On May 8, I had the honor of testifying before the U.S. Senate at the “Winning the AI Race” hearing alongside OpenAI CEO Sam Altman, AMD CEO Lisa Su, and Microsoft President Brad Smith. That Senate hearing made one thing crystal clear: we are not merely competing in an AI race with China and the rest of the world. It is a race to shape the future of the 21st-century global economy.
America’s ability to lead in AI hinges on a simple but urgent question: Can we build the computing infrastructure fast enough to unleash AI’s full potential and to ensure the U.S. maintains its competitive advantage?
Right now, we are standing at a pivotal moment where AI is transforming industries and reshaping innovation. And yes, it’s a race to the top. We’re racing to build the backbone of tomorrow’s global economy—and safeguard our national security in the process.
Building the foundation for AI is a monumental task made of silicon, steel, and megawatts. Running the 21st-century economy on 20th-century infrastructure isn’t an option.
The new world of AI demands that we rebuild and repurpose the AI infrastructure. AI workloads are unlike anything the world has seen until pretty recently. They involve trillions of simultaneous calculations, demanding unprecedented computing power, advanced cooling systems, cutting-edge chip technology, ultra-high-speed networking, and accelerated storage.
Take a look at this chart showing how newer GPU generations have evolved to meet rising AI demands, increasing power densities within data centers. Not long ago, 20kW per CPU rack was impressive. Today, rack power density is rising to 150kW in a single rack. General-purpose clouds, designed to perform simpler tasks like hosting websites, simply weren’t designed for this.
I spoke about CoreWeave’s role in building the AI cloud infrastructure that powers the next generation of AI while strengthening America’s leadership on the global stage. Today, we operate over 250,000 GPUs in more than 30 data centers located across 15 states with a more recent expansion in Europe, encompassing over 1.6GW of contracted power capacity. CoreWeave itself has spent billions of dollars in the last few years building it infrastructue.
AI cloud platforms like CoreWeave—those that are purpose-built for AI from ground to cloud—are the engine of the entire AI ecosystem. In 2025, Cohere, IBM, and Mistral AI have deployed thousands of Blackwell GPUs in NVIDIA GB200 NVL72 rack-scale systems to train and run reasoning models and agentic AI.
But the infrastructure at this scale doesn’t appear overnight. It requires billions of dollars of capital, a trained workforce consisting of thousands of plumbers, electricians, and engineers, and abundant supplies of affordable, clean sources of power. While these things are required at the national level, each individual company in the AI space must develop and implement a clear vision and strategy.
CoreWeaves’ strategy, as well as the private sector’s strategy, cannot be implemented without transparent, public policies that enable investment, train the AI workforce, and facilitate the addition of power supplies required to drive the infrastructure. Simply put, business is moving as fast as we can. We need the government to move forward with us.
I recently advocated a coordinated portfolio of policies in testimony before the U.S. Senate that I believe are required for the U.S. to maintain its leadership position in the AI race. The goals of these policies include but are not limited to:
• A stable and predictable policy environment to encourage long-term investment. • National strategies to expand power generation and modernize the grid to deliver it while accelerating infrastructure development. • Strategic export controls that protect national security while allowing U.S. firms to compete for the exploding demand for AI in foreign markets. • Public-private partnerships to develop the 21st-century workforce while leveraging the strengths of business and government to keep the U.S. at the forefront of innovation.
Getting more specific, these policies include a transparent, “light touch” regulatory framework across all jurisdictions. This would avoid a patchwork of fragmented regulation across all jurisdictions, which can raise business costs and stifle investment.
We should implement careful reforms of the permitting process to expedite the development of abundant, reliable power supplies using diverse fuel sources, which are required to power the data centers of the future, and modernize the grid to deliver it. The nation requires an enormous amount of energy to power its 21st-century infrastructure.
Congress and the executive branch can develop export control policies that protect our national security while enabling U.S. firms to compete to service the relentless demand for AI in foreign markets.
We can train workers through public-private partnerships to meet the growing demands for AI, creating well-paying jobs that will exist decades from today. We need more plumbers, electricians, and engineers. Earlier this year, CoreWeave joined Princeton University, Microsoft, and the New Jersey Economic Development Authority to launch the New Jersey AI Hub. This is just one model for how academia, industry, and government can work together to build the future. We need ten more just like it.
The U.S. must also use its vast engineering and scientific resources to remain at the bleeding edge of innovation. The public and private sectors have critical roles to play in the innovation cycle. The national laboratory system is the crown jewels of our research complex. This genius has helped to create the computing power that drives AI and the energy technologies that drive our data centers. Let’s use this talent to assist in creating the next great inventions and not take this resource for granted.
If we want to shape the future of AI, we must ensure it is developed here on American soil. Without these policies and investments from both the private and public sectors, the U.S. risks losing its current edge in AI to nations moving faster and with more clarity of purpose.
To be clear, this matters for every American. There’s more than a trillion-dollar economic opportunity on the line; IDC estimates up to $20 trillion of global GDP impact by 2030. Beyond raw growth, AI-enabled breakthroughs promise to ensure economic growth and national security while revolutionizing healthcare, education, and other vital services and industries. Most importantly, AI holds the promise of creating millions of good-paying jobs anchored in American communities.
But none of this is guaranteed. The next great leap in AI will go to the country that moves fastest, builds the boldest, and works together—government, industry, academia, and our communities. Let’s build our future on our own soil.
This is not just a race for technological power, it’s a race to define the future. CoreWeave is committed to building the cloud infrastructure to drive AI into the future. We invite you to go on this journey with us.
Michael Intrator, CEO, Coreweave, 15 May 2025
The importance of Coreweave’s technological expertise.
— CoreWeave (Nasdaq: CRWV), in collaboration with NVIDIA and IBM, delivered the largest-ever MLPerf® Training v5.0 submission on NVIDIA Blackwell, using 2,496 NVIDIA Blackwell GPUs running on CoreWeave’s AI-optimized cloud platform. This submission is the largest NVIDIA GB200 NVL72 cluster ever benchmarked under MLPerf, 34x larger than the only other submission from a cloud provider highlighting the large scale and readiness of CoreWeave’s cloud platform for today’s demanding AI workloads.
The submission achieved a breakthrough result on the largest and most complex foundational model in the benchmarking suite–Llama 3.1 405B–completing the run in just 27.3 minutes. When compared against submissions from other participants across similar cluster sizes, CoreWeave’s GB200 cluster achieved more than 2x faster training performance. This result highlights the significant performance leap enabled by the GB200 NVL72 architecture and the strength of CoreWeave’s infrastructure in delivering consistent, best-in-class AI workload performance.
“AI labs and enterprises choose CoreWeave because we deliver a purpose-built cloud platform with the scale, performance, and reliability that their workloads demand,” said Peter Salanki, Chief Technology Officer and Co-founder at CoreWeave. “These MLPerf results reinforce our leadership in supporting today’s most demanding AI workloads.”
These results matter because they translate directly to faster model development cycles and an optimized Total Cost of Ownership. For CoreWeave customers, that means cutting training time in half, scaling workloads efficiently, and training or deploying their models more cost-effectively by leveraging the latest cloud technologies, months before their competitors. With leading submissions for both MLPerf Inference v5.0 and Training v5.0 benchmarks and the sole cloud provider ranked in the Platinum tier of SemiAnalysis’ ClusterMAX, CoreWeave sets the standard for AI infrastructure performance across the entire cloud stack.
Coreweave news, 4 June 2025
Coreweave is a company moving at great speed in terms of sheer scale, in terms of technological expertise and in terms of its impact on America and the world. This is one of the most explosive growth stories I have ever encountered. Coreweave is going to become a stock market giant in the blink of an eye as it builds its state-of-the-art data centres/ AI factories across much of the globe.
As Intrator said in his submission to Congress, this is not something that can be done slowly. America is in a race that it must win. Companies like Coreweave need to exist to make this happen, and they are as well placed as anyone to know what is needed.
You could almost say that Coreweave needs to become huge to help the US and its allies win the race. The shares are rising rapidly (see chart) because of a dramatic change in the way the business is perceived by investors and a dawning realisation that it may continue to grow explosively for some time.
This chart does not reveal much about where Coreweave shares might go in the short term, but the explosive power of the move is consistent with the likelihood that much higher levels lie ahead, and this is the beginning of a sustained secular uptrend.
