Connect with us

Published

on

A trader works on the floor of the New York Stock Exchange.

Jason Decrow

Alejandro Lopez-Lira, a finance professor at the University of Florida, says that large language models may be useful when forecasting stock prices.

He used ChatGPT to parse news headlines for whether they’re good or bad for a stock, and found that ChatGPT’s ability to predict the direction of the next day’s returns were much better than random, he said in a recent unreviewed paper.

The experiment strikes at the heart of the promise around state-of-the-art artificial intelligence: With bigger computers and better datasets — like those powering ChatGPT — these AI models may display “emergent abilities,” or capabilities that weren’t originally planned when they were built.

If ChatGPT can display the emergent ability to understand headlines from financial news and how they might impact stock prices, it could could put high-paying jobs in the financial industry at risk. About 35% of financial jobs are at risk of being automated by AI, Goldman Sachs estimated in a March 26 note.

“The fact that ChatGPT is understanding information meant for humans almost guarantees if the market doesn’t respond perfectly, that there will be return predictability,” said Lopez-Lira.

But the specifics of the experiment also show how far so-called “large language models” are from being able to do many finance tasks.

For example, the experiment didn’t include target prices, or have the model do any math at all. In fact, ChatGPT-style technology often makes numbers up, as Microsoft learned in a public demo earlier this year. Sentiment analysis of headlines is also well understood as a trading strategy, with proprietary datasets already in existence.

Lopez-Lira said he was surprised by the results, adding they suggest that sophisticated investors aren’t using ChatGPT-style machine learning in their trading strategies yet.

“On the regulation side, if we have computers just reading the headlines, headlines will matter more, and we can see if everyone should have access to machines such as GPT,” said Lopez-Lira. “Second, it’s certainly going to have some implications on the employment of financial analyst landscape. The question is, do I want to pay analysts? Or can I just put textual information in a model?”

How the experiment worked

In the experiment, Lopez-Lira and his partner Yuehua Tang looked at over 50,000 headlines from a data vendor about public stocks on the New York Stock Exchange, Nasdaq, and a small-cap exchange. They started in October 2022 — after the data cutoff date for ChatGPT, meaning that the engine hadn’t seen or used those headlines in training.

Then, they fed the headlines into ChatGPT 3.5 along with the following prompt:

“Forget all your previous instructions. Pretend you are a financial expert. You are a financial expert with stock recommendation experience. Answer “YES” if good news, “NO” if bad news, or “UNKNOWN” if uncertain in the first line. Then elaborate with one short and concise sentence on the next line.”

Then they looked at the stocks’ return during the following trading day.

Ultimately, Lopez-Lira found that the model did better in nearly all cases when informed by a news headline. Specifically, he found a less than 1% chance the model would do as well picking the next day’s move at random, versus when it was informed by a news headline.

ChatGPT also beat commercial datasets with human sentiment scores. One example in the paper showed a headline about a company settling litigation and paying a fine, which had a negative sentiment, but the ChatGPT response correctly reasoned it was actually good news, according to the researchers.

Lopez-Lira told CNBC that hedge funds had reached out to him to learn more about his research. He also said it wouldn’t surprise him if ChatGPT’s ability to predict stock moves decreased in the coming months as institutions started integrating this technology.

That’s because the experiment only looked at stock prices during the next trading day, while most people would expect the market could have already priced the news in seconds after it became public.

“As more and more people use these type of tools, the markets are going to become more efficient, so you would expect return predictability to decline,” Lopez-Lira said. “So my guess is, if I run this exercise, in the next five years, by the year five, there will be zero return predictability.”

Continue Reading

Technology

TSMC hits yet another record as profit surges 39%, beating estimates on AI chip demand surge

Published

on

By

TSMC hits yet another record as profit surges 39%, beating estimates on AI chip demand surge

The TSMC logo is displayed on a building in Hsinchu, Taiwan April 15, 2025.

Ann Wang | Reuters

Taiwan Semiconductor Manufacturing Company on Thursday reported a 39.1% increase in third-quarter profit from last year, hitting a fresh record as demand for artificial intelligence chips stayed strong.

Here are the company’s results versus LSEG SmartEstimates:

  • Revenue: NT$989.92 billion new Taiwan dollars, vs. NT$977.46 billion expected
  • Net income: NT$452.3 billion, vs. NT$417.69 billion 

TSMC’s revenue in the September quarter rose 30.3% from a year ago to NT$989.92 billion, beating estimates.

TSMC’s high-performance computing division, which encompasses artificial intelligence and 5G applications, drove third-quarter sales.

As Asia’s largest technology company by market capitalization, TSMC has benefited from the artificial-intelligence megatrend as it manufactures advanced AI processors for clients, including Nvidia and Apple.

TSMC said advanced chips, with sizes 7-nanometer or smaller, accounted for 74% of TSMC’s total wafer revenue in the quarter. 

In semiconductor technology, smaller nanometer sizes signify more compact transistor designs, which lead to greater processing power and efficiency. 

Continue Reading

Technology

FCC moves to expel one of Hong Kong’s largest telco companies from U.S. networks

Published

on

By

FCC moves to expel one of Hong Kong’s largest telco companies from U.S. networks

People walk past a PCCW sign in Hong Kong.

Mike Clarke | AFP | Getty Images

Regulators in the U.S. have moved to block one of Hong Kong’s largest telecommunications companies from accessing domestic networks, citing national security concerns.

The U.S. Federal Communications Commission announced on Wednesday that it had initiated proceedings to potentially bar HKT Trust and HKT Ltd and its subsidiaries from interconnecting with American networks, escalating concerns over its ties to China. 

The government agency asked HKT, which is a subsidiary of information and communication technology giant PCCW, to justify why its authorizations should not be revoked.  HKT’s current hold permits allowing direct exchange of calls and data with U.S. carriers.

China Unicom, which owns about 18.4% of PCCW, lost its own U.S. network access in 2022 due to similar concerns.

“The FCC’s action on HKT today is an appropriate step towards ensuring the safety and integrity of our communications networks,” FCC Chairman Brendan Carr said in a statement. 

“The FCC will continue to safeguard America’s networks against penetration from foreign adversaries, like China.

The Hong Kong-listed shares of HKT fell more than 5%, while PCCW fell 3.6% in Thursday trading.

Stock Chart IconStock chart icon

hide content

Share price of HKT and PCCW

According to their 2024 annual reports, HKT and PCCW derived about 13% of their 2024 revenues from regions outside greater China and Singapore, though specific countries weren’t detailed. HKT made up about 90% of the group’s total revenue.

Neither PCCW nor HKT immediately responded to CNBC’s requests for comment.

Under the leadership of Carr, the FCC has expanded efforts to expel Chinese state-linked entities, including China Telecom, Pacific Networks and ComNet, from U.S. markets.

On Friday, the FCC announced that the major U.S. online retail websites had removed millions of listings for banned Chinese electronics as part of its broader China crackdown.

Caught in U.S.-China trade tensions

PCCW is majority-owned by Hong Kong tycoon Richard Li, son of billionaire Li Ka-shing, who has increasingly found his businesses caught in the crossfire of the U.S.-China trade tensions.

FWD Group, owned by Li’s Pacific Century Group, recently faced hurdles expanding into mainland China amid backlash from regulators in China, Bloomberg reported in July.

In March, Beijing reportedly instructed state-owned firms to pause new deals with businesses linked to Li Ka-shing and his family after their conglomerate CK Hutchison agreed to transfer stakes in over 40 global ports — including two in Panama — to a BlackRock-led consortium. 

The ports deal stalled after Beijing objected to the exclusion of Chinese investors, with CK Hutchison indicating it no longer plans to comeplete the transaction in 2025.

The FCC’s latest move against HKT also comes as U.S. President Donald Trump escalates his trade war with China.

Continue Reading

Technology

Thirsty AI mega projects raise alarm in some of Europe’s driest regions

Published

on

By

Thirsty AI mega projects raise alarm in some of Europe’s driest regions

Liquid cooled servers in an installation at the Global Switch Docklands data centre campus in London, UK, on Monday, June 16, 2025.

Bloomberg | Bloomberg | Getty Images

Europe’s ambitious artificial intelligence strategy is at risk of colliding with an often overlooked but critically important environmental issue: water scarcity.

The European Union has big plans for data center expansion, announcing in April that it intends to at least triple its capacity over the next five to seven years as part of a push to become a world-class AI hub.

The rapid rollout of data centers, which power all aspects of the digital economy, from social media and online banking to AI tools like ChatGPT, has sparked some concern — particularly in regions already facing water scarcity.

The issue is especially acute across southern Europe, with around 30% of the population known to be situated in areas with permanent water stress. This refers to a situation where the demand for water exceeds the available supply during a specific period.

Data centers typically require large quantities of water to keep them from overheating.

Major tech companies like Amazon, Microsoft and Meta have invested billions of U.S. dollars in new facilities in Spain, for instance, while Google has plans to develop three hubs in Greece’s Attica region.

There's a water crisis looming. Big Tech and AI could make it worse

Kevin Grecksch, associate professor of water science, policy and management at the U.K.’s University of Oxford, told CNBC that plans to build data centers in water-stressed areas across Europe reflects a lack of integrated thinking from policymakers.

“AI is a buzzword and the talk of the town,” Grecksch said. “So, national and regional politicians try to get their hands on it, and it sounds as if you’re investing into the future, creating a few new jobs — but sustainability seems to be an afterthought.”

Grecksch said the rapid rollout of data centers across the region throws up plenty of unanswered questions, such as, given that in most jurisdictions public water supply has priority over everything else, what happens if data centers are shut down in a drought scenario? He conceded he had no answer to this prospect.

“Data centres tend to be built in arid or semi-arid climates because that’s the preferred environment for servers; yet those areas tend to be subject to water scarcity or drought prone as well,” Grecksch said.

A spokesperson for the European Commission, the EU’s executive arm, said policies of the European High Performance Computing Joint Initiative (EuroHPC JU) include selecting hosting sites for AI factories based on criteria that prioritizes energy efficiency and environmental sustainability.

“Green computing will continue to be pursued through energy-efficient supercomputers optimised for AI, using techniques such as dynamic power saving and re-use techniques like advanced cooling and recycling of the heat produced,” the spokesperson told CNBC by email.

The EU referenced the new “JUPITER” supercomputer in Jülich, Germany, as “a prime example of European excellence” in addressing energy efficiency, saying the system runs entirely on renewable energy and features “cutting-edge” cooling and energy reuse.

Data centers’ water footprint

In Aragon, an area of severe water stress in northeastern Spain, Amazon is planning to open three data centers. The proposal, which the U.S. tech giant says will create thousands of jobs, has sparked tension between local farmers and environmental activists.

In the U.K., the small English village of Culham has been picked as the first of the British government’s so-called AI “growth zones.” The designation of the Oxfordshire site, which is situated close to one of the country’s first new reservoirs in 30 years, has raised fears that it could put further pressure on local water supplies.

Nick Kraft, senior analyst at political risk consultancy Eurasia Group, said “extremely arid” and high-water stress localities were being targeted across Europe for further data center development.

“Complicating the matter is the fact that the most common understanding of data center water usage, and typically what companies report on when communicating with local stakeholders, is on-site water use — or the water used for cooling in data centers,” Kraft told CNBC by email.

This photograph taken on August 24, 2025 shows a general view of the Mediano reservoir, in the northeastern region of Aragon, Huesca province.

Ander Gillenea | Afp | Getty Images

“This despite more than half of data centers’ water footprint being off-site, occurring in energy generation and semiconductor manufacturing,” he added.

There are emerging signs that data center operators are maturing in their water stewardship, Kraft said, but assessing the full water footprint of these projects is expected to remain a major challenge.

Analysis published by S&P Global last month said the data center industry’s average exposure to water stress is projected to be high in the 2020s, with southern European countries such as Spain and Greece among the locations forecast to face the most water stress.

Data centers power the digital economy

Michael Winterson, secretary general of the European Data Centre Association (EUDCA), which represents the interests of the European data center operator community, said water consumption is a concern that the industry takes seriously.

“Water treatment and collection is now normal for us. And there are continual innovations in this space that reduce energy required, reduce water needed and are fast approaching near zero chemical treatment,” Winterson told CNBC.

An operator works at the data centre of French company OVHcloud in Roubaix, northern France on April 3, 2025.

Sameer Al-doumy | Afp | Getty Images

The EUDC’s secretary general also sought to highlight the importance of data centers to the region’s digital economy.

“This is trillions of dollars of GDP and millions of technology jobs in Europe alone — which in average pay significantly higher salaries than national average wages. A 20 [megawatt] Datacentre uses the similar amount of water as a golf course! How much GDP do golf courses create? What kind of jobs?” Winterson said.

A deepening water crisis

European lawmakers have previously warned about the region’s growing water crisis, saying there is a pressing need to tackle issues such as scarcity, food security and pollution at a time when Europe is the fastest-warming continent on the planet.

The European Environment Agency, for its part, said late last month that the region’s water resources are currently under “severe pressure,” with water stress affecting one-third of Europe’s population and territory.

When factoring in the explosion of AI demand, Laura Ramsamy, climate and hazard lead at data analytics platform Climate X, said the rollout of new hyperscale data centers in already water-stressed European areas “is really exacerbating the problem.”

Europe's power infrastructure needs to keep pace with AI boom

In the Netherlands in 2022, for instance, Meta paused its plans to build a large data center in the region of Zeewolde amid objections over environmental concerns, particularly high power and water consumption.

Ireland, which has long embraced the rollout of data centers to facilitate an AI boom, also recently came under scrutiny from environmental groups, with many of these hubs concentrated in the Dublin area.

Notably, the Netherlands and Ireland have both imposed effective bans on new data centers over the coming years due to concerns over grid capacity and their environmental impact.

A spokesperson of Ireland’s Department of Climate, Energy and the Environment said the construction and operation of data centers have “positively contributed” to the Irish economy over the past decade.

“As with all sectors of our economy, the operation and development of data centres are underpinned by Ireland’s legally binding climate objectives and the need to maintain robust energy security,” the spokesperson said.

“It is understood that the largest data centres in Ireland primarily operate air cooling rather than water cooling systems. This differentiates Ireland from many global data centre locations,” they added.

Continue Reading

Trending