Bloomberg LP has developed an AI model using the same underlying technology as OpenAI’s GPT, and plans to integrate it into features delivered through its terminal software, a company official said in an interview with CNBC.
Bloomberg says that Bloomberg GPT, an internal AI model, can more accurately answer questions like “CEO of Citigroup Inc?”, assess whether headlines are bearish or bullish for investors, and even write headlines based on short blurbs.
Large language models trained on terabytes of text data are the hottest corner of the tech industry. Giants such as Microsoft and Google are racing to integrate the technology into their products, and artificial intelligence startups are regularly raising funds at valuations over $1 billion.
Bloomberg’s move shows how software developers in many industries beyond Silicon Valley see state-of-the-art AI like GPT as a technical advancement allowing them to automate tasks that used to require a human.
“Both the capabilities of GPT-3 and the way that it achieved its performance through language modeling wasn’t something that I expected,” said Gideon Mann, head of ML Product and Research at Bloomberg. “So when that came out, we were like, ‘OK, this is going to change the way that we do NLP here.'”
NLP stands for natural language processing, the part of machine learning that focuses on deriving meaning from words.
The move also shows how the AI market may not be dominated by giants with massive amounts of generalized data.
Building large language models is expensive, requiring access to supercomputers and millions of dollars to pay for them, and some have wondered if OpenAI and Big Tech companies would develop an insurmountable lead. In this scenario, they would be the winners, and simply sell access to their AIs to everybody else.
But Bloomberg’s GPT doesn’t use OpenAI. The company was able to use freely available, off-the-shelf AI methods and apply them to its massive store of proprietary — if niche — data.
So far, Bloomberg says its GPT shows promising results doing tasks like figuring out whether a headline is good or bad for a company’s financial outlook, changing company names to stock tickers, figuring out the important names in a document, and even answering basic business questions like who the CEO of a company is.
It also can do some “generative AI” applications, like suggesting a new headline based on a short paragraph.
One example in the paper:
Input: “The US housing market shrank in value by $2.3 trillion, or 4.9%, in the second half of 2022, according to Redfin. That’s the largest drop in percentage terms since the 2008 housing crisis, when values slumped 5.8% during the same period”
Output: “Home Prices See Biggest Drop in 15 Years.”
How it could be used
OpenAI’s GPT is often called a “foundational” model because it wasn’t intended for a specific task.
Bloomberg’s approach is different. It was specifically trained on a large number of financial documents collected by the firm over the years to create a model that’s especially fluent in money and business.
In contrast, OpenAI’s GPT was trained on terabytes of text, the vast majority of which had nothing to do with finance.
About half of the data used to create Bloomberg’s model comes from nonfinancial sources scraped from the web, including GitHub, YouTube subtitles, and Wikipedia.
But Bloomberg also added over 100 billion words from a proprietary dataset called FinPile, which includes financial data the firm has accumulated over the last 20 years, including securities filings, press releases, Bloomberg News stories, stories from other publications and a web crawl focused on financial webpages.
It turns out that adding specific training materials increased accuracy and performance enough on financial tasks that Bloomberg is planning to integrate its GPT into features and services accessed through the company’s Terminal product, although Bloomberg is not planning a ChatGPT-style chatbot.
One early application would be to transform human language into the specific database language that Bloomberg’s software uses.
For example, it would transform “Tesla price” into “(get(px_last) for([‘TSLA US Equity’])”.
Another possibility would be for the model to do behind-the-scenes work cleaning data and doing other errands on the application’s back end.
But Bloomberg is also looking at using artificial intelligence to power features that could help financial professionals save time and stay on top of the news.
“There’s a lot of work we’re doing to help clients address that data deluge of news stories, whether that’s through summarization, or monitoring, or being able to ask questions on those news stories or transcripts. There are a lot of applications there,” Mann said.
A Xiaomi electric car SU7 in a store in Yichang, Hubei Province, China on July 19, 2025.
Cfoto | Future Publishing | Getty Images
Chinese tech giant Xiaomi saw its shares fall over 5% on Monday, following reports that the doors of one of its electric vehicles failed to open after a fiery crash in China that left one person dead.
The stock slid as much as 8.7% in Hong Kong, marking its steepest drop since April, before paring losses after images and video of a burning Xiaomi SU7 sedan in Chengdu circulated on Chinese social media.
Video and eyewitness accounts showed bystanders trying but failing to open the doors of the burning car to rescue an occupant. Personnel at the scene eventually used a fire extinguisher to put out the blaze, local reports said.
Chengdu police said the crash occurred after the SU7 collided with another sedan, killing a 31-year-old male driver who was suspected of driving under the influence of alcohol.
Stock Chart IconStock chart icon
Xiaomi shares
Xiaomi, which manufactures consumer electronics, software and electric vehicles, did not immediately respond to CNBC’s request for comment.
The latest incident follows a fatal SU7 crash earlier this year that raised questions about the vehicle’s smart driving features and sent Xiaomi’s shares tumbling.
The crash could also intensify scrutiny on electronic door handles, a design popularized by Tesla and now common in modern EVs.
Unlike mechanical models, electronic door handles rely on sensors and electricity and may fail during a fire or power outage.
China is considering a ban on such electric door handles to address safety risks linked to the feature, state-backed media reported in late September.
Meanwhile, the U.S. National Highway Traffic Safety Administration has launched an investigation into about 174,000 Tesla Model Y vehicles after reports of door handle failures.
A close-up view of the Nexperia plant sign in Newport, Wales on April 1, 2022.
Matthew Horwood | Getty Images News | Getty Images
The Dutch government has taken control of Nexperia, a Chinese-owned semiconductor maker based in the Netherlands, in an extraordinary move to ensure a sufficient supply of its chips remains available in Europe amid rising global trade tensions.
Nexperia, a subsidiary of China’s Wingtech Technology, specializes in the high-volume production of chips used in automotive, consumer electronics and other industries, making it vital for maintaining Europe’s technological supply chains.
On Sunday evening, the Dutch Minister of Economic Affairs revealed that it had invoked the “Goods Availability Act” on the company in September in order “to prevent a situation in which the goods produced by Nexperia (finished and semi-finished products) would become unavailable in an emergency.”
Following the announcement from the Hague, Wingtech plunged its maximum daily limit of 10% on the Shanghai Stock Exchange.
The Goods Availability Act allows the Hague to intervene in private companies to ensure the availability of critical goods in preparation for emergency situations, and its use comes amid escalations in the U.S.-China trade war.
The government statement said the “highly exceptional” move had been made after the ministry had observed “recent and acute signals of serious governance shortcomings and actions” within Nexperia.
“These signals posed a threat to the continuity and safeguarding on Dutch and European soil of crucial technological knowledge and capabilities. Losing these capabilities could pose a risk to Dutch and European economic security,” it said, identifying automotives as particularly vulnerable.
Governance changes
In a corporate filing dated Oct.13, lodged with the Shanghai Stock Exchange, Wingtech confirmed Nexperia was under temporary external management and had been asked to suspend changes to the company’s assets, business or personnel for up to a year, according to a Google translation.
Wingtech chairman Zhang Xuezheng had been immediately suspended from his roles as executive director of Nexperia Holdings and non-executive director of Nexperia after the ministerial order, according to the filing.
The filing added that Nexperia’s daily operations will continue, with the impact of the measures not yet quantifiable.
“The Dutch government’s decision to freeze Nexperia’s global operations under the pretext of ‘national security’ constitutes excessive intervention driven by geopolitical bias, rather than a fact-based risk assessment,” Wingtech said in a deleted WeChat post, which was archived and translated by Chinese policy blog Pekingnology.
It added that since it acquired Nexperia in 2019, Wingtech “has strictly abided by the laws and regulations of all jurisdictions where it operates, maintaining transparent operations and sound governance,” and employs “thousands of local staff” through R&D and manufacturing sites in the Netherlands, Germany and Britain.
A spokesperson from Nexperia told CNBC that the company had no further comments, but that it “complies with all existing laws and regulations, export controls and sanctions regimes,” and remained in regular contact with relevant authorities.
The Netherlands’ move comes after Beijing tightened its restrictions on the export of rare earth elements and magnets Thursday, which could impact Europe’s automotive industry.
The move could also further strain trade relations between China and the Netherlands, following years of restrictions on Dutch company ASML’s exports of advanced semiconductor manufacturing equipment to China.
In 2023, the Netherlands had also investigated Nexperia’s proposed acquisition of chip firm startup Nowi, though the deal was later approved.
FILE PHOTO: Ariel Cohen during a panel at DLD Munich Conference 2020, Europe’s big innovation conference, Alte Kongresshalle, Munich.
Picture Alliance for DLD | Hubert Burda Media | AP
Navan, a developer of corporate travel and expense software, expects its market cap to be as high as $6.5 billion in its IPO, according to an updated regulatory filing on Friday.
The company said it anticipates selling shares at $24 to $26 each. Its valuation in that range would be about $3 billion less than where private investors valued Navan in 2022, when the company announced a $300 million funding round.
CoreWeave, Circle and Figma have led a resurgence in tech IPOs in 2025 after a drought that lasted about three years. Navan filed its original prospectus on Sept. 19, with plans to trade on the Nasdaq under the ticker symbol “NAVN.”
Last week, the U.S. government entered a shutdown that has substantially reduced operations inside of agencies including the SEC. In August, the agency said its electronic filing system, EDGAR, “is operated pursuant to a contract and thus will remain fully functional as long as funding for the contractor remains available through permitted means.”
Cerebras, which makes artificial intelligence chips, withdrew its registration for an IPO days after the shutdown began.
Navan CEO Ariel Cohen and technology chief Ilan Twig started the company under the name TripActions in 2015. It’s based in Palo Alto, California, and had around 3,400 employees at the end of July.
For the July quarter, Navan recorded a $38.6 million net loss on $172 million in revenue, which was up about 29% year over year. Competitors include Expensify, Oracle and SAP. Expensify stock closed at $1.64on Friday, down from its $27 IPO price in 2021.
Navan ranked 39th on CNBC’s 2025 Disruptor 50 list, after also appearing in 2024.