|
|
|
|
|
The Morning Download: AI’s Energy Demands Ripple Through Economy
|
|
|
|
|
|
What's up: TSMC raises revenue view again; big banks cite AI's impact on the workforce; Walmart to deploy millions of IoT sensors.
|
|
|
|
Data centers like Equinix’s San Jose, Calif., operation are bridging the energy gap themselves. JUNGHO KIM FOR WSJ
|
|
|
|
Good morning. There’s been a lot of discussion in recent months over how AI data centers’ massive need for energy can be met, given the limitations of the power grid. Well, it seems we now have the answer. We're building more power plants, people.
The WSJ explains the magnitude of the gap that needs to be filled:
America should be adding about 80 gigawatts of new power generation capacity a year to keep pace with AI as well as cloud computing, crypto, industrial demand and electrification trends, according to consulting and technology firm ICF. It’s currently building less than 65 gigawatts. That gap alone is enough electricity to power two Manhattans during the hottest parts of summer.
And it looks like a lot of money is being spent on fixing that imbalance:
The U.S. had around 522 hyperscale data centers at the end of the second quarter, which account for around 55% of global capacity, according to Synergy Research Group. Another roughly 280 are expected to come online through 2028 in the U.S.
|
|
|
|
|
This construction effort appears to be well underway, too.
In West Texas, natural-gas-fired power generation is under construction as part of the $500 billion Stargate project from OpenAI and Oracle. Gas turbines are in use at Colossus 1 and 2, the massive data centers Elon Musk’s xAI is building in Memphis, Tenn. More than a dozen Equinix data centers across the country are using fuel cells for power.
With the push for AI dominance at warp speed, the “Bring Your Own Power” boom is a quick fix for the gridlock of trying to get on the grid. It’s driving an energy Wild West that is reshaping American power.
Coding startup Poolside and cloud-infrastructure provider CoreWeave plan to build a data center in West Texas. [More on that below].
The energy sector is pushing ahead with nuclear energy, and that has spawned an investment bubble that is tied to the broader AI investment bubble, sort of a derivative bubble. Maybe banks can securitize bubble derivatives?
A group of non-revenue-generating energy companies have collectively ballooned in value to more than $45 billion in hopes that tech companies will one day pay for their yet-to-be-built power.
The biggest of these is the OpenAI CEO Sam Altman-backed nuclear startup Oklo, whose shares have risen about eightfold year to date. The company now has a market cap of roughly $26 billion, making it the biggest U.S.-incorporated public company that generated no revenue in the past 12 months, according to data from S&P Global Market Intelligence.’
Much attention has been focused on soaring AI valuations in the financial markets. But AI is having an impact on the real economy, too, as the data-center construction boom and the data-center power plant construction boom show. That knock-on effect is bound to grow, as support services sprout up near the data-centers and power plants.
|
|
|
Content from our sponsor: Deloitte
|
|
CCS CTO on AI: Powering a Shift From ‘Sick Care’ to Prevention
|
By serving as an “always-on” trusted adviser, AI can help empower patients to manage their health care proactively, according to CCS CTO Richard Mackey. Read More
|
|
|
|
|
|
|
|
Texas becomes Ground Zero for the Next AI power boom
|
|
|
|
Poolside plans to build a data center on more than 500 acres of land in West Texas. Poolside
|
|
|
|
A new wave of mega–data center projects, from Poolside and CoreWeave’s gas-fueled West Texas complex to Meta’s $1.5 billion El Paso hub, is transforming the Lone Star State.
|
|
Coding startup Poolside and cloud-infrastructure provider CoreWeave plan to build a data center in West Texas, taking advantage of natural gas produced in the Permian Basin to power its computing resources. The companies expect to complete construction in the first quarter of 2027.
Stats. When completed the facility will be about two-thirds the size of NYC's Central Park. It will host two gigawatts of computing firepower, the electric-generation capacity of the Hoover Dam.
An unusual setup. Poolside in the near time will gain access to a cluster of Nvidia AI computing resources provided by CoreWeave beginning in December. It plans to work with the company longer term on the broader build-out of a data center.
|
|
Artificial Intelligence Infrastructure Partnership, BlackRock’s new AI infrastructure consortium, has struck its first deal, agreeing to acquire Aligned Data Centers from Macquarie Asset Management. The group, founded by BlackRock, Global Infrastructure Partners, MGX, Microsoft and Nvidia, is paying roughly $20 billion for the Dallas-based company, people familiar with the matter said.
|
|
|
|
AI Lifts Global Chip Titans
|
|
|
|
It remains unclear how much U.S. tariffs will affect the Taiwanese chip maker’s business. Ann Wang/Reuters
|
|
|
|
Taiwan Semiconductor Manufacturing Company posted an all-time high net profit for the third quarter, up about 39 % year-over-year, driven largely by booming demand for AI chips. TSMC raised its full-year revenue outlook into the mid-30% growth range, citing sustained momentum in AI infrastructure investments.
Foxconn’s stock jumped after its chairman revealed he met with OpenAI’s CEO and plans to hold talks with Nvidia to explore AI collaboration. Investors responded enthusiastically, seeing this as a signal that Foxconn aims to deepen its role in the AI ecosystem.
The upbeat results followed similar updates from other major technology companies.
Dutch tech giant ASML on Wednesday reported better-than-expected orders of its chip-making equipment, while Samsung Electronics, the world’s largest memory-chip maker, said Tuesday that it expects its highest quarterly profit in three years.
|
|
|
|
“There’s just absolute spasm of almost daily, hourly, deal making”
|
— Former Meta executive and British politician Nick Clegg speaking to CNBC on the chance of an AI market correction
|
|
|
|
|
|
|
|
|
The Wall Street side of JPMorgan’s business continued to rake in money in the latest quarter. Seeger Gray/WSJ
|
|
|
|
Big banks JPMorgan Chase and Goldman Sachs this week dedicated part of their earnings calls to address how they are reorganizing their organizations for the age of AI. Both cited a desire to keep new headcount low, even as they reported strong quarters.
“It means taking a front-to-back view of how we organize our people, make decisions, and think about productivity and efficiency, Goldman Sachs CEO David Solomon told employees in a memo, according to CNBC.
| |