Nvidia announces plans to produce AI supercomputers at new Texas plants

Manufacturing News

Nvidia will produce AI supercomputers in the U.S. for the first time. Getty Images

Nvidia announced Monday that it will produce its artificial intelligence supercomputers in the United States for the first time.

The tech giant said it has commissioned more than 1 million square feet of manufacturing space to build and test its specialized Blackwell chips in Arizona and AI supercomputers in Texas — part of an investment the company said will produce up to half a trillion dollars of AI infrastructure in the next four years.

“The engines of the world’s AI infrastructure are being built in the United States for the first time,” Nvidia founder Jensen Huang said in a statement. “Adding American manufacturing helps us better meet the incredible and growing demand for AI chips and supercomputers, strengthens our supply chain and boosts our resiliency.”

Nvidia’s announcement comes as the Trump administration has said that tariff exemptions on electronics like smartphones and laptops are only a temporary reprieve until officials develop a new tariff approach specific to the semiconductor industry.

White House officials, including President Donald Trump himself, spent Sunday downplaying the significance of exemptions that lessen but won’t eliminate the effect of U.S. tariffs on imports of popular consumer devices and their key components.

“They’re exempt from the reciprocal tariffs but they’re included in the semiconductor tariffs, which are coming in probably a month or two,” U.S. Commerce Secretary Howard Lutnick told ABC’s “This Week” on Sunday.

Nvidia said in a post on its website that it has started Blackwell production at Taiwan Semiconductor Manufacturing Co. chip plants in Phoenix. The Santa Clara, California-based chip company is also building supercomputer manufacturing plants in Texas — with Foxconn in Houston and Wistron in Dallas.

Nvidia's AI super computers will serve as the engines for AI factories, “a new type of data center created for the sole purpose of processing artificial intelligence,” the company said, adding that manufacturing in the U.S. will create “hundreds of thousands of jobs and drive trillions of dollars in economic security over the coming decades."

Mass production at both plants is expected to ramp up in the next 12-15 months, Nvidia said. The company also plans on partnering with Taiwan-based company SPIL and Amkor for “packaging and testing operations” in Arizona.

In a statement Monday, the White House called Nvidia’s move “the Trump Effect in action.”

Trump “has made U.S.-based chips manufacturing a priority as part of his relentless pursuit of an American manufacturing renaissance, and it’s paying off — with trillions of dollars in new investments secured in the tech sector alone,” the White House said.

Earlier this year, Trump announced a joint venture investing up to $500 billion for infrastructure tied to artificial intelligence by a new partnership formed by OpenAI, Oracle and SoftBank. The new entity, Stargate, was tasked with building out data centers and the electricity generation needed for the further development of the fast-evolving AI in Texas, according to the White House.

The initial investment is expected to be $100 billion and could reach five times that sum.

Matthew Lamont is managing director at DownUnder GeoSolutions' which just opened its new, powerful data center west of Houston. Courtesy of DUG

Massive data center officially opens just west of Houston

Now online

DownUnder GeoSolutions has officially opened its new data centre in Skybox Houston in Katy, Texas. It's being billed as one of the most powerful supercomputers on earth.

The center, which houses DUG's geophysical cloud service, DUG McCloud, celebrated its grand opening on Thursday, May 16. The company's data hall has 15 megawatts of power and resides in a building designed to withstand hurricane-force winds up to 190 mph.

A second, identical hall is already planned to be built out later this year. Together, the two machines will have a capacity of 650 petaflop, which is a measurement of computing speed that's equal to one thousand million million floating-point operations per second.

In addition to the second hall, DUG is working to build another giant computing system with exaflop capacity — a billion billion calculations per second — by 2021.

"We are in a race to build the first exascale supercomputing system," says Phil Schwan, CTO for DUG, in a news release.

Australia-based DUG first started construction on Bubba, the nickname for the machine, last year and chose Skybox Datacenters as the facility to put Bubba in after a global search. The supercomputer landing in Houston represented the largest data center transaction in the Houston area's history. Dallas-Fort Worth, Austin, and San Antonio have long overshadowed Houston as hotspots for data center activity in Texas.

An differentiating asset of Bubba is the cooling process, which reduces energy usage and costs. Thirteen miles of pipes connect the hard drives to 20-foot cooling towers. Bubba uses "its own patented immersion system that submerges the computer nodes in more than 700 specially-designed tanks filled with polyalphaolefin dielectric fluid," according to the release.

"The complete DUG Insight software suite is available, and is fully-optimised to run on the cloud," says DUG's managing director, Matthew Lamont.

DUG's device is based on Intel® Xeon® processors, and the company uses Intel's technology to enhance its services, and there are more than 40,000 Intel Xeon processor nodes within the DUG McCloud network.

"The close collaboration between our two companies ensures DUG customers have access to the compute resources needed to obtain more meaningful insights from the geophysical landscapes they are exploring," says Trish Damkroger, vice president and general manager of Intel's Extreme Computing Organization, in a release.

"The Bubba supercomputer is a tremendous addition to the DUG McCloud network, and we look forward to our continued collaboration to build even more powerful systems to help accelerate this research and development."

Super-sized supercomputer

Natalie Harms/InnovationMap

Bubba, as the machine is called, has 15 megawatts of power and resides in a building designed to withstand hurricane-force winds up to 190 mph.

DownUnder GeoSolutions, which has its U.S. headquarters in Houston, is getting ready to flip the switch on what is being billed as the world's fastest supercomputer. Photo via DUG.com

World's fastest supercomputer is getting ready to power on in Houston

Booting up

An Australian company that provides geoscience and tech services to the oil and gas industry is gearing up to flip the switch in Katy on what's being billed as the world's fastest supercomputer.

At the 20-acre Skybox Houston data center campus in the Energy Corridor, DownUnder GeoSolutions is assembling a 15-megawatt data center that will house more than 40,000 servers to create the world's fastest supercomputer. Houston is the U.S. headquarters for DUG.

The data center will power a cloud computing service, known as DUG McCloud, that's tailored to the geosciences sector. The company says DUG McCloud will supply "enormous" computing capacity and high-performance storage for DUG's cloud business.

Construction on DUG McCloud — which has been delayed due to recent heavy rains — is set to be completed in April, according to the company's blog.

"DUG McCloud will be available to external companies to expand their computational resources on demand," the company says on its blog. "In addition, the cloud service will give clients access to DUG's proprietary software, with the option of source code, to accelerate their research, development, and production."

DUG McCloud is being touted as the world's biggest cloud computing service for the oil and gas industry. Among its prospective clients are global oil companies, government-owned oil producers, seismic contractors, and data companies.

"DUG McCloud is offering a wide range of companies the opportunity to significantly accelerate their oil and gas projects with cutting-edge geophysical software, stacked with extraordinary supercomputer power and services," Mick Lambert, the newly hired manager of DUG McCloud, said in December.

So, just how extraordinary will DUG's new supercomputer be?

DUG's equipment — contained in a building designed to withstand hurricane-force winds up to 190 mph — will offer more than 250 single-precision petaflops of computing speed, or 250,000 trillion calculations per second.

For now, the world's fastest supercomputer is Summit, a collaboration between the U.S. Department of Energy and IBM. Its top speed is 200 petaflops. Summit operates at Oak Ridge National Laboratory in Tennessee.

Over the long term, DUG envisions its data center being able to handle exascale computing, capable of generating at least 1 quintillion calculations per second. A quintillion has twice as many zeroes as a billion does. China is set to debut the world's first exascale supercomputer in 2020 — a year ahead of the first one to be established in the U.S., a $500 million public-private project called Aurora being developed by the Department of Energy and Intel.

DUG's deal for its data center in Katy represents the largest data center transaction in the Houston area's history. Dallas-Fort Worth, Austin, and San Antonio have long overshadowed Houston as hotspots for data center activity in Texas.

Matthew Lamont, co-founder of DUG, said in October that the company conducted an "exhaustive" search for the data center. "Houston was a natural choice," he said, "given the low cost of power and the fact that Skybox had the available infrastructure ready to go."

A unique feature of DUG's data center is how the servers will be cooled. The company's patent-pending DUG Cool system will immerse all of the servers in custom-designed tanks filled with an environmentally friendly cooling fluid.

DUG says this fluid enables condensed water-cooling chillers to be used to cool the servers, rather than server fans and refrigeration units. This will reduce energy consumption by 45 percent compared with traditional air-cooled systems, according to DUG.

"We like to call it the greenest cloud service in the world," Lamont said on DUG's blog. "DUG McCloud certainly offers more than just a silver lining."

The DUG center represents about 65 percent of the 23 megawatts of data center space under construction in the Houston area, according to a new report from commercial real estate services company CBRE.

"As high-performance computing continues to grow in importance to the energy sector, it is likely that additional latency-sensitive deployments will grow in the Houston market," Haynes Strader, senior associate at CBRE, says in a news release.

"Latency-sensitive" refers to the need for technology to act quickly in response to various events.

Ad Placement 300x100
Ad Placement 300x600

CultureMap Emails are Awesome

Property Showcase

Axiom Space taps solar array developer for first space station module

space contract

Houston-based Axiom Space is making progress on developing its commercial space station.

The company awarded Florida-based Redwire Corporation a contract to develop and deliver roll-out solar array (ROSA) wings to power the Axiom Payload Power Thermal Module (AxPPTM), which will be the first module for the new space station.

AxPPTM will initially attach to the International Space Station. AxPPTM will later separate from the ISS and rendezvous with Axiom’s Habitat 1 (AxH1) on orbit. Eventually, an airlock, Habitat 2 (AxH2) and finally the Research and Manufacturing Facility (AxRMF) will be added to the first two Axiom modules.

AxPPTM is anticipated to launch toward the end of 2027. The two-module station (AxPPTM and AxH1) is expected to be operational as a free-flying station by 2028, and the full four-module station around 2030.

The modules will be integrated and assembled at Axiom Space’s Assembly and Integration facility, making them the first human-rated spacecraft built in Houston.

Redwire’s ROSA technology was originally developed for the ISS, according to Space News. It has yielded a 100 percent success rate on on-orbit performance. The technology has also been used on NASA’s Double Asteroid Redirection Test mission, the Maxar-built Power and Propulsion Element for the Artemis Lunar Gateway and Thales Alenia Space’s Space Inspire satellites.

“As a market leader for space power solutions, Redwire is proud to be selected as a strategic supplier to deliver ROSAs for Axiom Space’s first space station module,” Mike Gold, Redwire president of civil and international space, said in a news release. “As NASA and industry take the next steps to build out commercial space stations to maintain U.S. leadership in low-Earth orbit, Redwire continues to be the partner of choice, enabling critical capabilities to ensure on-orbit success.”

Greentown Houston to add new AI lab for energy startups

AI partnership

Greentown Labs has partnered with Shoreless to launch an AI lab within its Houston climatetech incubator.

"Climatetech and energy startups are transforming industries, and AI is a critical tool in that journey," Lawson Gow, Greentown's Head of Houston, said in a news release. "We're excited to bring this new offering to our entrepreneurs and corporate partners to enhance the way they think about reducing costs and emissions across the value chain."

Shoreless, a Houston-based company that enables AI adoption for enterprise systems, will support startups developing solutions for supply-chain optimization and decarbonization. They will offer Greentown members climate sprint sessions that will deliver AI-driven insights to assist companies in reducing Scope 3 emissions, driving new revenue streams and lowering expenses. Additionally, the lab will help companies test their ideas before attempting to scale them globally.

"The future of climatetech is intertwined with the future of AI," Ken Myers, Founder and CEO of Shoreless, said in a news release. "By launching this AI lab with Greentown Labs, we are creating a collaborative ecosystem where innovation can flourish. Our agentic AI is designed to help companies make a real difference, and we are excited to see the groundbreaking solutions that will emerge from this partnership."

Greentown and Shoreless will collaborate on workshops that address industry needs for technical teams, and Shoreless will also work to provide engagement opportunities and tailored workshops for Greentown’s startups and residents. Interested companies can inquire here.

Recently, Greentown Labs also partnered with Los Angeles-based software development firm Nominal to launch the new Industrial Center of Excellence at Greentown's Houston incubator. It also announced a partnership with Houston-based EnergyTech Nexus, which will also open an investor lounge on-site last month. Read more here.

---

This article originally appeared on our sister site, EnergyCapitalHTX.com.

Houston medical institutions launch $6M kidney research incubator

NIH funding

Institutions within Houston’s Texas Medical Center have launched the Houston Area Incubator for Kidney, Urologic and Hematologic Research Training (HAI-KUH) program. The incubator will be backed by $6.25 million over five years from the National Institutes of Health and aims to create a training pipeline for researchers.

HAI-KUH will include 58 investigators from Baylor College of Medicine, Texas Children’s Hospital, the University of Texas Health Science Center at Houston, University of Houston, Houston Methodist Research Institute, MD Anderson Cancer Center, Rice University and Texas A&M University Institute of Biosciences and Technology. The program will fund six predoctoral students and six postdoctoral associates. Trainees will receive support in scientific research, professional development and networking.

According to the organizations, Houston has a high burden of kidney diseases, hypertension, sickle cell disease and other nonmalignant hematologic conditions. HAI-KUH will work to improve the health of patients by building a strong scientific workforce that leverages the team's biomedical research resources to develop research skills of students and trainees and prepare them for sustained and impactful careers. The funding comes through the National Institute of Diabetes and Digestive and Kidney Diseases.

The principal investigators of the project include Dr. Alison Bertuch, professor of pediatric oncology and molecular and human genetics at BCM; Peter Doris, professor and director of the Institute of Molecular Medicine Center for Human Genetics at UT Health; and Margaret Goodell, professor and chair of the Department of Molecular and Cellular Biology at Baylor.

“This new award provides unique collaborative training experiences that extend beyond the outstanding kidney, urology, and hematology research going on in the Texas Medical Center,” Doris said in a news release. “In conceiving this award, the National Institute of Diabetes and Digestive and Kidney Diseases envisioned trainee development across the full spectrum of skills required for professional success.”

Jeffrey Rimer, a professor of Chemical Engineering, is a core investigator on the project and program director at UH. Rimer is known for his breakthroughs in using innovative methods in control crystals to help treat malaria and kidney stones. Other co-investigators include Dr. Wolfgang Winkelmeyer (Baylor), Oleh Pochynyuk (UTHealth), Dr. Rose Khavari (Houston Methodist) and Pamela Wenzel (UT Health).

“This new NIH-sponsored training program will enable us to recruit talented students and postdocs to work on these challenging areas of research,” Rimer added in a release.

View All Listings