Capital · United States

Jev creator TypeSafe closes $870M round at $7.5B valuation

Credited to SiliconANGLE · siliconangle.com · 18h ago

Photograph · SiliconANGLE

TypeSafe Inc., the creator of the Jev artificial intelligence model, today announced that it has raised $870 million in funding at a $7.5 billion valuation.

Andreessen Horowitz led the round with contributions from Sequoia Capital, DCVC and unnamed angel investors. The cash infusion comes less than a month after TypeSafe launched Jev. The company’s funding momentum reflects the surging popularity of its model: It has already been adopted by about a third of the Fortune 500.

When an enterprise application needs to perform a task using a large language model, it automatically sends a description of the task to the model. The LLM then responds with its output. That output often takes the form of natural language text. Applications must condense the natural language text into a structured, standardized format before they can put it to use.

Jev partly owes its popularity to the fact that it skips the latter step. When an application asks Jev to perform a task, the model generates structured output instead of natural language text. That removes the need for applications to reformat and condense Jev’s output before using it.

It supports just three types of requests. Applications can ask the model to answer a question with the equivalent of “yes” or “no,” pick an item from a list or generate a score. Developers can customize what that score measures. For example, Jev can be configured to rate cybersecurity alerts based on their severity or quantify the urgency of a support ticket.

The fact that the model’s terse output is easy for applications to process reduces the amount of data preparation code that developers must write. As a result, software projects can be completed faster. Furthermore, simplifying an application’s code base reduces the risk of errors.

Jev also boosts application reliability in other ways. When the model generates a score or chooses from a list of user-provided options, it outputs a number that indicates its confidence in the accuracy of the response. Applications can use that number to mitigate the impact of hallucinations.

TypeSafe says that it developed Jev using a new training approach called reinforcement learning for calibrated decisions. It’s a variation of reinforcement learning, a popular LLM training method. TypeSafe also developed a new model architecture.

The company claims that its custom technologies enable Jev to process user requests in under 700 milliseconds. That makes the model up to 200 times faster than some frontier LLMs. Furthermore, TypeSafe says that it’s up to 100 times more cost-efficient.

Jev is part of a planned model series dubbed System One. TypeSafe will use the proceeds from its funding round to make more additions to the lineup. Additionally, the company plans to roll out unspecified “enterprise features” that will make it easier for large organizations to use its models.

Support our mission to keep content open and free by engaging with theCUBE community. Join theCUBE’s Alumni Trust Network, where technology leaders connect, share intelligence and create opportunities.

Are you an AWS customer? Support SiliconANGLE financially by buying your AWS services from our Marketplace portal page and links: https://siliconangle.com/aws-marketplace/

Founded by tech visionaries John Furrier and Dave Vellante, SiliconANGLE Media has built a dynamic ecosystem of industry-leading digital media brands that reach 15+ million elite tech professionals. Our new proprietary theCUBE AI Video Cloud is breaking ground in audience interaction, leveraging theCUBEai.com neural network to help technology companies make data-driven decisions and stay at the forefront of industry conversations.

Oxide Computer raises $445M to step up data center rack production

Warehouse robot maker Ultra Robotics bags $62M

IBM connects enterprise AI orchestration to production readiness ahead of TechXchange

Seismora builds a control plane to route AI workloads across devices and clouds

McKinsey connects enterprise data through a knowledge graph for AI

OpenAI revenue falls short, models play hopscotch and Trump cracks down on tech green cards

INFRA - BY MARIA DEUTSCHER. 13 HOURS AGO

Join our community

CapitalUnited States

Fisk Electric wins $24M modular contract for Houston data center

Credited to Construction Dive · constructiondive.com18h ago

Useful

The Tutor Perini subsidiary’s win reflects a priority shift toward the booming data center sector that the heavy civil contractor has emphasized on earnings calls this year.

Dive Insight:

The Los Angeles-based megacontractor has historically focused on large civil infrastructure projects, including the Gateway tunnel project in New York and New Jersey and California’s high-speed rail project between Los Angeles and San Francisco.

On recent earnings calls, however, the firm has said it was exploring opportunities in the data center construction space, as the sector has boomed to represent a bigger share of large contractors’ backlogs.

Tutor Perini is not alone. Watsonville, California-based Granite Construction, another contractor focused on civil infrastructure, has also noted its growing exposure to data center projects. Its work includes site preparation and construction of access roads and parking lots at the facilities.

Fisk’s electrical construction work for the Houston data center is expected to begin this year, according to the release. The contractor anticipates substantial completion in the first quarter of 2027.

Original · Construction Dive

Front Method

Capital · United States

Black Hills to build 564 MW of new gas, manage 2.1 GW microgrid for Google data center in Wyoming

Credited to Power Engineering · power-eng.com · 22h ago

Photograph · Power Engineering

Google’s next big data center needs serious power. Black Hills Corp. is bringing the turbines — and a $1.8 billion investment — to Cheyenne, Wyoming. The South Dakota-based utility holding company would build 564 MW of new natural gas-fired generation, while managing about 2.1 GW of third-party resources to serve Google’s planned data center load. Black Hills said the project would draw on a total resource mix of about 2.7 GW, including reserve margins and excluding transmission expansion.

The new gas capacity would be built at Black Hills’ existing Cheyenne Prairie Generating Station between 2027 and 2029 for $1.8 billion. The new units would be owned by a non-regulated Black Hills affiliate, not the Wyoming utility.

But the gas expansion is only part of the power play. The project also includes 2.1 GW of third-party resources, which the project release suggests would be powered by turbines and fuel cells. Those resources would operate through a privately managed microgrid under Wyoming’s Large Power Contract Service tariff. Black Hills would not own them but instead collect a microgrid management fee for coordinating grid operations, reliability services and dispatch across the contracted resources and market purchases.

Google picks up the tab

Black Hills said Google would cover the costs of serving the data center, with contractual protections intended to shield existing utility customers from project-related expenses. Those protections address stranded assets, early termination and unexpected costs. Google has advanced $399 million for long-lead equipment purchases, which Black Hills expects to reimburse before July 2027. The utility projects about $150 million in net income from the project in 2030 and $2.4 billion in unlevered free cash flow through 2048, after accounting for its generation investment. Black Hills expects to start earning returns on the investment next year by introducing microgrid management fees. The company has secured an air quality permit for the Cheyenne Prairie expansion and is pursuing additional siting and transmission approvals.

Google’s gas deal draws criticism

There is no settled answer to what kind of generation should serve the wave of data center loads being built. Utilities facing gigawatt-scale requests have largely turned to new gas-fired capacity which they can site at existing plants.

Tech companies driving the data center demand have pledged to run on carbon-free power. Google, for example, has a goal of matching its consumption with clean energy around the clock by 2030.

There are a range of approaches between those two positions. These include nuclear uprates and restarts, long-term PPAs for solar, wind and storage, on-site or behind-the-meter (BTM) generation and hybrid arrangements like the one in Cheyenne. State regulators, grid operators and developers are seemingly working through the trade-offs between speed, reliability, cost and emissions one deal at a time.

Constellation, Vistra uprate deals signal fastest path to new nuclear megawatts

But Google’s latest agreement is drawing criticism for its reliance on new gas. Jesse Jenkins, a Princeton University energy researcher and co-founder of Firma Power, criticized the Wyoming deal as a departure from Google’s emphasis on bringing new clean electricity online. Firma Power, develops software and power procurement strategies to assemble renewable energy and storage portfolios for large electricity customers. His criticism came as Google announced a separate agreement with Constellation Energy to add 890 MW of nuclear capacity through upgrades to existing reactors — a deal Jenkins praised on X as a model for powering data centers.

Jenkins argued that data center developers should instead tap wind, solar and battery storage projects awaiting grid interconnection rather than build new gas-fired generation to meet surging electricity demand.

Yesterday I praised @Google 's deal with Constellation for 890 MW of new clean capacity from uprated reactors as a prime example of data center power done right. Unfortunately, Google also announced a separate deal yesterday with Black Hills Corp. in Wyoming that directly invests… pic.twitter.com/MW2jNEZf6k — Jesse D. Jenkins (@JesseJenkins) October 7, 2026