, supercomputing capacity, research universities, NVIDIA, partners, Jersey

The study, published in the journal Discover Agriculture by Erica Auguste of Trinidad and Tobago’s Ministry of Agriculture and Fisheries and colleagues Martin Hughes and Sivakumar Karuppusamy at the University of the West Indies, set out to answer a question with real economic stakes. Dairy farming across the tropics is dominated by smallholders, and their productivity is chronically low. In Trinidad and Tobago, national milk production collapsed from roughly 9.4 million kilograms in 2001 to about 1.6 million kilograms in 2024. Jamaica saw annual yields fall by around 57 percent over a comparable period. Understanding which biological factors actually constrain output on these farms is therefore not an academic exercise but a matter of livelihood and food security.

The Centeno Livestock Station occupies about 502 acres of flat to gently sloping land in El Carmen, St. Helena, in the county of Caroni. Its soils belong to the Cunupia series — heavy clays with impeded drainage and a pH between 4.5 and 5.4 — and the climate is classically tropical, with average annual rainfall near 1,877 millimetres and temperatures ranging from 22.7 to 31.3 degrees Celsius. A wet season runs from July to December and a dry season from January to June, during which pasture quality and quantity typically decline. The dairy herd of roughly 60 cows grazes rotationally on Brachiaria arrecta and unimproved native grasses, is milked twice daily with an automated system, and receives about 2.0 to 2.5 kilograms of an 18 percent crude protein concentrate at each milking.

Two breeds dominate the herd: the Centeno Crossbred, a Bos taurus × Bos indicus animal developed in the late 1970s by crossing Holstein-Friesians with Zebu-type breeds to combine high yield with heat tolerance and resistance to tick fever, and the Jersey. The researchers deliberately excluded a small group of graded Holstein-Friesians because so few were lactating and in poor condition. Breeding is carried out by artificial insemination, and the cows range from three to twelve years old. It was this herd, managed under consistent conditions, that provided the retrospective dataset spanning 2014 to 2019.

The team analysed the records using analysis of variance in SPSS version 27, treating calf sex, breed, parity and the breed-by-parity interaction as fixed effects, with Bonferroni-adjusted comparisons and Tukey’s test to separate means. Data normality was confirmed with the Ryan-Joiner test, and lactations shorter than 140 days in milk were excluded from total-yield calculations. Early lactation was defined as the first 90 days in milk. The statistical verdict was unambiguous: calf sex, breed and the breed-by-parity interaction had no significant effect on milk yield or lactation length.

Parity told a different story. Cows in their fourth through sixth parities produced 849 ± 66 kilograms of milk in early lactation, and those beyond their sixth parity produced 1,005 ± 116 kilograms — both figures significantly higher than the 673 ± 41 kilograms recorded for cows in their first to third parities. Daily milk yield followed the same gradient, at 5.6 ± 0.33 kilograms for the youngest parity group compared with 8.4 ± 1.0 kilograms for cows beyond their sixth lactation. The pattern makes physiological sense: young cows are still growing, so nutrients and energy must be split between body development, gestation and milk synthesis, often pushing them into a deeper negative energy balance. Mature cows, with larger mammary glands and more developed secretory tissue, can channel a greater share of their intake into milk.

Perhaps the most striking finding came from regression analysis. Early lactation milk yield accounted for 73 percent of the variability in total lactation yield — a coefficient of determination of 0.73. In other words, a cow’s output in her first three months of milking is a powerful predictor of her entire lactation. This aligns with a well-established body of literature showing that early lactation is a physiologically critical window in which milk yield peaks while energy demands often exceed dry matter intake. It also suggests that phase-specific management — concentrating feed resources and attention on freshly calved cows — could pay dividends even on smallholder, pasture-based tropical farms, where such approaches are rarely practised.

The null result on calf sex is notable because the literature on sex-biased milk synthesis is genuinely contested. Some studies, including a well-known analysis of Holsteins, reported that cows carrying female fetuses produced around 143 kilograms more milk than those carrying males, possibly because longer gestations for male calves shorten the subsequent lactation. Others found the opposite, attributing higher yields after male births to larger fetal size and elevated placental estrogen, which stimulates prolactin release — though bigger calves also raise the risk of difficult calvings that depress yield. The Centeno findings suggest that such sex effects, where they exist, are contingent on factors like calf size, dam health, parity and production system, and that within a uniform herd under consistent management, they may simply wash out.

The equivalence of the Jersey and Centeno Crossbred cows is equally consequential. Despite their different genetic backgrounds, the two breeds performed comparably under identical conditions, with the crossbreds posting numerically higher early lactation and daily yields while Jerseys showed numerically higher total yield and longer lactations. Since neither difference reached significance, the results hint that both breeds carry similar genetically linked performance characteristics when mature body size and management are held constant. For farmers choosing breeding strategies, that is reassuring: the locally adapted crossbred, bred for heat tolerance and disease resistance, gives up little in productive potential to the imported Jersey.

The sobering counterpoint is that absolute productivity at the station remains very low. Jersey cows elsewhere in subtropical climates yield around 11 kilograms of milk per day, and 20 to 25 kilograms in temperate environments — between 27 and 68 percent more than the CLS herd. Worryingly, low productivity in Centeno Crossbreds was flagged four decades ago, and the new data suggest performance may have declined further, pointing to persistent inadequate nutrition, unimproved pastures, dry-season feed gaps, costly concentrate supplementation and possible genetic depletion. The authors argue that improving pasture quality and quantity, supplemental feeding, reproductive efficiency and heat-stress management — measures that meta-analyses suggest can lift smallholder tropical milk yields by as much as 29 percent — must target cows in fourth and later parities and, above all, cows in early lactation, because that is where milk production is maximized and where the whole lactation is won or lost.

Subject of Research: Effects of calf sex, breed and parity on milk yield and lactation length in a tropical dairy herd

Article Title: Effects of calf sex, breed and parity on the lactation performance of dairy cows in Trinidad and Tobago

Article References: Auguste, E., Hughes, M., & Karuppusamy, S. (2026). Effects of calf sex, breed and parity on the lactation performance of dairy cows in Trinidad and Tobago. Discover Agriculture, 4(1), Article 295. https://doi.org/10.1007/s44279-026-00779-x

Image Credits: AI Generated

DOI: 10.1007/s44279-026-00779-x

Keywords: dairy cattle, milk yield, parity, calf sex, lactation length, early lactation, Centeno Crossbred, Jersey, smallholder farming, Trinidad and Tobago, tropical agriculture, animal breeding