Power & Sites · United States

Nvidia-backed data centre firm scraps IPO as AI valuation concerns deepen

Credited to BBC · bbc.co.uk · 9h ago

Photograph · BBC

Artificial intelligence (AI) data centre company Firmus has scrapped its plans for what would have been one of Australia's biggest-ever stock market listings.

The Nvidia-backed firm said it had made the decision due to "recent market volatility and prevailing market conditions" and that going public would not be in the company or shareholders' best interests.

Firmus had initially announced plans for a stock market debut that valued the company at more than $30bn (£22.65bn).

One investment firm told the BBC that it had decided not take part in the initial public offering (IPO) over concerns about its valuation.

"Firmus will now pursue capital from the private markets and consider alternative public and private market options. We will provide additional information to shareholders as those options progress," the company said.

Firmus builds and operates liquid-cooled data centres, or what it calls "AI factories", for clients including OpenAI and Meta.

It has operations in Australia, Singapore and other parts of the Asia-Pacific region.

The company's backers include Nvidia and major investment firms Blackstone and Jane Street.

Blackstone declined to comment when contacted by the BBC.

Nvidia and Jane Street have also been contacted for comment.

The decision by Firmus to scrap its stock market listing comes as investors and industry analysts have raised concerns about the hundreds of billions of dollars being poured into AI as the prospects for long-term returns remain unclear.

A humming annoyance or jobs boom? Life next to 199 data centres

AI data centres are booming in Australia - but at what cost?

UniSuper, one of Australia's biggest pension funds, was among the institutional investors that decided not take part in the IPO.

"We think that Firmus indeed has a compelling story. It just doesn't have a compelling valuation," UniSuper's chief investment officer John Pearce said in an update to investors.

He also said UniSuper was concerned that Firmus would have to go further into debt to fund its growth plans.

"It's disappointing. The [Australian Securities Exchange] needs new stories and this could have been one if it was correctly priced," Pearce told the BBC.

This kind of listing is a test of how willing companies are to keep "ploughing money" into AI, said Phillip Wool, chief research officer from Rayliant Investment Research.

Investors in Firmus would be buying into a company that is still in its early stages and is likely to need to borrow large sums of money, Wool said.

"It's more like a bet on a dream", he added.

Australia has become an attractive destination for data centre investment, thanks to the country's abundant clean energy, natural gas supplies and land availability.

Australia has become an attractive destination for data centre investment

OpenAI boss Sam Altman said earlier this year that Australia could become a world leader in the data centre industry if it wanted to.

There are currently over 160 data centres in the country and plans for more.

That is despite resistance from some Australians who have raised concerns about their environmental impact and the noise they create.

In September, Altman said OpenAI did not aim to list on the stock market this year, citing concerns over the technology's safety that make it "an ill-advised moment" to go public.

The ChatGPT-maker, along with rival Anthropic, have been eyeing blockbuster stock market debuts that would value the firms at more than $1tn each.

AI-related stocks, Nvidia and Oracle, fell in US trading on Thursday after reports that OpenAI's revenues were lower than previously thought.

Firmus Explores $3 Billion Private Round After IPO Flop

Credited to Bloomberg$ · bloomberg.com9h ago

Useful

Firmus Grid Ltd. is exploring a private funding round to raise as much as $3 billion after the collapse of its planned initial public offering left the data center operator looking for other financing options.

View the article · $Bloomberg

Front Method

Power & Sites · United States

Data Center Energy Trends Force a Rethink of Chip Power Delivery

Credited to SemiWiki · semiwiki.com · 23h ago

Intel Foundry published a two article series on chip power delivery that are worth reading. Here is a quick summary:

Artificial intelligence is changing the requirements for semiconductor innovation. Faster transistors and denser packaging remain essential, but their benefits increasingly depend on another engineering challenge: delivering stable power to the places where computation happens. Intel Foundry’s two-part series on power delivery explains why electricity distribution and voltage regulation are becoming central to the future of data center performance.

The scale of projected demand makes this shift significant. In the first article, Intel Foundry cites Gartner estimates that worldwide data center power demand will rise from 104 gigawatts in 2025 to 132 gigawatts in 2026, reaching 290 gigawatts by 2030. AI-optimized servers are expected to account for 31% of consumption in 2026 and surpass conventional servers by 2027.

These are forecasts, but they highlight a practical constraint: computing expansion depends on available electricity and cooling capacity. Operators must divide limited power budgets among processors, memory, networking, and storage. Improving how electricity travels through compute hardware could help increase useful output within those budgets. That makes power delivery relevant to infrastructure economics as well as chip design.

Conventional frontside power delivery routes electricity and signals through shared metal layers. As designs become more complex, those functions compete for space. Rising current increases voltage losses through resistance, while inductance complicates responses to sudden workload changes. Engineers compensate with wider connections, additional capacitors, and operating margins, consuming resources and potentially sacrificing efficiency to maintain reliable operation.

Intel Foundry’s RibbonFET Gate-all-around (GAA) transistor architecture with Intel 18A, separates these functions by moving power routing to the chip’s backside. Dedicated power layers create shorter paths to transistors and free frontside resources for signals. Intel says this arrangement reduces resistance, inductance, and IR drop, the voltage lost as current passes through resistance, while improving the stability of the supply.

Local energy storage complements that architecture. Intel Foundry describes Omni MIM on-die capacitors, substrate-embedded capacitors with through-silicon vias, and embedded deep trench capacitors. These components provide nearby charge when demand suddenly increases. For bursty AI inference and high-performance computing workloads, that support can help limit brief voltage dips that otherwise threaten timing or require more conservative operating settings.

The second article extends the problem to chiplets and stacked dies. Combining multiple dies enables greater bandwidth and functional integration, but also concentrates current demand and increases the number of voltage domains. Package connections must carry power alongside dense signal traffic. Consequently, advanced packaging can create electrical bottlenecks even as it solves other constraints on performance scaling.

Intel Foundry’s response is to move voltage conversion closer to the load. Distributing power at a higher voltage reduces the current required to carry a given amount of power; conversion near the compute die then supplies its lower operating voltage. This approach can reduce distribution losses, although its overall benefit depends on regulator efficiency and the complete delivery network.

The authors describe several implementations. Fully integrated voltage regulators place the final conversion stage on the processor, supported by CoaxMIL magnetic inductors designed for higher current density. Continuous capacitive voltage regulators, or C2VR, use capacitors instead of traditional inductors. Dedicated regulator chiplets offer placement flexibility, including beneath compute dies or within package structures, potentially shortening demanding power distribution paths.

Why it matters is the connection between electrical stability and usable performance. A processor cannot consistently exploit its capabilities if supply voltage fluctuates beyond acceptable limits. Better routing, localized charge storage, and nearby regulation could reduce the margins required for safe operation. That creates opportunities for more efficient computing, while helping designers manage increasingly complex systems.

Bottom line: Scaling computation now requires coordinated advances across transistors, wiring, capacitors, regulators, and packaging. Power delivery is becoming a decisive factor in whether additional silicon produces worthwhile performance within increasingly constrained energy budgets.

Driving Power Delivery Innovations for the AI Data Center Era

Intel Foundry: AI performance now depends on the whole system

ASML Has High-NA and Chipmakers Can’t Say No

High-NA EUV Moves From Experiment to Manufacturing