OpenAI has unveiled a new device known as the Codex Micro at the AI Engineer World Fair. According to spokesperson Dominik Kundel, this innovative keyboard aims to significantly enhance users' experience with Codex. It is designed to optimize and elevate productivity, allowing developers and coders to make the most of OpenAI's groundbreaking technology.

Check out this image of OpenAI's Codex hardware.

A bright, minimalist AI research lab bathed in warm natural light streaming through a skylight. In the center, a sleek black-and-silver server unit labeled “Codex” with “OpenAI” text etched cleanly on its side (no logo). A lone technician in a crisp white lab coat stands beside it, adjusting a glowing circuit board inside the open chassis. In the background, neatly organized server racks hint at vast computing power.

Here’s a photo of OpenAI’s Codex hardware. - The Verge

- OpenAI introduces the Codex Micro at the AI Engineer World Fair.  
- Codex Micro is designed to enhance users' interaction with Codex technology.  
- OpenAI's spokesperson emphasizes the gadget's potential to supercharge Codex usage.  

Here’s a photo of OpenAI’s Codex hardware.

The gadget is called the Codex Micro, and it’s being shown at the AI Engineer World Fair, where it’s being described as a keyboard that’s “designed to supercharge people’s Codex usage,” according to OpenAI spokesperson Dominik Kundel.

Here’s a photo of OpenAI’s Codex hardware.

Investors looking to engage with future IPOs from tech giants like OpenAI and Anthropic are facing a unique roadblock: there’s no direct method for the everyday person to buy shares ahead of these launches. In response, some individuals are turning to an unconventional market, purchasing used apparel—such as t-shirts and hoodies—that were originally given to employees or attendees at events, now fetching impressive resale prices.

On platforms like eBay, there's a noticeable increase in demand for merchandise branded by these AI companies. For example, a collection of OpenAI Codex items, which includes a hat and a hoodie, is being listed for as much as $475, while a pre-owned OpenAI sweatshirt recently sold for $246.99. Similarly, an Anthropic pen reached a selling price of $39.99, and an employee t-shirt from OpenAI garnered $120.

Mitchell Brody, who operates an eBay t-shirt business with his daughter, comments, “I believe these will become the vintage concert shirts of the future.” They currently feature an employee-issued OpenAI tee priced at $249.99, with Brody envisioning a future where people proudly share stories about their unique OpenAI merchandise from this era.

Conversations with various sellers reveal a striking trend: items from OpenAI and Anthropic are significantly more valuable compared to merchandise from other tech firms. For instance, an employee tank top from Google is available for only $24.95, and even a SpaceX polo shirt is listed for a mere $11.92. 

This trend resembles another recent occurrence when Silicon Valley Bank’s merchandise spiked in value amidst its financial turmoil; however, interest rapidly declined once the situation stabilized. One seller noted that items tied to companies experiencing dramatic highs and lows, like SVB, often create a temporary bubble that eventually bursts.

While a t-shirt or pen won't replace the prestige of company stock or influential titles on a resume, they fulfill a social desire to belong to the unique culture of Silicon Valley. Many buyers hail from regions outside major tech hubs, such as Ohio and upstate New York, where ownership of these exclusive items provides a sense of connection to the tech world.

The real draw for these garments isn’t merely their branding but their association with limited production runs tied to significant product launches or achievements within the companies. To validate authenticity, buyers pay close attention to manufacturing tags and original packaging. Many items, often sourced from thrift stores or gifted by industry friends, carry a narrative of insider knowledge.

“The most sought-after shirts are those issued to employees—they carry an ‘insider’ appeal,” Brody points out, highlighting that shirts tied to specific SpaceX missions, for instance, command higher prices. 

Capitalizing on the growing buzz around its merchandise, OpenAI recently launched OpenAI Supply Co., which was initially intended for staff but is now open to the public. This online store offers custom garments and highlights limited-edition items, akin to exclusive sneaker releases, further solidifying the connection between their brand identity and consumer culture.

Historically, interest in Silicon Valley memorabilia has surged during times of company turmoil. From the early days of the dot-com bubble to post-IPO phenomenon items from firms like Theranos and WeWork, many notable tech artifacts have achieved significant resale values. For example, items associated with Apple and its founder Steve Jobs have auctioned for millions—an indication that nostalgia and historical significance have made tech memorabilia a lucrative market.

Bobby Livingston, executive vice president at RR Auction, describes collectibles from notable technological breakthroughs as "high-end participation trophies," valuing them as tangible connections to pivotal moments in innovation.

Though not every tech company’s merchandise carries equal weight—failed ventures may go unnoticed in history—items from successful firms like Apple remain revered. According to Livingston, it may take decades before AI-related items reach their own recognition in the world of collectibles. Yet, he notes early signs of market enthusiasm, such as signed graphics cards from Nvidia fetching substantial prices.

For those hoping to profit quickly, specialized items, particularly those from Anthropic, tend to be scarce. Buyers like Alejandro Navia have even tailored their plans just to acquire unique pieces, such as the sought-after “thinking” baseball hat from Anthropic, which typically sells for $50 to $100 on resell platforms. Although Navia initially considered selling his hat for profit, he ultimately opted to keep it, valuing the experience and pride it represents within the tech community.

The Thriving Resale Market for Used OpenAI and Anthropic Merchandise

An outdoor flea market stall in warm morning sunlight: a weathered wooden booth displays gently worn OpenAI hoodies, mugs, and stickers under a hand-painted “OpenAI” sign (text only), while a large Anthropic logo banner hangs behind. Handwritten “used” price tags dangle from each item as a vendor hands a collector a scuffed OpenAI notebook amid rows of secondhand swag.

The Booming Secondary Market For Used OpenAI and Anthropic Swag - Forbes

- AI company swag, including t-shirts and hoodies, is reselling for hundreds of dollars.  
- OpenAI's merchandise is viewed as a future collectible, akin to vintage concert shirts.  
- Speculative swag trends in Silicon Valley mirror past tech memorabilia resale phenomena.  

N ormal people can’t buy shares of OpenAI or Anthropic ahead of the companies’ looming IPOs. So some are buying into a different kind of secondary transaction: used t-shirts and hoodies that were doled out to employees or conference-goers as a perk, that are now reselling for hundreds of dollars.

On eBay, there’s a bubble in AI company swag. Recently, boxes of OpenAI Codex swag consisting of a hat, hoodie and a few small tchotchkes, were commanding as much as $475. A used OpenAI sweatshirt, $246.99. An Anthropic-branded pen sold for $39.99; An authentic employee t-shirt from the $956 billion AI company, fetched $120.

“I do think these are the vintage concert rock shirts of tomorrow,” says Mitchell Brody, who operates a t-shirt shop on eBay with his college-aged daughter, Sunny. They’re currently selling an employee-issued OpenAI t-shirt for $249.99. “It’s geek chic. There’s going to be a day when people brag about the OpenAI double stitch shirt they got in the roaring twenties.”

Forbes spoke with sellers who have recently sold AI company swag on eBay, who confirmed that it, particularly OpenAI and Anthropic items, is going for many multiples above other types of company swag. On Brody’s shop, for instance, an employee tank top from tech stalwart Google is listed for $24.95. Even SpaceX lacks the new AI swag hotness premium—a legit company polo is priced at $11.92.

It’s another bubble inside the bubble.

“There’s going to be a day when people brag about the OpenAI double stitch shirt they got in the roaring twenties.” EBay seller Mitchell Brody

“The only time I've seen it like that is when Silicon Valley Bank collapsed, and their swag was selling at big prices,” said one Bay Area-based seller, who declined to be identified because his employer is not aware of his side hustle. But when the bank survived, its swag bubble deflated. Recently an appropriately fatuous SVB wine glass and cheese board set sold for less than $23.

A used OpenAI sweatshirt that sold for nearly $250. The seller found it at a thrift store for less than $15. courtesy eBay seller Travis Deuel

A shirt or pen is a poor substitute for equity or an AI big on your resume. But perhaps there’s something in the illusion of belonging in the machine washable theater of Silicon Valley clout.

“You’ll look cool if you wear the swag, particularly if you don’t live in San Francisco where it’s commonplace to know people who work at these companies,” the seller says. Many buyers are from outside major tech hubs.

“Ohio is a place I hear from a lot,” he says. “Same with upstate New York.”

The real value, sellers say, isn't just the logo: it is tied to hyper-limited, small-batch runs pegged to product launches and company milestones. The garments are Silicon Valley street cred, proof that you attended a special event — or bought a shirt from someone who did.

Because anyone can print a corporate logo onto a generic blank t-shirt, buyers meticulously inspect garments for manufacturing tags and original conference packaging to verify their authenticity, sellers told Forbes. Many are found in thrift shops. Some are simply gifted to them by friends in the industry.

“The shirts that are worth the most are employee-issued. It’s insider-ish,” Brody said, who said that SpaceX shirts specific to a single flight number command massive premiums. “If you can get something others can’t buy on a public website, those are the real bragging rights.”





courtesy Mitchell Brody

OpenAI is opportunistically leaning into the hype to create a revenue-generating business line. OpenAI Supply Co, which was formerly for internal employees, quietly launched to the public on Wednesday. The site peddles custom garments, such as a $40 hat and $15 socks, and on an archive page, lists old swag with “decommissioned” dates, similar to how Nike would orchestrate a limited release SNKR scarcity event.

The store is a monument to its own vainglory, peddling “commemorative” swagriculture.

“Supply Co. began as a small line of merch for employees and almost immediately blew up. OpenAI-ers developed concerningly strong, excited opinions about collectible cards, blue folding chairs, and graphic hoodies as emerging currency,” the store’s About page explains, noting that OpenAI merch isn’t just corporate logo loot, but rather a “material embodiment of company culture” and “physical expressions of research energy.”

This isn’t the first speculative swag frenzy Silicon Valley has experienced. Collectors have always paid top dollar for mementos of spectacular startup crashes, where it’s from early dot-com busts Pets.com and Webvan or the more recent wreckage of Theranos and WeWork. And Apple has always been on the forefront of tech hypebeast culture: in the early 2000s, enthusiasts would wait in line at Apple store openings for limited edition free tshirts, which they would then flip on eBay for $200 or even $500.

But it’s only in the past decade that relics from past tech success stories have turned into a multi-million-dollar asset class — in particular, vintage Apple and Steve Jobs mementos. RR Auction, a Boston-based auction house that has overseen some of the largest tech memorabilia sales in history, has sold an original Apple-1 computer board handmade in Jobs’ garage for $2.75 million, a 1976 Wells Fargo check signed by Jobs and Steve Wozniak for $2.4 million, and a rare early Apple promotional poster for $650,000. Google and Microsoft memorabilia also sell, though they command prices in the low thousands, not millions.

Once the purview of Hollywood and rock-n-roll fans, the memorabilia market has spent the last decade making room for tech artifacts the same way it made plastic Star Wars action figures high-value assets.

Bobby Livingston, executive vice president at RR Auction, describes it as a sort of high-end participation trophy. “That moment when a creator comes up with a paradigm-shifting product, that’s the touchstone for them emotionally,” he says. “It means something.”

But the market is brutally binary. Apple relics are king, and Google and Microsoft might command respectable prices, but everything else is garage dregs.

“Failed companies’ t-shirts and swag have no value if nobody’s heard of that product anymore,” Livingston says. “Apple has a cult, and the rest are lost to history.”

Livingston thinks it will take another 25 years for AI swag to mature into a formal auction category, but he’s already seeing signs of froth. Early graphics cards and computing boards hand-signed by Nvidia CEO Jensen Huang are currently selling for between $8,000 and $10,000. Last year, AI influencer Linus Ekenstam received a Google Gemini TPU, one of the computer chips that the search giant makes for training AI models, entombed in an acrylic box. He views it as a family heirloom and plans to pass it down through generations of his family.

Those looking to make a quick buck today should target Anthropic swag, as it has been historically hard to find. Alejandro Navia, who works at an AI company, said that he switched his flight to New York City just so he could stand in line to receive a free Anthropic baseball hat embroidered with the words “thinking.” They now run between $50 and $100 on eBay.

When he learned he could make some cash from selling it, he considered it. But he quickly decided against it.

“I’d rather have it for the lore; the bragging rights,” he said. “The Valley is documented via swag. ‘Oh you were there for that one? Yeah. So cool!’”





More from Forbes

Forbes Israel’s Palantir Rival Is Selling $1 Million Spy Vans To U.S. Cops

Forbes These Gen Z Founders Are Reinventing Dating Apps, Without The Swipe

Forbes AI Data Centers Are Set To Electrify This $13 Billion Family’s 76-Year-Old Company

The Booming Secondary Market For Used OpenAI and Anthropic Swag

Microsoft appears to be gearing up its sales force to enhance its competitive stance in the rapidly evolving AI sector.

During an internal meeting held on Tuesday, company leaders revealed a strategy aimed at positioning their AI offerings more favorably in comparison to those from rivals such as OpenAI, Google, and Anthropic. According to a Bloomberg report, this session, which focused on strategies for the upcoming fiscal year, emphasized the advantages of Microsoft’s proprietary AI models in terms of efficiency and cost-effectiveness.

Jay Parikh, Executive Vice President, reportedly urged attendees, “While others are offering pieces, we’re presenting a complete end-to-end solution. This narrative is essential for us in fiscal year 27.”

Adding to the discussion, Executive Vice President Jacob Andreou delivered a presentation that placed Microsoft’s Copilot side by side with Anthropic’s chatbot, Claude. Bloomberg noted that Andreou pointed out specific performance shortcomings in Anthropic's model, highlighting its slower speed, lower accuracy, and insufficient security features when integrated with Microsoft’s office applications.

TechCrunch has reached out to both Microsoft and Anthropic for reactions and will provide updates should they respond.

While it’s not unusual for a company to coach its sales team on competitive tactics, what stands out is Microsoft’s choice of targets—rival firms it has historically relied on for the AI technology behind its products.

This aligns with a broader shift within the company. A recent report indicated that Microsoft has been removing OpenAI and Anthropic's models from flagship applications like Word and Excel, opting for its proprietary solutions instead as a cost-saving measure.

In the past, Microsoft and OpenAI had a tightly-knit partnership, which included financial backing and computational resources provided by Microsoft in exchange for exclusive access to OpenAI’s API and models. However, changes to this partnership in April lifted exclusivity, now permitting OpenAI to collaborate with Microsoft’s competitors.

This evolving landscape may shed light on the new approach of Microsoft’s sales team. Over the past year, the company has faced pressure regarding its stock performance, as investors have raised concerns about its substantial investments in AI development. By stressing the competitiveness of its products, Microsoft aims to instill confidence among stakeholders about its future in the AI arena.

Microsoft is said to be training its sales staff to undermine OpenAI and Anthropic.

A dimly lit, shadowy corporate training room: a large Microsoft logo glows softly on a dark projector screen. In the foreground, silhouettes of suited salespeople huddle over a podium strewn with briefing packets—one shows plain “OpenAI” text in black, another bears the Anthropic logo. The mood is conspiratorial, whispers and cautionary glances under a single overhead lamp.

Microsoft is reportedly training salespeople to talk down OpenAI and Anthropic - TechCrunch

- Microsoft is preparing sales teams to competitively promote its in-house AI models.  
- Executives emphasized the efficiency and cost-effectiveness of Microsoft’s AI versus competitors.  
- Microsoft is replacing OpenAI and Anthropic models in its flagship applications for cost savings.

Microsoft appears to be prepping its sales team to get more competitive with the other major players in the AI industry.

At an internal meeting on Tuesday, the company’s executives outlined a plan for salespeople to negatively compare AI products from companies like OpenAI, Google, and Anthropic to its own, according to a new report from Bloomberg. The meeting, billed as a strategy session for the new fiscal year, reportedly leaned heavily on pitching the efficiency and cost-effectiveness of Microsoft’s in-house models against those of its rivals.

“Everyone else is selling parts — we’re selling the full end-to-end system. That’s the story that we all need to get out there and tell in FY27,” Executive Vice President Jay Parikh reportedly told the room.

Executive Vice President Jacob Andreou reportedly went further, delivering a presentation comparing Copilot directly to Anthropic’s chatbot Claude. According to Bloomberg, Andreou noted that, when it came to performance within Microsoft’s office apps, Anthropic’s model was “slower and less accurate, and lacked the proper security integrations,” Bloomberg writes.

TechCrunch has reached out to Microsoft and Anthropic for comment and will update this story if we hear from either outfit.

A company coaching its sales team on how to trash-talk competitors isn’t particularly surprising. What’s more notable is who Microsoft is now targeting — the same companies it has long depended on for the AI models powering its own products.

It’s just the latest move in that direction. A report earlier this month found that Microsoft has been swapping OpenAI and Anthropic’s models out of flagship apps like Word and Excel in favor of its own — a cost-cutting move, according to that report.

There was a time when Microsoft and OpenAI were attached at the hip. The two companies entered into a very unique agreement years ago that saw Microsoft provide capital and compute to OpenAI while allowing Microsoft to enjoy exclusive access to OpenAI’s API and models. The companies amended the partnership in April, dropping the exclusivity clause and clearing OpenAI to sell to Microsoft’s competitors.

That revised relationship may help explain the sales team’s new pitch. Microsoft has been battling a less-than-optimal stock outlook over the past year, as investors question the company’s massive spending on the buildout of its AI business. Talking up how competitive those products actually are is likely an attempt to calm those waters and build confidence in Microsoft’s long-term AI plan.

Microsoft is reportedly training salespeople to talk down OpenAI and Anthropic

Just over a year ago, OpenAI's CEO Sam Altman revealed ambitious plans for the company to create an AI-centric device with the collaboration of renowned Apple designer, Jony Ive. Recent insights, as reported by Bloomberg on Tuesday, suggest we might finally glimpse what this innovative gadget could entail.

While specifics have been limited, sources indicate that the anticipated device will function as a “screen-free smart speaker,” envisioned to serve as a novel type of home computer tailored for the AI-driven landscape. This device aims to interact in a “human-like” manner, featuring smart speaker attributes that facilitate control over household appliances, media playback, message responses, and information retrieval through OpenAI's ChatGPT technology.

According to Bloomberg's insiders, the device will possess AI capabilities that enable deeper personal interactions with users. It is expected to become increasingly personalized and proactive, utilizing data such as personal emails. The inclusion of mechanical components designed to move autonomously could provide users with the impression that the device possesses a semblance of life, rather than merely reacting to commands.

As of now, OpenAI has not responded to inquiries from Fortune regarding the device's development.

The landscape for AI-focused home gadgets is still nascent. While Amazon and Google have integrated AI functionalities into their pre-existing smart speakers, Google recently introduced a new Google Home speaker specifically developed for its Gemini AI. However, initial reviews for this product have been mixed; critics have acknowledged its advanced conversational abilities and features like the Gemini Live chat but have also pointed out issues with the device sometimes misinterpreting questions or becoming confused during exchanges.

Currently, there are only a limited number of true AI-first devices on the market. One such example, which resembles what OpenAI is reportedly planning, is ElliQ. Developed by Intuition Robotics and launched in 2022, ElliQ is designed for older adults, presenting a lamp-like appearance complete with moving parts. Similar to OpenAI's concept, it aims to provide companionship by engaging users throughout the day with humor and questions. Its AI technology evolves to become more tailored to individual users and assists with monitoring wellness metrics such as heart rate and blood pressure.

Startup Humane also unveiled a hands-free AI pin as a smartphone alternative, but faced backlash over performance issues and overheating. Ultimately, HP discontinued the product after acquiring the startup last year for $116 million.

While there is no confirmed launch date for OpenAI's device, the company’s chief global affairs officer, Chris Lehane, indicated earlier this year that a reveal could happen in the latter half of 2023. Despite being in developmental stages, this new AI-centric product has already generated a competitive stir among established technology companies.

Recently, Apple initiated legal action against OpenAI and two of its former employees who transitioned to the startup, filing a suit in the U.S. District Court for the Northern District of California. Apple accuses OpenAI of utilizing insider information from ex-employees to boost its fledgling hardware division, alleging that there were inappropriate inquiries into confidential projects and the solicitation of prototypes.

In response, an OpenAI spokesperson stated that the company is not interested in other organizations' trade secrets and is dedicated to crafting innovative technology aimed at benefiting users globally.

The introduction of OpenAI’s new device could enhance the company’s prospects as it gears up for a potentially record-setting initial public offering (IPO) after filing necessary documentation last month. Though the timeline for a public listing remains uncertain, OpenAI is keen to leverage the substantial investments it has secured over recent years.

Last year, OpenAI made headlines by acquiring Jony Ive’s design firm, IO, for an impressive $6.5 billion, marking the beginning of a partnership to realize its AI-first vision. Altman has expressed that the company’s goal is to create a range of products that could potentially replace traditional smartphones and computers. As he noted during an OpenAI podcast in June, the conventional understanding of computers, software, and hardware was shaped in an era devoid of AI; now, as society evolves, so too must the technology we utilize.

OpenAI aims for its speaker to have a lifelike quality, while Apple claims it's a copied concept.

A dim, moody room with a single warm spotlight on a sleek smart speaker perched on a pedestal. The speaker’s grille glows like a beating heart, and “OpenAI” text (no logo) is subtly printed at its base. Behind it on a shadowy wall, a large Apple logo looms above a cracked patent document pinned with dramatic emphasis. Low-key, dramatic lighting underscores the legal clash.

OpenAI wants its speaker to feel alive. Apple says it’s a stolen idea - Fortune

- OpenAI plans to create a screen-free smart speaker as an AI-first home device.  
- The device aims for personal interaction, drawing on user information for responses.  
- Apple has sued OpenAI over alleged misuse of trade secrets from former employees.  

It’s been just over a year since OpenAI CEO Sam Altman announced the company would build an AI-first device with the help of legendary Apple designer Jony Ive. Now we may have an idea of what it might look like.

Firm details about the device have been scant, but on Tuesday, Bloomberg reported, cited people familiar with the matter, what the device may look like: a “screen-free smart speaker designed to be a new type of home computer for the AI era.”

The device is meant to be “human-like” with smart speaker-like features that help control appliances in a user’s home, play media, respond to messages and answer queries by using OpenAI’s ChatGPT, the people added.

The device’s AI capabilities could allow it to better interact with users on a personal level, according to the report. It will also become increasingly personalized and proactive over time as it draws on information like a users’ personal emails. Mechanical parts that move on their own also give users “a sense that it is alive and not just an object responding to commands,” the outlet reported.

OpenAI did not immediately respond to Fortune’s request for comment.

AI device market

The market for AI-first home devices is still in its infancy. Both Amazon and Google have incorporated AI capabilities into their existing smart speakers. Meanwhile, Google last month also released a new Google Home speaker designed specifically for Gemini.

Still, reviews for Google’s product have thus far been mixed. While critics praised the AI-enabled speaker’s more natural conversations, and especially a paywalled live chat feature called Gemini Live, they also noted the device has a tendency to ignore some questions and can get confused during interactions.

Only a handful of truly AI-first devices exist on the market today. One example that could resemble OpenAI’s planned product—according to Bloomberg’s description—is ElliQ, an AI device designed for older adults that resembles a desk lamp and also has moving parts. First released by Intuition Robotics in 2022, ElliQ, like OpenAI’s device, is meant to be humanlike and comfort elderly people by prompting users several times a day with jokes or questions. The device uses AI to become more personalized to a user over time, and can serve as a companion but also can help track wellness information like heart rate and blood pressure as well as appointments.

In addition, startup Humane advertised its hands-free AI pin as an alternative to smartphones when it released the product in 2024, yet the product was criticized for slow performance and overheating. Ultimately, HP shuttered the product after acquiring the startup last year for $116 million.

OpenAI’s device still has no release date, but in January, OpenAI’s chief global affairs officer Chris Lehane said it was set to unveil the product in the second half of this year. Even while still in development, the prospect of a new AI-first device to compete with products from established tech giants already has competitors on the defensive.

Last week, Apple sued OpenAI and two of its former employees who joined the startup in the U.S. District Court for the Northern District of California. The company alleges OpenAI leveraged the insider knowledge of its ex-employees to build up its fledgling hardware division. The company alleged that OpenAI asked ex-Apple employees in interviews about secret projects and asked them to bring in prototypes.

“OpenAI’s nascent hardware business now rests on the shakiest of foundations, rotten to its core by its illegal reliance on misappropriated trade secrets,” Apple wrote in the lawsuit.

For its part, an OpenAI spokesperson said in a statement to multiple outlets that it has no interest in other companies’ trade secrets and that the company remains “focused on building innovative technology that empowers people everywhere.”

OpenAI’s new device could supercharge the company’s prospects as it approaches a potential record-breaking IPO after filing the paperwork to go public last month.

While it’s still unclear when OpenAI might list its shares publicly, the company is looking for ways to capitalize on the hundreds of billions of dollars it has raised from investors in the past several years.

Last year, the company bought Ive’s design studio IO for a whopping $6.5 billion and announced a partnership with the Apple designer to bring its AI-first product to fruition. At the time of the announcement, the pair said they had been collaborating for about two years.

Altman has said the company is aiming to develop a new family of products that can replace static smartphones and computers.

“Computers, software, and hardware, just the way we think of current computers, were designed for a world without AI,” Altman said on the OpenAI podcast last June. “Now we’re in, like, a very different world, and what you want out of hardware and software is changing quite rapidly.”

OpenAI wants its speaker to feel alive. Apple says it’s a stolen idea

Thinking Machines Lab, an AI startup established by former OpenAI CTO Mira Murati, unveiled its first proprietary AI model, Inkling, on Wednesday morning. This model diverges from the major offerings of OpenAI, Anthropic, and Google by being open-weight, allowing external developers and companies to download and modify it freely.

Inkling is a mixture-of-experts system boasting a total of 975 billion parameters, although it employs approximately 41 billion for specific tasks. This design is common among large models as it enhances efficiency and cost-effectiveness. The model was trained on an extensive dataset of 45 trillion tokens encompassing text, images, audio, and video, enabling it to natively reason across all four modalities. Currently, however, Inkling is limited to text outputs, which include code, styled outputs, and structured data.

This launch marks Thinking Machines Lab’s first public demonstration of its AI capabilities after a year and a half of working largely behind the scenes on its infrastructure. Earlier this year, the company shared a preview of its “interaction models,” intended to engage users more dynamically than traditional chatbots. Inkling also tests the core assumption of the startup: that AI models which organizations can adapt themselves will ultimately prove more effective than the generalized models offered by larger labs.

Inkling is designed to provide nuanced responses, alerting users to uncertainty rather than generating random guesses. It also allows users to adjust the “thinking effort” applied during tasks, providing a trade-off between response speed and the depth of analysis required. In comparative benchmarks, the company claims that Inkling operates using a third of the tokens needed by Nvidia’s Nemotron 3 Ultra to achieve similar coding results.

It is important to note that Thinking Machines does not assert that Inkling is the leading model in the market. In its documentation, the company explicitly states that Inkling is "not the strongest model available today, closed or open," and rather aims for balanced performance and adaptability.

The primary target audience for Inkling appears to be enterprises, as Thinking Machines is promoting the model as a foundation for customization rather than a final product. Organizations will be able to refine Inkling through Tinker, the company’s model-customization platform, although this places some responsibility on users to ensure their modifications are secure, which necessitates skilled machine learning knowledge.

Unlike OpenAI’s ChatGPT or Anthropic’s Claude, which focus on general-purpose chatbot functionalities, Inkling is positioned to be a more versatile tool for organizations wishing to tailor their AI solutions.

In a post released by Thinking Machines prior to the launch, the company argued that centrally trained AI models often fail to meet the nuanced needs of specific enterprises, as much of the expertise needed is unique to those organizations. This sentiment echoes growing concern about the value of centralized models, as articulated by Microsoft CEO Satya Nadella, who pointed out that companies effectively incur dual costs when relying on proprietary AI models — initially through subscription fees and subsequently by relinquishing proprietary business insights through usage data.

Hugging Face’s CEO, Clem Delangue, echoed these thoughts, stating during a recent discussion that mainstream AI operations are likely to veer toward private or open-source models as opposed to centralized solutions, aligning with the path that Thinking Machines is pursuing.

Evidence supporting Thinking Machines’ hypothesis emerged from a collaboration with Bridgewater Associates, the largest hedge fund globally (which does not invest in Thinking Machines). The team trained an existing open-source model with Bridgewater's financial expertise, achieving a score of 84.7% on financial reasoning tests that surpassed leading proprietary AI models while incurring significantly lower operating costs — though it’s crucial to mention these results originated from the companies’ own assessments, not independent validation.

Thinking Machines emphasizes its rapid development timeline, claiming to bring Inkling to market in about nine months, substantially quicker than the timelines observed for OpenAI and Anthropic.

Questions have arisen around whether Inkling utilized outputs from competitor models, known as “distillation,” a practice under scrutiny in the industry. According to the company, Inkling was predominantly trained from scratch but did incorporate some early post-training data from other open-weight models, including Moonshot AI’s Kimi K2.5, before transitioning to large-scale reinforcement learning. Future models are anticipated to rely exclusively on self-contained post-training methods.

On the financial front, Thinking Machines has maintained a level of caution. Earlier this year, the company formed a strategic partnership with Nvidia, harnessing substantial computing capacity for training Inkling on Nvidia’s advanced systems. However, the company has been reticent regarding its revenue plans, which reportedly have not been a primary focus. There were discussions about a potential $50 billion fundraising round, but updates indicate those efforts have faced delays.

Another key consideration is the scalability of Thinking Machines’ spending relative to competitors like OpenAI and Anthropic. The startup seems to favor a more efficient spending model, possibly eliminating the necessity to match the financial outlays of larger rivals. The open-weight framework of Inkling means that once the model is in the public domain, there is no obligation for users to pay for its operation, contrasting the subscription models employed by OpenAI and Anthropic. Instead, revenue for Thinking Machines will largely stem from Tinker, focusing on training, fine-tuning, and the surrounding ecosystem.

Thinking Machines’ workforce appears to have stabilized, currently employing around 200 individuals, a recovery from earlier attrition, including the exit of two co-founders who joined OpenAI.

The company cultivates a culture that prioritizes stability over individual prominence, a calculated approach that minimizes disruption when changes occur within the team. This is notably significant as much of the company’s narrative is still linked to its co-founder’s well-known name, balancing recognition with a commitment to organizational continuity.

Thinking Machines enhances its commitment to customized AI with the launch of its initial open model, Inkling.

A bright, sunlit AI research lab with a large Thinking Machines logo on the wall; at its center a sleek, vault-style device swung open, letting glowing blueprint pages labeled “Inkling” swirl out like paper birds; a lone researcher adjusts a dial as warm natural light streams through tall windows for an optimistic, creative vibe.

Thinking Machines amps up its bet against one-size-fits-all AI with its first open model, Inkling - TechCrunch

- Thinking Machines Lab launched Inkling, an open-weight AI model for customization by developers.  
- Inkling features 975 billion parameters but utilizes only 41 billion per task efficiently.  
- The startup aims to outperform centralized models by enabling enterprise-level AI customization.  

Thinking Machines Lab, the AI startup founded by former OpenAI CTO Mira Murati, released its first proprietary AI model Wednesday morning, called Inkling. And unlike the flagship models from OpenAI, Anthropic, or Google, it’s open-weight, meaning outside developers and companies can download it and modify it directly.

Inkling is a mixture-of-experts system with 975 billion total parameters, though it only draws on a fraction of that — about 41 billion — for any given task, a common design that keeps very large models faster and cheaper to run. It was trained on 45 trillion tokens of text, image, audio, and video, and reasons natively across all four, , according to the company’s own release materials. Its current generation capabilities are limited to text output, though, which includes code, styled artifacts, and structured data.

The model is Thinking Machines Labs’ first public proof point after a year and a half spent building AI infrastructure largely out of public view. Some of that work surfaced already, in a May research preview of “interaction models” — AI designed to listen and speak (and even interrupt) instead of stop and wait as with typical chatbots. It’s also a test of the central bet behind the startup, which is that AI that organizations can adapt for themselves will outperform the one-size-fits-all models the biggest labs currently sell.

Inkling is an interesting model, one that’s designed to give calibrated answers, including flagging uncertainty rather than guessing, and which lets users dial “thinking effort” up or down when they want to trade for speed. On one benchmark, the company says, Inkling uses a third as many tokens as Nvidia’s Nemotron 3 Ultra — its latest generation open-weight models — in order to hit the same coding performance.

It’s worth noting that Thinking Machines doesn’t claim Inkling is best-in-class. Its briefing materials state explicitly that Inkling is “not the strongest model available today, closed or open.” What it’s evidently going for instead is well-rounded performance and customizability.

Of course, that raises the question of who this product is targeting, beyond being decidedly an enterprise product. Thinking Machines is, for now, marketing Inkling less as a finished work than as a starting point, something for organizations to fine-tune themselves through Tinker, the company’s model-customization platform, but this means that their own customers have to make sure their customizations are safe, for example. (Fine-tuning requires serious machine learning talent.)

OpenAI, Anthropic, and Google have all taken a very different approach with ChatGPT, Claude, and Gemini, respectively, which were all built to compete as general-purpose chatbots first, with agentic, autonomous features layered on top.

A post published by Thinking Machines last week was clearly meant as the backdrop for this release. AI that’s trained centrally by one company and then set in stone, the company argued in that post, underperforms AI that organizations shape themselves because so much expertise is specific to the people who hold it. The broader idea is that centralized labs are selling everyone the same product, repeatedly refined by the lab that built it, while enterprises willing to own and customize their own models can wring far more value from them.

It’s an argument that’s gaining steam. In a blog post published Sunday, Microsoft CEO Satya Nadella — whose company has invested billions in both OpenAI and Anthropic — warned that enterprises using proprietary AI models effectively pay twice: once in subscription costs, and again by handing over business knowledge embedded in their thousands of prompts and corrections, which can be absorbed into future model versions.

Hugging Face CEO Clem Delangue made a similar prediction in conversation with TechCrunch last week. Frontier models, he said, will increasingly be reserved for experimentation and high-value tasks, while most production AI work shifts to private or open-source alternatives — the exact split Thinking Machines is building around.

The clearest evidence for Thinking Machine’s argument came from a recent project with Bridgewater Associates, the world’s largest hedge fund (which is not, for what it’s worth, a Thinking Machines investor). Researchers from both companies took an existing open-source model and trained it further on Bridgewater’s own financial expertise. The result was said to score 84.7% on financial reasoning tests, beating top proprietary AI models, while costing roughly a fourteenth as much to run — though those results come from the two companies’ own evaluation, not an independent one.

Either way, Thinking Machines is emphasizing how quickly it got here. OpenAI took roughly five years to bring tech to market and show revenue, and Anthropic roughly three. Thinking Machines says it did the same in about nine months.

Some will wonder whether Inkling was trained on outputs from competitors’ models, a practice known as “distillation” that has drawn scrutiny industry-wide. The short answer, per the company’s own materials, is partly. Thinking Machines pre-trained Inkling from scratch, but it says it used other open-weight models — including Moonshot AI’s Kimi K2.5 — to help generate some of its early post-training data before large-scale reinforcement learning took over. The next model, the company insists, will use fully self-contained post-training instead.

On the cost side, Thinking Machines has been more guarded. It struck a strategic partnership with Nvidia in March to deploy a gigawatt of Vera Rubin computing capacity, and says Inkling itself was trained entirely on Nvidia’s GB300 NVL72 systems. But the company hasn’t said how it plans to balance that against revenue that, by most accounts, hasn’t been a primary focus so far. (A reported $50 billion fundraising round was said to be coming together last November, which multiple outlets reported had stalled by January; the company has declined to talk about its funding picture since, though Nvidia said it made a “significant investment” in Thinking Machines when the companies announced that March partnership.)

A related question is whether Thinking Machines’ spending will ever reach the scale of OpenAI’s or Anthropic’s, or whether its efficiency-driven approach means the economics look different. Put another way, the company’s bet may be less that it will eventually spend like its larger rivals than that it won’t need to at all — because once weights are public, nothing obligates anyone who downloads them to pay Thinking Machines to run them, unlike the metered access OpenAI and Anthropic sell. It’s Tinker, not the model itself, where the company’s revenue has to come from, via training, fine-tuning, and, now, a cut of the hosting ecosystem built around it.

Headcount, at least, looks more settled. Thinking Machines now employs roughly 200 people, up from levels reported after a wave of departures earlier this year, including two co-founders who left for OpenAI in January.

Thinking Machines, for its part, doesn’t seem interested in playing up individual moves the way much of the industry does. According to a source inside the company, its culture, by design, favors continuity over reliance on any one personality. It makes sense: it’s less of a setback when people change teams if they were never put on a pedestal to begin with. It’s also a remarkable thing for a company to insist on, given how much of its own story is still associated with the name of its now-famous co-founder, whether she planned it or not.

Thinking Machines amps up its bet against one-size-fits-all AI with its first open model, Inkling

Anthropic, a leading player in the field of artificial intelligence, is adopting a competitive strategy that encourages states to implement increasingly stringent AI regulations, rather than rallying around a unified national standard.

This stance sharply contrasts with that of its primary competitor, OpenAI, which has been advocating for a more harmonized approach to AI regulation among lawmakers. 

Cesar Fernandez, who heads U.S. state and local government relations for Anthropic, articulated this perspective in a recent interview with POLITICO, part of the Axel Springer Global Reporters Network: "While some in the industry see state policy as a way to establish a limit for federal legislation, we are not simply aiming to back the same bill in every state. Our goal is to advance legislation that genuinely enhances the safety of the most advanced AI systems."

Fernandez's remarks were prompted by inquiries about OpenAI's ongoing efforts to influence state-level AI regulations. OpenAI's chief lobbyist, Chris Lehane, has coined the term “reverse federalism” to describe their strategy of bypassing a stagnant Congress and creating a national AI framework by adapting regulations state by state.

This subtle critique of OpenAI aligns with Anthropic’s history; several of its executives departed OpenAI in 2020 due to concerns over a lack of priority on safety. Since then, Anthropic has actively advocated for tougher AI safety regulations at both the federal and state levels—an initiative that some critics, particularly from the Trump administration and venture capital circles, have perceived as an effort to weaken regulatory bodies and eliminate competition.

In defense of their approach, OpenAI spokesperson Liz Bourgeois emphasized that “reverse federalism, where effective state safeguards shape national standards, assists regulators in enforcing the law, clarifies protections for the public, and enables developers to concentrate on safety rather than conflicting standards.”

The difference in tactics between Anthropic and OpenAI surfaces at a pivotal moment for AI regulation. With Congress hesitant to take action and the White House oscillating between a lenient and strict stance, the AI sector is increasingly looking at state governments for regulatory guidance. The eventual decision by state legislators to either consolidate around a unified AI safety framework or engage in a competitive regulatory landscape will significantly influence the future of AI rules in the United States.

Fernandez expressed a desire for a federal regulatory framework but acknowledged the need for immediate action: “A government response to the challenges posed by advanced AI models cannot wait for Washington's intervention.”

Setting Anthropic apart, Fernandez highlighted the company’s proactive stance in state policy debates. Notably, Anthropic was the first major AI laboratory to endorse California's upcoming 2025 law aimed at regulating advanced AI models, establishing a precedent in the country.

In contrast, OpenAI did not take a public stance on the California proposal before it was passed, but has since pointed to the law as a model for others to follow due to its focus on increasing transparency regarding companies' safety measures.

Anthropic views the California regulations as an opportunity to amplify its AI safety initiatives, with Fernandez noting, "Each of those bills has been more robust than its predecessor, driving forward real safety obligations." He believes that mere transparency and self-reporting are no longer sufficient.

Fernandez also referred to Anthropic’s Claude Mythos model, which reportedly exhibited the ability to exploit vulnerabilities in major operating systems during its evaluations. These concerns over cybersecurity prompted significant alarm within the Trump administration, leading to the imposition of export controls until potential risks were addressed.

Last year, OpenAI successfully lobbied New York Governor Kathy Hochul to modify her state's AI safety legislation so that it mirrored California’s guidelines more closely. However, it surprised many safety advocates by aligning with Anthropic on an Illinois initiative that implemented stricter requirements than those in New York and California. This newly enacted law mandates annual independent audits of safety plans for leading AI corporations.

Continuing its push for rigorous standards, Anthropic also endorsed evolving regulations in Massachusetts, which it regards as the most robust AI safety proposal in the country. The proposed legislation would require leading AI firms to engage independent evaluators to assess risks related to catastrophic outcomes, such as the facilitation of bioweapons development, and empower the state's attorney general to enforce these mandates.

Bourgeois noted that OpenAI is still evaluating the Massachusetts proposal but supports the emphasis on AI safeguards put forth by the state legislature.

Beyond regulatory approaches, both companies have engaged in political maneuvering. They maintain competing super PAC networks that have invested millions into political campaigns throughout the nation. Recently, Anthropic began making direct contributions to California legislators.

Fernandez shared that, “We support candidates for election and re-election who share our perspective on AI safety regulation, as it aligns with our mission to ensure the transition to a world with powerful AI benefits society at large.” He stated that they do not coordinate with employees who have also donated to political candidates, explaining, “We don’t direct our employees to contribute, but they are drawn to Anthropic because of their concerns about AI’s future and the need for effective safety policies.”

Exploring Anthropic's strategy for enhancing AI regulations in each state

A dimly lit strategy room centered on a giant digital map of the United States, each state outlined in glowing red chains. In front of it, an adult executive’s silhouette studies the map, hand raised as if plotting new boundaries. The Anthropic logo glows prominently on the wall-mounted display above the map, casting deep shadows around the room.

Inside Anthropic's State-by-State Plan to Ratchet up AI Rules - Business Insider

- Anthropic advocates for stronger AI regulations, contrasting OpenAI's push for unified standards.  
- California's AI law is a model for Anthropic’s push for stricter state regulations.  
- Anthropic endorses strict measures, like independent audits, to ensure AI safety compliance.  

Artificial intelligence giant Anthropic is pursuing a strategy of one-upmanship that encourages states to impose increasingly tougher AI guardrails, rather than align around a single set of regulations.

The approach stands in stark contrast to the one favored by the company's archrival, OpenAI, which has pushed state lawmakers toward common ground on regulating the breakthrough technology.

"While there are some in the industry that think of state policy as a way to create a ceiling for federal legislation, Anthropic is not just looking to support the same bill across the country in every single state," Cesar Fernandez, the company's head of U.S. state and local government relations, said in an interview with POLITICO — which, like Business Insider, is part of the Axel Springer Global Reporters Network — on Tuesday. "We're looking for legislation that meaningfully raises the bar on safety for the most capable AI systems."

Fernandez's comments came in response to questions from POLITICO about OpenAI's ongoing campaign to shape states' AI regulations. The ChatGPT maker's top lobbyist, Chris Lehane, has coined the term "reverse federalism" to describe its attempts to bypass a paralyzed Congress and build a national AI framework by mirroring bills state-by-state.

The veiled jab at OpenAI is on-brand for Anthropic, whose executives left OpenAI in 2020 over concerns the company wasn't prioritizing safety. Anthropic has consistently pushed for stronger AI safety rules at both the federal and state level — an effort that some critics, particularly those close to the Trump administration and in venture capital, frame as an attempt to hamstring regulators and lock out competitors.

In a statement, OpenAI spokesperson Liz Bourgeois defended its approach, saying "reverse federalism, where effective state safeguards shape national standards, helps regulators enforce the law, gives the public clearer protections, and allows developers to focus resources on safety rather than conflicting requirements."

The split between OpenAI and Anthropic's approach to statehouses comes at a critical time for AI regulation. With Congress reluctant to act and the White House flip-flopping between a light touch and a heavy hand, the AI industry is increasingly looking to states for regulatory clarity. Whether state legislators ultimately coalesce around a single AI safety framework or work to outdo each other over time will have a massive impact on the final shape of AI rules in the U.S.

Similar to Lehane, Fernandez said he wants a federal framework, but that a government response to the risks posed by advanced AI models "can't wait for action in Washington."

The Anthropic lobbyist also set his company apart by touting its early inroads into state policy debates. Anthropic was the only leading AI lab to endorse California's 2025 law to regulate advanced AI models, the first such law in the country.

OpenAI didn't take a position on the California proposal ahead of its passage. But it has since turned to the law, which aims to foster greater transparency into companies' safety plans, as an example for other states to replicate.

Anthropic, on the other hand, saw the California law as a springboard to ratchet up its efforts on AI safety. Fernandez said the rapid development of increasingly powerful AI models was the main factor behind his company's endorsement of more ambitious bills — in New York, Illinois and now Massachusetts — and its move to weigh in earlier in the legislative process.

"Each one of those bills was stronger than the previous bill, and the bills all moved real safety obligations forward," Fernandez said. "Transparency and self-reporting, we don't believe are sufficient anymore."

He pointed to Anthropic's powerful Claude Mythos model, which the company found to be capable of exploiting security flaws in every major computer operating system during its testing. The cybersecurity concerns raised by Mythos (and its public-facing version, known as Fable) sparked panic inside the Trump administration, which slapped export controls on the technology until Anthropic and the government could address alleged vulnerabilities.

Late last year, OpenAI lobbyists successfully pressed New York Gov. Kathy Hochul to amend her state's AI safety bill to more closely resemble California's rules. But to the surprise of some safety advocates, it joined Anthropic in backing an Illinois measure seen as stricter than those in New York and California. That proposal, signed into law this month by Gov. JB Pritzker, requires leading AI companies to submit to annual independent third-party audits of their safety plans — a first-of-its-kind mandate.

Anthropic is pushing the bar further. In late June, it endorsed regulations under development in Massachusetts for an economic development bond bill that Anthropic calls the nation's strongest state AI safety proposal. The language it supported included a requirement for leading AI companies to hire independent evaluators to assess the potential for catastrophic risks such as the technology assisting in the development of bioweapons, as well as a provision empowering the state's attorney general to enforce that mandate.

Bourgeois, the OpenAI spokesperson, said the company is still reviewing the Massachusetts proposal, but added OpenAI supports the state legislature's focus on AI safeguards.

The AI giants have also clashed on the campaign trail. Each is associated with dueling super PAC networks that so far have sunk tens of millions of dollars into political campaigns across the country. And in June, Anthropic started cutting checks directly to California legislators.

"We back candidates for election and re-election when their point of view of AI safety regulation is aligned with our mission to make sure that the transition to a world with powerful AI does well for people in this country and throughout the world," Fernandez said. "We're very much supporting candidates where there's ideological alignment."

Fernandez said the company isn't coordinating with employees who also have made contributions to political candidates in California and elsewhere.

"We don't direct our employees to make contributions, but they work at Anthropic because they're concerned about the future of AI and where this is headed if there's not proper safety policy that's enacted by governments," said Fernandez. "I would assume that that drives them to engage in the political process."

Inside Anthropic's state-by-state plan to ratchet up AI rules

A cohort of regular employees from OpenAI has collectively contributed over $215,000 to a super PAC advocating for enhanced regulations on cutting-edge AI laboratories. The initiative, known as Guardrails Alliance, launched last month with an initial funding of $5 million and positions itself as a grassroots movement supported by tech professionals, labor unions, and other organizations. The Alliance aims to serve as a balance to Leading the Future, a super PAC that champions the AI industry and has amassed over $100 million in backing from prominent figures in the tech world, including OpenAI's president and co-founder, Greg Brockman.

As reported by WIRED, seven current staff members and one former employee from OpenAI have made donations to the Guardrails Alliance. The super PAC disclosed some of its supporters to WIRED ahead of its first quarterly report to the Federal Election Commission, which will be public on July 15. Among those expected to be included in the filing are two OpenAI employees, while future disclosures will reveal five additional contributors.

One of the most significant contributions came from Juan Felipe Cerón Uribe, a research engineer at OpenAI since 2022, who donated $200,000 to the political organization. Cerón Uribe expressed his concerns that the critical research conducted over the past four years aimed at reducing potential societal risks associated with AI may be in vain without appropriate safeguards holding private companies accountable for ethical AI development. “Tech billionaires, like Greg Brockman, have funded the super PAC Leading the Future to advocate for a lack of regulation. Learning about Guardrails Alliance’s initiative to counter LTF was refreshing; it made my decision to support them an easy one,” he remarked.

The contributions from OpenAI's current and former employees represent just a small segment of Guardrails Alliance’s ambitious target to raise $15 million during this election cycle. Their donations pale in comparison to the $50 million that Brockman and his wife, Anna, have pledged to Leading the Future.

Nonetheless, these donations from OpenAI's grassroots workforce underscore a rising discontent within the company regarding its influence on AI policy. Brockman’s financial support for Leading the Future has raised alarms among some employees, prompting them to request clarity on OpenAI's association with the super PAC. While company leaders have attempted to distance themselves from this group, a segment of the workforce is actively using their own funds to oppose Leading the Future's agenda.

OpenAI Employees Are Backing a Competing Super PAC Against Their Employer

A dimly lit office corridor under warm tungsten lights, seen from behind: Sam Altman’s silhouette facing a makeshift fundraising table piled high with stamped envelopes and oversized checks labeled “Rival Super PAC” in red, white, and blue. A group of determined OpenAI staffers crowd around, passing cash-filled folders. On the far wall above Sam Altman, the word “OpenAI” appears in simple black text (no logo). Moody shadows and tense atmosphere capture the internal showdown.

OpenAI Staffers Are Funding a Rival Super PAC to Take on Their Boss - WIRED

- OpenAI employees donated over $215,000 to Guardrails Alliance for stricter AI regulations.  
- Juan Felipe Cerón Uribe contributed $200,000 to promote responsible AI development and accountability.  
- Donations reflect rising tensions among OpenAI employees regarding the company's AI policy influence.

A group of rank-and-file OpenAI employees have donated more than $215,000 to a super PAC pushing for stricter regulations on frontier AI labs. Guardrails Alliance, which launched last month with $5 million in total initial funding, bills itself as a populist effort supported by tech workers, labor unions, and other groups. It’s aiming to be a counterweight to Leading the Future, a pro-AI industry super PAC bankrolled with more than $100 million from technology industry leaders, including OpenAI president and cofounder Greg Brockman.

Seven current OpenAI employees have donated to Guardrails Alliance, as well as one former employee, WIRED has learned. The super PAC exclusively shared some of its donors with WIRED before its first quarterly filing with the Federal Election Commission is made public on July 15. Two OpenAI employees are expected to appear in that filing, while five more are scheduled to be named in Guardrails Alliance’s future disclosures.

One of the largest donations from an OpenAI employee came from Juan Felipe Cerón Uribe, who gave $200,000 to the political spending group. Cerón Uribe, who has been a research engineer at OpenAI since 2022, tells WIRED he spent the last four years working on the company’s strategies for mitigating potential societal harms caused by AI.

“In this time, I’ve become concerned that all that research will have gone to waste if it doesn’t translate to guardrails that hold private companies accountable for the responsible development of AI,” Cerón Uribe said in a statement. “Tech billionaires, such as Greg Brockman, funded the super PAC Leading the Future to keep AI unregulated. I was very happy to learn that Guardrails Alliance is pushing back against LTF; my decision to donate to them was easy.”

Contributions from current and former OpenAI employees make up a small portion of Guardrails Alliance’s larger goal to raise $15 million this election cycle. They are also paltry compared to the $50 million commitment Brockman and his wife Anna have made to Leading the Future.

But regardless of their size, the donations from OpenAI’s rank-and-file workers highlight growing tensions inside the company over its efforts to shape AI policy. Brockman’s donations to Leading the Future have sparked concern among some OpenAI employees, who have pressed executives to explain the company’s ties to the super PAC. OpenAI leaders have since tried to distance themselves from the group, but now some workers are using their own money to directly counter Leading the Future.

OpenAI Staffers Are Funding a Rival Super PAC to Take on Their Boss

OpenAI is set to venture into the consumer tech arena with an innovative portable smart speaker designed without a screen, as reported by Bloomberg on Tuesday. This device aims to redefine personal computing in the age of artificial intelligence, acting as a "humanlike AI companion" nestled within the home, according to Mark Gurman of Bloomberg.

Currently under development, this smart speaker will function similarly to existing devices, responding to inquiries, playing audio content, managing smart home devices, and utilizing features from OpenAI's ChatGPT. However, it is anticipated to elevate the user experience by incorporating a distinct "personality" that fosters a more engaging relationship with users. Bloomberg notes that it will feature mechanical components that exhibit movement, lending an illusion of sentience.

This news follows closely on the heels of a legal dispute, as Apple has filed a lawsuit against OpenAI, alleging the company misappropriated its trade secrets. Apple claims to have discovered a consistent pattern of information theft by former employees who now work for OpenAI, suggesting the issue could be more widespread than initially thought.

In response, OpenAI asserts that its forthcoming product diverges significantly from any Apple offerings, and key sources within the company express confidence that it is "unlikely" to infringe on any Apple trade secrets.

Speculation surrounding OpenAI's foray into hardware has intensified, with rumors indicating potential AI-powered earbuds could be launched later this year or early next year. Furthermore, the company is reportedly collaborating with tech giants MediaTek, Qualcomm, and Luxshare on a smartphone project.

OpenAI's hardware undertakings also feature a noteworthy collaboration with renowned former Apple designer Jony Ive, and there are suggestions that the company could introduce several distinct devices, possibly totaling five, by the close of 2028, including one that could be worn in the ear.

OpenAI has yet to respond to requests for further information on this development.

(Note: Ziff Davis, the parent company of CNET, has filed a lawsuit against OpenAI alleging copyright infringement related to the training and operation of its AI systems.)

OpenAI's Initial Product is Expected to Be a Portable Smart Speaker

A sleek cylindrical portable smart speaker with subtle “OpenAI” text (no logo) engraved on its side sits on a sun-drenched wooden coffee table in a minimalistic living room. Warm natural morning light streams through floor-to-ceiling windows, illuminating a potted plant nearby and casting soft, inviting shadows that evoke a sense of groundbreaking smart-home innovation.

OpenAI's First Device Will Reportedly Be a Portable Smart Speaker - CNET

- OpenAI plans to release a portable smart speaker positioned as a humanlike AI companion.  
- The device will operate like current smart speakers, answering questions and controlling smart-home appliances.  
- Apple has sued OpenAI for alleged trade secret theft, but OpenAI denies any violations.  

OpenAI's entry into the world of consumer devices is expected to begin with the release of a portable smart speaker that lacks a screen, according to a Bloomberg report on Tuesday. The device is being positioned as "a new type of computer for the AI era," Bloomberg's Mark Gurman reported, adding that it would "serve as a humanlike AI companion that lives in the home."

The device, which Gurman wrote is still in development, is expected to operate in much the same way today's smart speakers do: answer questions, play media, operate smart-home appliances, as well as respond to messages and make use of capabilities offered by OpenAI's ChatGPT.

But OpenAI's device is expected to go a bit further into homes by offering a "personality" that would make a humanlike connection with users. It will include "mechanical elements that can move on their own, creating a sense that it is alive," Bloomberg reported.

The product leak comes days after Apple sued OpenAI, accusing the company of stealing its trade secrets. Apple alleged it "uncovered a pattern of theft of Apple's trade secrets by OpenAI employees who were formerly at Apple." Apple also called the instances it uncovered "the tip of the iceberg."

However, OpenAI feels its new product "veers significantly" from any Apple device currently on the market and that it is "unlikely" its new device violates any Apple trade secrets, according to anonymous sources described as knowledgeable of OpenAI's plans.

There's been much speculation lately about OpenAI's hardware ambitions. The company has been linked to a possible release of AI earbuds later this year or early 2027 and is also said to be working on a smartphone in partnership with component-makers MediaTek, Qualcomm and Luxshare.

OpenAI's hardware ambitions have included a high-profile partnership with former Apple designer Jony Ive. One recent rumor was that the devices could include one worn in the ear, but that OpenAI could have as many as five different devices ready by the end of 2028.

OpenAI didn't immediately respond to a request for comment.

(Disclosure: Ziff Davis, CNET's parent company, in April 2025 filed a lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.)

OpenAI's First Device Will Reportedly Be a Portable Smart Speaker

OpenAI is reportedly venturing into the hardware arena with a new mobile smart speaker that will be equipped with AI features, enabling it to connect with ChatGPT and offer various home AI functionalities.

According to a Bloomberg report released Tuesday, this upcoming device, which is still in development, is intended to operate without a screen and is being internally described as a “humanlike AI companion” for the home.

OpenAI has expressed interest in producing a hardware product for a while, with some speculation suggesting ambitions to create its own phone, a move that could directly challenge Apple’s dominance in the market.

What sets OpenAI’s smart speaker apart from typical models is its ability to exhibit a “personality” and to learn proactively about its users, tailoring interactions based on individual preferences. Sources noted that the device would connect to various aspects of a user's digital life, incorporating elements such as email data.

In an intriguing twist, the device reportedly features “mechanical elements that can move autonomously,” and the details suggest it aims to embody a companion-like presence, acting as a tangible representation of OpenAI’s ChatGPT.

Bloomberg highlights that the creation of this device benefits from the expertise of several former Apple engineers, who played key roles in developing iconic products like the iPhone and Mac. However, OpenAI faces significant challenges, as it is currently embroiled in legal disputes related to hardware.

Recently, Apple filed a lawsuit against OpenAI, accusing the company of misappropriating its trade secrets. Apple has indicated that the allegations may only represent a portion of broader misconduct that could come to light during the discovery phase of the trial. OpenAI has refuted these claims.

Sources familiar with OpenAI's endeavors shared with Bloomberg that the company believes its product is distinct from anything Apple currently offers, suggesting that it is unlikely to infringe on Apple’s trade secrets.

This initiative by OpenAI coincides with a growing interest in consumer AI hardware across the tech industry. For example, Hark, an AI lab initiated by Brett Adcock, recently secured $700 million in a Series A funding round at a $6 billion valuation to develop what it describes as “personal intelligence” — a blend of proprietary AI models and custom hardware serving as a “universal interface between humans and machines.”

While Hark has yet to reveal specifics about its device, the significant investment in this sector indicates a robust market interest even before products have launched.

OpenAI's initial hardware product is said to be a speaker without a screen that has mobility.

A sleek, white orb-shaped speaker on soft rubber wheels gliding across a modern, sunlit living room floor; warm daylight streams through large windows, casting gentle shadows on minimalist furniture and potted plants; subtle motion blur around the wheels to show movement; “OpenAI” text etched on its side with no logo.

OpenAI’s first hardware device is reportedly a screenless speaker that can move - TechCrunch

- OpenAI is developing a mobile smart speaker that integrates with ChatGPT and offers AI services.  
- The device aims to provide a personalized experience and interact like a human companion.  
- Apple has sued OpenAI for allegedly stealing trade secrets related to this hardware project.  

OpenAI’s first foray into hardware devices is reported to be a mobile smart speaker with integrated AI capabilities that can sync with ChatGPT and provide other home AI services.

Bloomberg reported Tuesday that the device — which is still currently under development — is designed to be screen-free and is being pitched internally as a “humanlike AI companion that lives in the home.”

OpenAI has long claimed that it wants to launch a hardware product — with some rumors being that it wants to launch its own phone, a move that would put it in competition with Apple.

OpenAI’s newly surfaced device sounds like something of a departure from traditional smart speakers — as sources described the device to Bloomberg as having a “personality” and being able to proactively learn about its owner over time, providing more personalized service. The machine would have access to a user’s digital life, drawing off things like emails, sources said.

The device is also weirdly described as involving “mechanical elements that can move on their own” and the Bloomberg report includes the detail that the device is designed to “feel like a companion and become a physical manifestation of OpenAI’s ChatGPT.”

The device was developed with help from many former Apple engineers who were instrumental in “creating products such as the iPhone and Mac,” Bloomberg writes. Indeed, OpenAI may be attempting to launch a new hardware line, but the company is currently up to its eyeballs in trouble over hardware-related legal problems.

Apple last week sued OpenAI, accusing the AI company of stealing its trade secrets. Apple further claimed that the allegations involved in the suit are merely “the tip of the iceberg” and that more misconduct will be revealed during the legal discovery process. OpenAI has denied wrongdoing.

Citing anonymous sources with knowledge of OpenAI’s plans, Bloomberg writes that the company feels its new product “veers significantly from anything Apple has on the market today” and that it is “unlikely that it violates trade secrets” belonging to Apple.

OpenAI’s push comes as the tech world grows more excited about consumer AI hardware more broadly. Hark, an AI lab founded by Brett Adcock, raised an oversubscribed $700 million Series A back in May at a $6 billion valuation to build what it calls “personal intelligence” — proprietary AI models paired with custom hardware designed as a “universal interface between humans and machines.”

The company hasn’t yet detailed its device’s form factor, underscoring how much capital is chasing this category even before products ship.

OpenAI’s first hardware device is reportedly a screenless speaker that can move

On Tuesday, OpenAI responded to allegations made by Apple in a lawsuit regarding trade secrets, asserting that the claims lack substantiation. 

"We take these allegations seriously, but we are not aware of any evidence supporting their validity," OpenAI stated, as reported by Bloomberg journalist Ed Ludlow on X. "We are advocates for fair competition and the freedom of individuals to work in the environment they choose, and our focus remains on creating innovative technology that benefits people globally."

This response follows Apple’s recent legal action against the AI research organization, accusing OpenAI employees—many of whom previously worked at Apple—of orchestrating a campaign to secure confidential information and proprietary technology. The lawsuit, spanning 41 pages, was lodged on Friday in the U.S. District Court for the Northern District of California and contains serious accusations against OpenAI executives, including Chief Hardware Officer Tang Tan, who has a lengthy history with Apple, having served there for 24 years in high-level roles, including vice president of product design overseeing the iPhone and Apple Watch.

This marks the first instance where OpenAI has publicly addressed the lawsuit. In an earlier statement made shortly after the lawsuit was filed, OpenAI expressed disinterest in the technology of other companies. “We have no interest in other companies’ trade secrets. Our aim is to develop groundbreaking technology that empowers individuals everywhere,” the company told TechCrunch.

In its lawsuit, Apple asserts that its internal investigation revealed evidence suggesting OpenAI and its collaborators utilized the company’s proprietary information in the development of their own hardware product.

Speculation, along with OpenAI's recent acquisition of Jony Ive’s startup, io, indicates the tech lab might be developing a new device that could compete with Apple. Bloomberg reported that OpenAI is in the process of creating a mobile, screen-free smart speaker, described by insiders as a “humanlike AI companion” intended for home use.

This innovative device is said to lack a display, incorporate mechanical components, and is being developed with contributions from several former Apple engineers who have experience with the iPhone and Mac.

TechCrunch has reached out to OpenAI for additional comments and will provide updates as soon as the company responds.

OpenAI challenges Apple's lawsuit over trade secrets.

A dimly lit corporate hallway with warm, low overhead sconces casting long, tense shadows. At the far end, a massive, backlit Apple logo dominates a heavy wood-paneled door. In the foreground, Sam Altman stands seen only from behind, in a sharp suit, clutching a stack of legal briefs. On the polished marble floor ahead of him, the word “OpenAI” is projected in crisp white sans-serif letters (OpenAI text with no logo), surrounded by scattered folders and papers. Dark theme, no artificial blue tones.

OpenAI pushes back on Apple trade secret lawsuit - TechCrunch

- OpenAI denies Apple's allegations of trade secret theft, claiming the lawsuit lacks merit.  
- Apple accuses OpenAI of using confidential information from former employees for its hardware.  
- OpenAI is reportedly developing a mobile, screen-free smart speaker to compete with Apple.  

OpenAI pushed back Tuesday against allegations made by Apple in a trade secret lawsuit, suggesting the complaint lacks merit.

“While we take these allegations seriously, we’re not aware of any evidence that this complaint has merit,” OpenAI said in a statement, first shared by Bloomberg reporter Ed Ludlow on X. “We believe in fair competition and allowing people the freedom to work wherever they choose, and we’re focused on building innovative technology that empowers people everywhere.”

The statement comes several days after Apple filed a lawsuit against the AI lab, alleging that OpenAI employees, who previously worked at the iPhone maker, engaged in a coordinated effort to obtain confidential information and intellectual property. The 41-page complaint, filed Friday in the U.S. District Court for the Northern District of California, contains a string of allegations against OpenAI leadership, including Chief Hardware Officer Tang Tan. Before joining OpenAI, Tan was a veteran at Apple, where he worked for 24 years and held top positions, including vice president of product design for the iPhone and Apple Watch.

This is the first time OpenAI commented on the case itself. In its initial statement hours after Apple filed its lawsuit, it proclaimed a lack of interest in technology developed by other companies, telling TechCrunch: “We have no interest in other companies’ trade secrets. We remain focused on building innovative technology that empowers people everywhere.”

Apple claims in its lawsuit that its internal investigation uncovered evidence that OpenAI and its partners used the company’s confidential information as it develops its own hardware product.

Reports, along with OpenAI’s recent acquisition of Jony Ive’s startup io, suggest the company is working on a device that could directly compete with Apple’s business. Bloomberg reported on Tuesday that OpenAI is working on a mobile, screen-free smart speaker, described by people familiar with the plans as a “humanlike AI companion” designed to live in the home.

The device reportedly has no screen, includes moving mechanical elements, and is being built with input from several former Apple engineers who worked on the iPhone and Mac.

TechCrunch has reached out to OpenAI for further comment and will update this article when the company responds.

OpenAI pushes back on Apple trade secret lawsuit

Searching online has increasingly become a frustrating experience, inundated with AI-generated falsehoods, low-quality sites forcefully climbing search rankings, and an overwhelming number of paid results. As a consequence, many users have turned to ChatGPT as a primary means of finding information. This shift suggests an impending commercialization of the platform. According to a report from the New York Times, OpenAI has entered a partnership with prediction market firm Kalshi to integrate real-time odds for World Cup games directly into responses generated by ChatGPT.

Although the specifics of the agreement between OpenAI and Kalshi have not been publicly disclosed, the Times highlights that while results featuring Kalshi data will credit the source, there is no identification clarifying if this data is part of a paid placement or sponsorship. OpenAI’s documentation regarding ads in ChatGPT states that "ads can appear below a response," and assures users that these ads are labeled as sponsored and kept visually distinct from ChatGPT's content. Thus, it seems OpenAI does not consider this arrangement to be merely a monetization tactic.

On the support page dedicated to ChatGPT's search functionalities, additional context has been provided: "For any queries related to the 2026 World Cup, ChatGPT may present predictive data from Kalshi to offer relevant insights into upcoming matches. When this information is displayed, it will always be accompanied by the label ‘Source: Kalshi.'"

OpenAI emphasizes that this data is intended solely for "informational purposes," clarifying within its usage policies that "you cannot place bets through ChatGPT." This language seems to give OpenAI some degree of separation from Kalshi, even though the primary utility of Kalshi's information revolves around betting. The data reflects public sentiment and expert opinions, but it’s fundamentally linked to a gambling framework.

Regardless of how OpenAI and Kalshi define their collaboration, this is not an isolated incident for Kalshi. Recently, the company, which aims to "trade on anything," has secured various partnerships to expand its market information across multiple platforms, including recent arrangements with major outlets like CNN and CNBC.

The rising popularity of prediction markets is clearly attracting more individuals to participate. While engaging in betting is ultimately a personal choice, it's crucial to understand the risks involved. A recent analysis of bettors from a competitor of Kalshi, Polymarket, disclosed that approximately 70% of accounts have experienced net losses since 2022, with the top 1% of users capturing 77% of all winnings.

Gizmodo attempted to contact both OpenAI and Kalshi for further insights regarding this apparent partnership, but did not receive a response before the article went to print.

Kalshi Odds in ChatGPT is the Unnecessary Combination of Peanut Butter and Chocolate.

Warm sunlight streams onto a cozy wooden desk. Centered is a small plate with a gooey swirl where peanut butter and chocolate bar have melted into one absurd treat. To the right, a laptop screen shows a ChatGPT chat bubble labeled “Kalshi Odds” (OpenAI text with no logo). In the foreground, a betting slip branded with a large Kalshi logo lies casually beside a spilled spoon, hinting at an unnecessary pairing.

Kalshi Odds in ChatGPT is the Peanut Butter and Chocolate of Things You Don’t Need - Gizmodo

- OpenAI and Kalshi are integrating World Cup odds into ChatGPT responses for users.  
- Kalshi data will be marked with "Source: Kalshi," but may not be sponsored content.  
- Research shows high losses among bettors, with top users winning most of the money.  

Online search has become a nightmare, littered with AI-generated misinformation, spammy sites that brute-force their way to the top of results, and tons of sponsored results. It’s been enough to lead people to start using ChatGPT as their de facto search engine, which can only mean one thing: time to enshittify that, too (you know, more than it already is). According to the New York Times, OpenAI struck a deal with prediction market Kalshi to show the current odds for World Cup games right in ChatGPT’s responses.

The agreement between Kalshi and OpenAI has not been announced publicly, the Times notes. Results that include Kalshi data note the source, but offer no disclosure if the data is part of a paid advertisement or sponsorship. According to OpenAI’s own documentation about advertisements in ChatGPT: “Ads can appear below the end of a response. Ads are clearly labeled as sponsored and visually separated from ChatGPTâs response.” So presumably, whatever the arrangement between the companies, OpenAI is not considering it to be Kalshi simply paying for placement.

On a help page specific to ChatGPT Search, the company offered a slight bit of additional clarity: “For queries related to the 2026 World Cup, ChatGPT may show predictive information sourced from Kalshi to provide timely context about upcoming matches. If and when this information appears, it will always be clearly labeled with ‘Source: Kalshi.'”

OpenAI claims the data is for “informational purposes only,” and notes that “You cannot place bets through ChatGPT,” per the company’s usage policies.

That’s a handy bit of distance that the company places between itself and Kalshi, but it’s worth noting that there is little informational purpose for Kalshi data other than betting. It is a reflection of what people are betting their money on and is meant to be used by other people to bet their own money. It may provide a snapshot of public and “expert” or insider sentiment, but even that is in what is effectively a gambling context.

Whatever OpenAI and Kalshi would like to call their arrangement, it’s another in a line of deals that Kalshi has entered into in recent months to make sure their data is everywhere. Late last year, the company that says it wants to “trade on anything” entered into deals to provide market information to CNN and CNBC, which have since integrated that information into their platforms.

Given the widespread normalization of prediction markets, it’s no wonder more people are jumping in all the time. And look, you’re welcome to do whatever you want with your money, just know that the odds are not in your favor. A recent study that analyzed bettors on Kalshi competitor Polymarket found that since 2022, about 70% of accounts have a net loss across their activity, and 77% of all winnings have been funneled to the top 1% of users on the platform.

Gizmodo reached out to both OpenAI and Kalshi for additional information about the apparent partnership, but did not receive a response at the time of publication.

Check out this image of OpenAI's Codex hardware.

More from OpenAI

The Thriving Resale Market for Used OpenAI and Anthropic Merchandise

Microsoft is said to be training its sales staff to undermine OpenAI and Anthropic.

OpenAI aims for its speaker to have a lifelike quality, while Apple claims it's a copied concept.

Thinking Machines enhances its commitment to customized AI with the launch of its initial open model, Inkling.