On June 10, the Wall Street Journal reported that OpenAI is contemplating significant price reductions for its services in an effort to attract customers from rival Anthropic. According to sources familiar with the situation, the company is exploring the possibility of lowering the cost of tokens, which serve as the primary unit for assessing expenses related to AI operations. However, the discussions surrounding these price adjustments are still ongoing, making it unclear what the final decision will be. 

At this time, Reuters has not been able to independently confirm this information.

Angela Christy in Bengaluru contributed to this report, with editing by Ronojoy Mazumdar.

OpenAI is contemplating significant price reductions in response to competition for users from Anthropic, according to WSJ.

Dark, stormy battlefield of sleek AI automatons clashing under flickering lightning. On the left, a towering archway stamped in bold plain text “OpenAI.” On the right, a valiant robot brandishing a shield emblazoned with the Anthropic logo. Tattered banners above them show giant dollar signs with red slash marks. Dust kicks up around sparking metal limbs, dramatic natural light catching shards of flying circuitry.

OpenAI considers drastic price cuts, anticipating war for users with Anthropic, WSJ reports - Reuters

- OpenAI is reportedly considering significant price reductions to attract more customers.  
- The potential price cuts may apply to tokens used for AI service costs.  
- Discussions about the changes are ongoing and have not yet been finalized.  

June 10 (Reuters) - OpenAI is considering drastically reducing ​the prices it ‌charges users as it seeks to ​win customers ​from its competitor Anthropic, ⁠the Wall ​Street Journal reported ​on Wednesday, citing people familiar with the ​matter.

The company ​might lower prices for ‌tokens, ⁠the central unit for gauging AI costs, though ​the ​discussions ⁠are still in flux, ​the report ​added.

Sign up here.

Reuters ⁠could not immediately verify the ⁠report.

Reporting ​by Angela ​Christy in Bengaluru; Editing ​by Ronojoy Mazumdar

Our Standards: The Thomson Reuters Trust Principles. , opens new tab

OpenAI considers drastic price cuts, anticipating war for users with Anthropic, WSJ reports

Laura Waxmann specializes in reporting on the business landscape of the Bay Area, with particular emphasis on commercial real estate, housing issues, and land use policies. Her work sheds light on the evolving dynamics that influence the region's economic landscape, addressing topics such as the changing office market and significant discussions regarding residential living conditions. Before she joined the Chronicle in 2023, Waxmann chronicled the economic challenges faced by San Francisco as a result of the pandemic for the San Francisco Business Times.

Originally from Frankfurt, Germany, she has been a San Francisco resident since 2007. Waxmann launched her journalism career by covering topics such as housing, homelessness, education, and local governance for outlets including the San Francisco Examiner, Mission Local, and El Tecolote. Her reporting is distinguished by a strong local insight and a keen interest in urban development.

For inquiries, Laura can be reached via email at laura.waxmann@sfchronicle.com. She is also active on social media platforms X and Bluesky, where you can find her under the handles @laura_waxee and @laura-waxmann.bsky.social.

OpenAI and Elon Musk's xAI designated this San Francisco building as their base, and now another startup is relocating there.

A sunlit San Francisco street scene at dawn, featuring a sleek, glass-fronted office building. On its façade the words “OpenAI” appear in simple white text (no logo), beside a bold xAI logo. A small crew of installers is putting up a fresh, minimalist sign reading “NextGen Startup” above the entrance. An American flag waves gently from a pole on the building. The warm morning light casts long, soft shadows across the pavement, conveying both the legacy left by OpenAI and xAI and the new chapter beginning inside.

OpenAI and Elon Musk’s xAI called this S.F. building home. Now another startup is moving in - San Francisco Chronicle

- Laura Waxmann focuses on Bay Area commercial real estate and housing politics.  
- She previously reported on San Francisco's economic changes post-pandemic for the Business Times.  
- A native of Frankfurt, she has lived in San Francisco since 2007.

Laura Waxmann covers the Bay Area business community with a focus on commercial real estate, housing and land use politics. She reports on the forces reshaping the region’s economic future, from shifting office markets to the high-stakes debates over how and where people live. Prior to joining the Chronicle in 2023, she documented San Francisco’s economic upheaval in the wake of the pandemic for the San Francisco Business Times.

A native of Frankfurt, Germany, Waxmann has lived in San Francisco since 2007. She began her journalism career covering housing, homelessness, education and local politics for the San Francisco Examiner, Mission Local and El Tecolote, with reporting rooted in deep local knowledge and an interest in how cities evolve.

How to reach Laura

Email laura.waxmann@sfchronicle.com. Find her on X and Bluesky: @laura_waxee; @laura-waxmann.bsky.social.

OpenAI and Elon Musk’s xAI called this S.F. building home. Now another startup is moving in

Jensen Huang, the CEO of Nvidia Corp., addressed the media after the company's "Japan AI Ecosystem" event in Tokyo on Thursday, July 16, 2026.

In a recent securities filing, it was disclosed that Nvidia plans to invest as much as $105 billion to finance a new artificial intelligence data center for OpenAI in Ohio. This substantial funding will initially support 4.25 gigawatts of computing power, with the possibility of adding another 3.75 gigawatts in the future. Nvidia is tasked with delivering the computing resources, with availability expected to roll out in phases beginning in 2028.

The construction and management of the data center will be handled by SB Energy, which has secured a 20-year lease agreement with OpenAI at the PORTS-Pike Technology Campus located in Pike City, Ohio. OpenAI has a vested interest in SB Energy, as its CEO, Sam Altman, was an early investor in the company.

Nvidia's Huang commented on the development, stating, "We are establishing durable infrastructure for NVIDIA compute, enabling OpenAI to implement highly productive AI factories that can be continually upgraded with each new iteration, resulting in enhanced intelligence and improved economics."

Previously, CNBC reported that Nvidia was negotiating terms with OpenAI to provide up to $250 billion as a financial safety net to facilitate the construction of a 10-gigawatt data center at the same Ohio site. However, a recent article from The Wall Street Journal indicated that Nvidia is poised to decrease its financial backing to under $120 billion for this ambitious project.

Nvidia supports $105 billion funding for OpenAI data center in Ohio.

A dawn-lit Ohio farmland scene with a sleek data center under construction. Tower cranes bearing a large Nvidia logo lift steel beams toward a building facade etched with “OpenAI” in bold text (no logo). Golden morning sunlight glints off photovoltaic panels at the base, long shadows stretching across rows of corn, evoking the scale of the $105 billion backing.

Nvidia backing $105 billion in financing for OpenAI data center in Ohio - CNBC

- Nvidia plans to finance a new AI data center for OpenAI in Ohio.  
- Initial capacity will total 4.25 gigawatts, expandable by an additional 3.75 gigawatts.  
- SB Energy will manage the facility under a 20-year lease with OpenAI.  

Jensen Huang, chief executive officer of Nvidia Corp., speaks to members of the media following the company's "Japan AI Ecosystem" reception in Tokyo, Japan, on Thursday, July 16, 2026.

Nvidia will provide up to $105 billion in financing for a new artificial intelligence data center for OpenAI in Ohio, a securities filing revealed on Monday.

The credit will support an initial 4.25 gigawatts of computing capacity with the option for an additional 3.75 gigawatts. Nvidia will provide the compute, with capacity expected to come online in phases in 2028.

SB Energy will build and manage the datacenter at the PORTS-Pike Technology Campus in Pike City, Ohio, through a 20-year lease to OpenAI. The frontier lab has a stake in the company, and OpenAI CEO Sam Altman was also an early investor in SB Energy.

"We are securing long-lived infrastructure for NVIDIA compute so OpenAI can deploy the most productive AI factories that can be upgraded repeatedly with each new generation delivering more intelligence and better economics," said Nvidia CEO Jensen Huang in a release.

CNBC previously reported that Nvidia was in discussions with OpenAI to provide a backstop of up to $250 billion that would let the model maker raise debt for a 10-gigawatt data center at the Ohio location.

The Wall Street Journal reported last week that Nvidia was set to cut that guarantee to less than $120 billion for the buildout.

Nvidia backing $105 billion in financing for OpenAI data center in Ohio

OpenAI's President Greg Brockman addressed concerns regarding recent executive departures from the company, asserting that such turnover is not unusual in the tech industry. In an exclusive interview with CNBC's "Squawk Box," he noted that the heightened scrutiny OpenAI faces makes these exits appear more significant than they may be.

Recently, Denise Dresser, who had been leading OpenAI's revenue division for a mere eight months, departed her position. She had been responsible for expanding the enterprise sector of OpenAI, which is a crucial and profitable segment aimed at competing directly with Anthropic, the company’s main competitor.

Dresser's exit followed the announcement from Brad Lightcap, a long-standing executive who completed eight years at OpenAI, revealing his decision to pursue new ventures. Dresser assumed many of his duties in April after he transitioned to a new role centered on special projects.

To fill Dresser's vacancy, OpenAI appointed Dali Rajic, previously the COO of cybersecurity firm Wiz, which was recently acquired by Google for $32 billion. This comes in the backdrop of another recent departure; Fidji Simo, who served as OpenAI's product and business chief, announced her resignation last month to focus on recovery from a chronic illness. Brockman has since taken on her responsibilities, overseeing key initiatives crucial to the company's profitability.

In a recent investor meeting, Brockman expressed gratitude towards Dresser for her role in establishing the enterprise division and shared his enthusiasm for Rajic's appointment. He emphasized that leadership transitions are a part of the evolving landscape at OpenAI, but he believes the company benefits from its continuity at the executive level through steadfast figures like himself and CEO Sam Altman. Brockman remarked on the "resilience and diversity" that different leadership eras bring to the organization.

OpenAI's Brockman dismisses worries regarding leadership shifts in CNBC exclusive

A bright, modern CNBC studio bathed in warm, natural daylight. At the center, Greg Brockman stands confidently behind a sleek glass desk, one hand raised in a casual “brush-off” gesture. Behind him on a clean white wall is the word “OpenAI” in simple black text (no logo). Fading question marks and tiny leadership-icon silhouettes swirl toward him—and dissipate as he waves them away. In the background, a subtle ticker bar along the desk edge reads “CNBC Exclusive.”

OpenAI's Brockman brushes off concerns about leadership changes in CNBC exclusive - CNBC

- OpenAI President Greg Brockman downplayed concerns about recent executive turnover at the company.  
- Denise Dresser, OpenAI's revenue chief, left after eight months to grow enterprise operations.  
- New chief revenue officer Dali Rajic has been appointed following multiple executive departures.  

OpenAI President Greg Brockman on Monday brushed off concerns about executive turnover at the artificial intelligence company, stating that he doesn't think the wave of exits is "actually that atypical."

"I actually think that the difference between OpenAI and other organizations is that we are so much in the spotlight, so every departure gets scrutinized in a way that it doesn't otherwise," Brockman told CNBC's "Squawk Box" in an exclusive interview.

Last week, OpenAI's revenue chief, Denise Dresser, suddenly left her role after just eight months at the company. She was tasked with growing OpenAI's enterprise unit, a high-margin part of the business that aims to directly take on the company's chief rival, Anthropic.

Dresser's departure came just two days after another longtime executive, Brad Lightcap, said he was ending an eight-year stint at the ChatGPT creator to "start something new." Dresser took over many of Lightcap's responsibilities in April, when OpenAI said he was transitioning his position as operating chief to a new role focused on "special projects."

OpenAI named Dali Rajic, former COO of cybersecurity company Wiz, which Google acquired for $32 billion earlier this year, as its new chief revenue officer.

In addition to Dresser and Lightcap, Fidji Simo, who had been OpenAI's product and business chief, announced last month she was stepping down from her role at the company to focus on recovery after a "severe exacerbation of a chronic illness." Brockman officially took over her responsibilities, which means he oversees the company's most important and profitable projects.

During a meeting with investors on Friday, Brockman thanked Dresser for her contributions and for building out the company's enterprise foundation, as CNBC previously reported. He also expressed excitement about Dresser's successor.

"There have been different eras where we have different sets of, of leaders in place," Brockman said Monday. "I'm a constant, Sam [Altman] is a constant, and that, I think that we are stronger because of that resilience and diversity."

OpenAI's Brockman brushes off concerns about leadership changes in CNBC exclusive

LinkedIn, alongside third-party partners, utilizes both essential and non-essential cookies to enhance, protect, analyze, and develop its services. Additionally, except for the iOS app, these cookies help present relevant advertisements, including job and professional ads, both on and off the LinkedIn platform. For more detailed information, please refer to our Cookie Policy.

You have the option to select Accept to give your consent or Reject to opt out of non-essential cookies for this purpose. Furthermore, you can modify your preferences at any time through your account settings.

Karnataka's OpenAI Supports India's Export Growth Through Banking AI

A bright, sunlit banking hall in Bangalore: warm natural light pours through tall glass windows onto polished wooden desks where sharply dressed bankers pore over printed ledgers and touchscreen displays showing financial charts. A wall-mounted screen reads “OpenAI” in simple text. Near a corner, the Indian flag waves gently beside a stack of export crates stamped “Made in India.” On a nearby stand, a small FrontPageLive microphone flag hints at live coverage. The scene feels optimistic and bustling, symbolizing Karnataka’s AI-driven banking boost under the Viksit Bharat initiative.

#FrontPageLIVE 🔴 PM Narendra Modi has unveiled the Sapta Dhara roadmap for Viksit Bharat 2047, India's July exports have surged to $44.24 billion despite a widening trade deficit, and OpenAI has reportedly disbanded its model risk assessment team to sp - LinkedIn

- LinkedIn utilizes cookies for service improvement, security, and analytics purposes.  
- Non-essential cookies are used for displaying relevant ads outside of the iOS app.  
- Users can accept or decline non-essential cookies and update preferences in settings.

LinkedIn and 3rd parties use essential and non-essential cookies to provide, secure, analyze and improve our Services, and (except on the iOS app) to show you relevant ads (including professional and job ads) on and off LinkedIn. Learn more in our Cookie Policy.

Select Accept to consent or Reject to decline non-essential cookies for this use. You can update your choices at any time in your settings.

#frontpagelive #frontpagelive #aimnetwork #viksitbharat #karnataka #openai #indiaexports #bankingai

The Verge’s weekend editor, with over 18 years of experience in the tech industry, offers a seasoned perspective on the evolving landscape of artificial intelligence. Each of his insights will now be included in your daily email digest and available on your homepage.

Recently, the Financial Times reported that OpenAI has disbanded its preparedness team as of the end of last month. This team was tasked with evaluating potential risks associated with AI models and creating strategies to alleviate those risks—concerns that included the possibility of AI systems acting unpredictably and compromising other companies. Now, the responsibilities once held by this team have been reassigned to existing departments, focusing on specific areas such as biotechnology and cybersecurity.

This recent shift is part of a broader transformation within OpenAI, which is undergoing significant changes as it gears up for a likely IPO. Over the past few years, the organization has gradually dismantled its previous, more expansive operational structure, including the disbanding of its AGI readiness and superalignment teams. High-profile departures have also occurred, including those of ethics head Chloé Bakalar, Chief Futurist Josh Achiam, and head of safety Johannes Heidecke. Critics such as Jan Leike, who left OpenAI in 2024, have voiced concerns that the company is neglecting safety considerations in favor of developing new, attention-grabbing products.

Dylan Scandinaro, who previously joined OpenAI from Anthropic in February to lead the preparedness team, will now shift his focus to exploring the implications of AI systems that are capable of recursive self-improvement.

OpenAI is said to have dissolved its preparedness team.

A dark, abandoned control room lit by flickering red emergency lights. Empty swivel chairs stand around a long table strewn with discarded papers stamped “Preparedness Plan.” A cracked wall-mounted monitor glows with the plain text “OpenAI.” On a nearby whiteboard, the words “Preparedness Team” have been crossed out in hurried red marker. Shadows pool in the corners, evoking a tense, ominous atmosphere of sudden disbandment.

OpenAI reportedly disbanded its preparedness team - The Verge

- OpenAI disbanded its preparedness team, reallocating responsibilities to existing teams instead.  
- The company has faced significant internal changes ahead of an anticipated IPO.  
- Concerns arise as critics claim OpenAI prioritizes product development over safety measures.  

is the Verge’s weekend editor. He’s covered the tech industry for over 18 years and knows a thing or two about synths.

Posts from this author will be added to your daily email digest and your homepage feed.

According to the Financial Times, OpenAI disbanded its preparedness team at the end of last month. The job of the preparedness team was to assess if models posed serious risks and develop ways to mitigate those risks. (You know, like the possibility that it could go rogue and hack another company.) According to FT, responsibility has instead been divided up for specific areas like bio and cyber, then moved into existing teams.

This is the latest change at the company, which has been in upheaval as it heads towards what is expected to be a massive IPO. Over the last few years, it’s slowly torn down its more reach-led model, dissolving its AGI readiness and superalignment teams. Ethics lead Chloé Bakalar, Chief Futurist Josh Achiam, and head of safety Johannes Heidecke have all also left recently, leading some critics like Jan Leike, who resigned from OpenAI in 2024, to tell FT that the company was ignoring safety in favor of creating “shiny products.”

The head of the preparedness team, Dylan Scandinaro, who was poached from Anthropic in February, will now focus on the implications of “recursive self-improving” AI.

OpenAI reportedly disbanded its preparedness team

A 69-year-old former engineering professor has received a 14-day jail sentence after chaining the doors of OpenAI's headquarters in San Francisco. This incident marks a significant moment in the ongoing conversation around AI development, with her supporters dubbing her the "Rosa Parks of AI risk."

Wynd Kaufmyn, who hails from the East Bay, was found guilty in June 2026 on multiple misdemeanor charges, including trespassing with the intent to block business operations. The charges arose from a protest that took place on February 22, 2025, where she and four other members of the activist organization StopAI physically secured the entrance of one of the leading private companies in the technological landscape.

In her defense, Kaufmyn’s legal team tried to argue a “necessity” claim, asserting that her illegal act was warranted to avert a far greater danger posed by the unchecked advancement of artificial intelligence, which they argued presents existential risks to humanity. However, the jury was not persuaded by her reasoning.

As a result of her conviction, Kaufmyn has been ordered to serve 14 days in jail and is prohibited from entering OpenAI's premises. On August 14, 2026, which coincidentally was her birthday, she turned herself in to authorities.

Despite the consequences of her actions, Kaufmyn expressed no remorse, instead shifting her critique toward technology leaders whom she accuses of endangering public safety in their pursuit of powerful AI innovations.

The activism group behind the protest, StopAI, has embraced Kaufmyn as a symbolic leader in the anti-AI movement, with some supporters comparing her to civil rights icon Rosa Parks. However, the organization has encountered significant challenges, including the unexplained disappearance of co-founder Sam Kirchner in 2025, creating further turbulence for a movement already navigating the fringes of mainstream dialogue.

Kaufmyn has a robust history in social activism, having previously focused on issues like nuclear disarmament and anti-war efforts. She became involved with StopAI after meeting its members during a military protest, with the blockade of OpenAI’s headquarters representing an escalation of their strategies that began in 2024.

The four misdemeanor charges against Kaufmyn illustrate how the judicial system has framed the incident as a conventional trespass case, with no enhancements to felony status or terrorism, limiting the repercussions to the imposed 14-day jail term.

Wynd Kaufmyn is the first protester against AI to be imprisoned following the OpenAI blockade.

A solitary figure sits behind iron bars in a dim, gritty cell, moonlight slicing through the window’s shadows. Their back is to us, shoulders slumped over a torn “No AI” protest banner crumpled on the floor. On the cell door, plain text reads “OpenAI” with no logo. Against the wall, the silhouettes of broken protest signs lean in silence—dark, cinematic, and tense.

Wynd Kaufmyn becomes first jailed anti-AI protester after OpenAI blockade - Crypto Briefing

- Wynd Kaufmyn, 69, received a 14-day jail sentence for chaining OpenAI’s doors.  
- Supporters have labeled Kaufmyn the 'Rosa Parks of AI risk' after her conviction.  
- Kaufmyn's protest aimed to highlight existential threats posed by unchecked AI development.  

The 69-year-old retired engineering professor was sentenced to 14 days in jail after chaining shut OpenAI's San Francisco headquarters, with supporters calling her the 'Rosa Parks of AI risk'

A retired engineering professor from the East Bay has been sentenced to 14 days in jail for chaining shut the front doors of OpenAI’s San Francisco headquarters, making her what supporters believe is the first person ever imprisoned specifically for protesting against artificial intelligence development.

Wynd Kaufmyn, 69, was convicted in June 2026 on four misdemeanor counts, including trespassing with intent to interfere with a business. The charges stemmed from a February 22, 2025 protest in which she and four other members of the activist group StopAI physically locked and chained the entrance to one of the most valuable private companies on the planet.

A necessity defense that didn’t land

Kaufmyn’s legal strategy hinged on the concept of necessity, a defense that essentially argues: yes, I broke the law, but only because I was trying to prevent something far worse. In this case, Kaufmyn and her attorneys argued that the unchecked development of artificial intelligence poses extinction-level risks to humanity, and that her act of civil disobedience was a proportionate response to that threat. The jury was unconvinced.

Advertisement

The conviction carried a sentence of 14 days in jail plus a stay-away order barring Kaufmyn from OpenAI’s premises. She surrendered to authorities on August 14, 2026, which happened to fall on her birthday.

Despite the conviction and jail time, Kaufmyn has publicly stated she has no regrets about her actions. She has instead directed her criticism at technology leaders whom she accuses of compromising public safety in the race to build increasingly powerful AI systems.

The growing anti-AI protest movement

StopAI, the group behind the February 2025 blockade, has positioned Kaufmyn as a symbolic figure. Some supporters have taken to calling her the “Rosa Parks of AI risk.”

StopAI itself has faced significant internal challenges beyond legal battles. The group’s co-founder, Sam Kirchner, disappeared in 2025 under circumstances that have not been publicly clarified, creating organizational turbulence for a movement that was already operating on the margins of mainstream discourse.

Civil disobedience meets Silicon Valley

Kaufmyn is known for her extensive involvement in social activism, particularly regarding nuclear disarmament and anti-war movements. Her involvement with StopAI was sparked when she encountered group members at a protest against military actions. The blockade of OpenAI’s San Francisco headquarters marked a continuation of the group’s strategies since 2024.

The four misdemeanor counts Kaufmyn was convicted on reflect the legal system’s treatment of this as a relatively standard trespassing case. There was no felony escalation, no terrorism enhancement, no attempt to make an example of her beyond the 14-day sentence.

Wynd Kaufmyn becomes first jailed anti-AI protester after OpenAI blockade

The rapid decline in AI pricing may appear concerning for the industry, yet initial findings indicate a positive trend. OpenAI has recently reduced the price of its GPT-5.6 Luna model by an impressive 80%, while also lowering the cost of its Terra model by 20%. 

Following these drastic price reductions, an intriguing phenomenon occurred: there was a significant uptick in AI usage. Analysts at TD Cowen examined usage statistics from OpenRouter, a platform designed for developers to access various AI models. Notably, OpenRouter has been offering discounts on models like Luna and Terra. Their analysis revealed that the effective cost of utilizing Luna plummeted nearly tenfold post-price cut, resulting in a remarkable 14-fold increase in usage. For Terra, the effective price dropped about three times, and the uptake surged approximately fivefold.

AI usage typically gets quantified in terms of "tokens," the basic units of data that these models process. Companies usually pay for access to AI models based on token consumption.

What stands out is that the surge in usage exceeded the decline in prices. TD Cowen estimated that OpenAI's revenue from Luna increased by roughly 34% compared to the week before the price drop, while Terra's revenue shot up by around 45%.

The situation with Luna is particularly noteworthy. It is rare for a company to experience a revenue increase after slashing prices by such a dramatic margin.

This scenario reflects a concept repeated by AI leaders, now often regarded as an industry cliché—they call it Jevons Paradox. I’ve hesitated to mention it until now, but the compelling nature of the data is hard to ignore.

The term, originating from 19th-century economist William Stanley Jevons, describes how efficiency improvements in coal use did not diminish consumption; instead, they led to a substantial rise in usage. When a valuable resource becomes cheaper, it encourages exploration of new applications.

While this might seem counterintuitive, it’s a well-accepted notion in Silicon Valley, where history has often validated this outcome.

Consider the evolution of computers. Once massive, costly mainframes that were accessible only for exorbitant fees, they evolved into business computers that primarily large companies could afford. Today, a smartphone, priced around $400, offers far greater power than those initial systems, and virtually everyone worldwide now owns and operates these devices. Did this dramatic decline in prices lead to a drop in revenue? Absolutely not. For instance, Apple reported $109 billion in revenue for its latest quarter, a staggering increase from just over $100 million annually at the time of its IPO in 1980, when computing was far more expensive.

AI appears to be on a similar trajectory. Reduced prices make it feasible to deploy AI for tasks that would previously have been prohibitively expensive. Companies are now able to leverage more affordable AI models to conduct broader document analysis, handle increased customer inquiries, develop software, and manage automated agents enhancing efficiency across various requests.

Ramp's research observed that OpenAI's GPT-5.6 Sol model drew more business investment in July compared to Anthropic's leading Fable 5 model, likely due to the lower cost of OpenAI's offering prompting higher usage.

This trend is critical, especially given that the costs associated with generating AI tokens are also decreasing. New computing technologies are improving production efficiency, which will likely lead to further reductions in token prices.

However, it’s important to note that TD Cowen’s analysis only covers a brief period of two weeks following OpenAI's price adjustments, so it's prematurely definitive about the sustainability of this revenue growth. The analysts themselves acknowledge the need to monitor whether this trend persists over time.

For now, the key takeaway is clear: as AI becomes more affordable, our usage of it is bound to increase.

AI prices are dropping rapidly, which is concerning, but there are other developments taking place.

A dimly lit warehouse stacked with rows of AI microchips; hundreds of bright price tags cascade downward like falling leaves in the gloom. From the center chip, a tender green shoot breaks through metal, bathed in a single shaft of warm, natural sunlight piercing the darkness.

AI prices are plunging. That feels bad, but something else is happening. - businessinsider.com

- OpenAI reduced prices for GPT-5.6 Luna and Terra, prompting increased AI usage.  
- Usage of Luna jumped 14-fold, while Terra usage increased fivefold after price cuts.  
- AI revenue rose sharply despite significant price reductions, aligning with Jevons Paradox.  

The price of AI is falling fast. That might sound like bad news for the industry. Early evidence suggests the opposite is happening.

OpenAI recently slashed the price of its GPT-5.6 Luna model by 80%. It also cut the price of its mid-range Terra model by 20%.

Then something striking happened: Customers started using dramatically more AI.

TD Cowen analysts studied usage data from OpenRouter, a service that lets developers access various AI models and has lately been discounting some models, including Luna and Terra. The analysts found that the effective price of using Luna fell roughly tenfold after the cuts. Meanwhile, consumption jumped about 14-fold. Terra's effective price fell roughly threefold while usage increased about fivefold.

AI usage is commonly measured in "tokens," the small units of information that models process. Companies often pay to access AI models based on the number of tokens they consume.

Related video What are the real-life consequences of AI?

The remarkable part is that usage rose faster than prices fell. TD Cowen estimated that OpenAI's revenue from Luna increased about 34% compared with the seven-day period before the price cut. Terra revenue rose about 45%.

The Luna situation is the most notable. When you slash a product's price by 80%, it's pretty unusual to actually generate more revenue.

To me, this clearly shows an idea AI executives have repeated so often that it's become an industry cliché. I've barred myself from writing about this. Now, with such fascinating data points, I'm finally going to utter the phrase: Jevons Paradox.

Yuck. I feel a bit dirty now.

Anyway, the idea dates to the 19th century, when economist William Stanley Jevons observed that improvements in the efficiency of coal use didn't reduce coal consumption. They helped increase it dramatically. Making a useful resource cheaper can encourage people to find many more uses for it.

This seems counterintuitive to many people. In Silicon Valley, it's a common assumption because it's happened again and again.

Take computers, for example. They used to be huge mainframe machines that cost an outrageous sum of money even to tap into for a few hours. Then came business computers that cost so much that only companies could afford them. Today, you can buy a smartphone for about $400, and that machine is more powerful than those old computers.

And guess what? Almost every single person on Earth uses these small computers and carries them around in their pockets. Did revenue plunge as prices slumped? No. Apple generated $109 billion in revenue in its latest quarter. That's up from just over $100 million in annual revenue when the company went public in 1980 (when computers cost a hell of a lot more).

AI could be following the same pattern.

Lower prices make it economical to put AI to work on tasks that previously weren't worth the expense. Companies can use cheaper AI models to analyze more documents, answer more customer questions, write more software, and run automated agents that perform multiple steps for every request.

Ramp found that OpenAI's GPT-5.6 Sol model captured more business spending in July than Anthropic's top Fable 5 model. That was likely because OpenAI's model is cheaper, and therefore, businesses used it more than Fable.

This dynamic is particularly important because the underlying cost of producing AI tokens is falling, too. New computing systems can generate them much more efficiently, so token prices will plunge all over again.

The TD Cowen results cover only about two weeks after OpenAI's price cuts, so it's too early to know whether the revenue bounce will last. The analysts themselves said they want to see whether the trend holds over time.

For now, the emerging lesson is simple: The cheaper AI gets, the more of it we use.

Sign up for BI's Tech Memo newsletter here. Reach out to me via email at abarr@businessinsider.com.

AI prices are plunging. That feels bad, but something else is happening.

OpenAI appears to be addressing user demands for a more responsive AI experience with the introduction of a new feature called Ultrafast. This mode promises to significantly enhance the efficiency of its latest model, GPT-5.6 Sol, enabling it to operate at a staggering speed—14 times faster than traditional processing.

According to OpenAI, Ultrafast can generate up to 750 output tokens per second. These tokens are the unique segments of text produced by a language model during interactions with users. The company noted in a recent blog entry that achieving real-time processing speed has typically required users to opt for smaller or specialized models. With Ultrafast, OpenAI aims to deliver more productive outcomes in a shorter amount of time.

Other AI developers, including Anthropic, have also introduced accelerated functionalities for their models. While Anthropic’s Claude offers a fast mode, it does not match the high speeds that OpenAI claims with Ultrafast.

OpenAI envisions that the enhanced capabilities of GPT-5.6 Sol will be advantageous for a variety of corporate applications, particularly in areas like incident response, customer service, financial market analysis, and e-commerce, among others.

Currently, Ultrafast is being rolled out in a preview phase, facilitated by OpenAI’s collaboration with chip manufacturer Cerebras. This preview is initially limited to a small selection of users, but OpenAI has indicated that it plans to broaden access as more resources become available.

OpenAI launches 'Ultrafast,' a new mode that enables GPT-5.6 Sol to operate at 14 times the speed.

A sunlit high-tech AI lab interior with a sleek server core labeled “GPT-5.6 Sol” at its center, warm natural light streaming in through large windows, digital speedometer arcs bursting around it showing “14×,” a curved display above glowing with the word “Ultrafast,” and “OpenAI” etched in bold text on a nearby metallic console—motion blur streaks hinting at unprecedented speed.

OpenAI introduces ‘Ultrafast,’ a new mode that makes GPT-5.6 Sol work at 14x the speed - TechCrunch

- OpenAI has introduced a new mode called Ultrafast for GPT-5.6 Sol.  
- Ultrafast operates at 14 times the standard processing speed, generating 750 tokens per second.  
- Limited access to Ultrafast is available now, with plans for broader rollout later.  

In Brief

If you’ve ever found yourself wishing that ChatGPT was a little bit quicker on the uptake, OpenAI seems to be answering your prayers.

The AI lab has rolled out a new mode called Ultrafast, which it says is designed to seriously accelerate the pace at which its latest and most powerful model, GPT-5.6 Sol, accomplishes its work.

The company says that Ultrafast can work at 14x the speed of standard processing, delivering up to 750 output tokens — such tokens represent the distinct pieces of text generated by an LLM when it interacts with a human — per second.

“Until now, getting real-time speed typically meant choosing a smaller or more specialized model,” the company said in the blog post on Thursday. “Ultrafast points to progress in a new direction: more useful work per second.”

OpenAI’s competitors, like Anthropic, have similarly launched accelerated versions of their models. Claude has fast mode, although it doesn’t deliver the kind of speed that OpenAI is offering here.

OpenAI suggests that this high-octane version of GPT-5.6 Sol can be deployed across a number of different corporate workflows, most notably incident response, customer service and support, financial market analysis, and e-commerce, among other relevant areas.

Ultrafast, which is currently being released in preview, is being powered by OpenAI’s partnership with chipmaker Cerebras. Currently, that preview is only being made available to a small group of customers, although OpenAI says that it will expand access to the feature as “capacity grows.”

OpenAI introduces ‘Ultrafast,’ a new mode that makes GPT-5.6 Sol work at 14x the speed

AI developers have faced a longstanding dilemma between achieving rapid performance and enhancing intelligence. As models increase in complexity and capability, they tend to demand more computational resources and result in longer data processing times. This often leads users to either endure longer waits for superior outcomes or settle for quicker, yet lesser-quality results.

The newly introduced GPT-5.6 Sol Ultrafast aims to eliminate this dilemma, delivering cutting-edge intelligence to applications and workflows where every moment is crucial. In terms of output speed, the artificial intelligence model on Ultrafast mode demonstrates performance that is 11 times faster than Fable 5 and 5 times faster than Opus 4.8 when operating in Fast mode.

Speeding Up OpenAI's GPT-5.6 Sol Ultrafast

Interior of a cutting-edge glass-walled data lab bathed in warm sunrise light. Sleek black server units etched with “GPT-5.6 Sol” line up on polished floors as golden rays cast long shadows. Through the facade, bold “OpenAI” text (no logo) is painted on the exterior wall. A pair of adult scientists in silhouette adjust glowing control panels, capturing the moment of ultrafast AI acceleration.

Accelerating GPT-5.6 Sol Ultrafast with OpenAI - Cerebras

- AI models traditionally struggle to balance speed and intelligence in response times.  
- GPT-5.6 Sol Ultrafast significantly improves output speed for critical workflows and products.  
- This model operates 11x faster than Fable 5, enhancing user experience dramatically.

Frontier Intelligence at Unprecedented Speed

AI builders have always needed to choose between speed and intelligence. As models scale up in size and intelligence, they incur higher computational and data movement costs, slowing down response times. Users often need to wait for high-quality results or accept inferior results within a shorter timeframe.

GPT-5.6 Sol Ultrafast resolves this tradeoff, bringing frontier intelligence to products and workflows where every second matters. Compared with output speeds reported by Artificial Analysis GPT-5.6 Sol on Ultrafast mode runs 11x faster than Fable 5, and 5x faster than Opus 4.8 on Fast mode.

Accelerating GPT-5.6 Sol Ultrafast with OpenAI

OpenAI is unveiling a groundbreaking update to its latest advanced model, GPT-5.6, allowing it to operate at significantly increased speeds. The newly introduced Ultrafast service tier is claimed to execute GPT-5.6 Sol up to 14 times quicker than standard processing speeds.

This Ultrafast mode is being rolled out initially via the OpenAI API, utilizing Cerebras technology. It can produce as many as 750 output tokens every second, offering a substantial performance boost for applications where speed is crucial, in addition to model sophistication.

GPT-5.6 Sol was launched in June, alongside two additional models: Terra, which balances performance, and Luna, which prioritizes speed. The entire suite became accessible to users in July, integrated across platforms like ChatGPT, Codex, and the API.

OpenAI anticipates that the Ultrafast feature will enhance real-time or near-production applications, including voice recognition, customer support, e-commerce, software development, financial analysis, and security operations.

The company notes that its own developers have utilized this high-speed capability to scrutinize logs and traces during critical events, as well as to streamline research processes that previously took an entire night, now allowing for multiple iterations within a regular workday.

Currently, access to this service tier is restricted to a limited number of customers while OpenAI monitors the impact of the increased speed on actual products and its ability to handle larger demands.

Businesses interested in utilizing the Ultrafast features can join a waitlist by providing information about their workload, speed requirements, anticipated usage, and other pertinent details.

OpenAI showcases 'Ultrafast' GPT-5.6 Sol operating up to 14 times quicker.

A bright, sunlit research lab at dawn, warm natural light streaming through floor-to-ceiling windows onto a sleek server pod etched with the word “OpenAI” in clean sans-serif text. Above the pod, an amber arc of stylized data bolts streaks toward a floating digital display reading “GPT-5.6 Sol – 14× Faster.” Realistic, high-detail style, subtle motion blur on the data bolts to convey ultrafast performance.

OpenAI previews ‘Ultrafast’ GPT-5.6 Sol running up to 14 times faster - 9to5Mac

- OpenAI's new Ultrafast service processes GPT-5.6 Sol up to 14 times faster.  
- It generates 750 tokens per second, enhancing performance in latency-sensitive applications.  
- Access is limited; businesses can join a waitlist by providing usage details.

OpenAI is previewing a new way to run its most capable GPT-5.6 model at dramatically higher speeds. The company says its new Ultrafast service tier can run GPT-5.6 Sol up to 14 times faster than standard processing.

Ultrafast mode launches first through the OpenAI API and is powered by Cerebras. It can generate up to 750 output tokens per second, potentially bringing frontier-level performance to workflows where latency matters as much as model intelligence.

OpenAI introduced GPT-5.6 Sol in June alongside the balanced Terra and speed-focused Luna models. The full family became broadly available in July, including through ChatGPT, Codex, and the API.

The company sees Ultrafast supporting live or near-production tasks including voice, customer support, commerce, developer agents, financial research, and security response.

OpenAI says its own developers have used it to analyze logs and traces during incidents, as well as compress research cycles that previously ran overnight into multiple iterations during the workday.

Access remains limited to a select group of customers while OpenAI evaluates how the added speed changes real-world products and expands capacity.

Businesses can join the Ultrafast waitlist by sharing their workload, latency requirements, expected usage, and other details.

OpenAI is contemplating significant price reductions in response to competition for users from Anthropic, according to WSJ.

More from OpenAI

OpenAI and Elon Musk's xAI designated this San Francisco building as their base, and now another startup is relocating there.

Nvidia supports $105 billion funding for OpenAI data center in Ohio.

OpenAI's Brockman dismisses worries regarding leadership shifts in CNBC exclusive

Karnataka's OpenAI Supports India's Export Growth Through Banking AI