OpenAI Falls Into AI Hardware Trap: Apple Lawsuit, Youth-oriented Layoffs, and Out-of-Control Product Line
TechFlow Selected TechFlow Selected
OpenAI Falls Into AI Hardware Trap: Apple Lawsuit, Youth-oriented Layoffs, and Out-of-Control Product Line
For investors, this reveals systemic problems in management, compliance, and product execution at a company with a trillion-dollar valuation, while its two major allies, Microsoft and Apple, have both been betrayed.
Navigating Web3 tides with focused insightsAuthor: Michael Spencer from AI Supremacy
Translation: TechFlow
TechFlow Editor's Note: OpenAI is undergoing an unprecedented trust crisis: Apple accuses it of organized theft of trade secrets, executives are leaving one after another, and products are being shut down consecutively. This is not just about a lawsuit, but relates to the entry ticket for the 2027 AI wearable device war. For investors, this reveals systemic issues in management, compliance, and product execution for a company valued at trillions of dollars, while its two major allies, Microsoft and Apple, have both been betrayed.
Is AI Exacerbating Age Discrimination in the Workplace?
Just as the 2027 consumer AI economy is about to kick off the "Physical AI" and wearable device race, OpenAI's possibility of participating faces serious threats—legal disputes and increasing scrutiny are coming head-on. The Apple lawsuit news over the weekend dealt a further blow to OpenAI's already severely damaged brand image and ability to form business partnerships. In 2026, I'm not sure which enterprise client would want to do business with them. But what about ChatGPT users?
OpenAI's Past 24 Hours (Saturday, July 11):
Executive unexpectedly resigns (claiming health reasons)
Shut down Atlas, a browser tool operating for 9 months—another failed product
Sued by Apple for stealing trade secrets
Exposed for selling products to China in violation of sanctions
Finally: Head of Security Systems Johannes Heidecke resigns
Apple's complaint accuses OpenAI of meticulously orchestrating a systematic action to steal and utilize Apple's confidential information and trade secrets, which is very serious in nature. The evidence appears conclusive.
How can this be recovered? I expect Microsoft will also sue them in the coming months, on the grounds that OpenAI violated Microsoft terms to switch to Amazon.
Apple is seeking (explained by Ed Ludlow):
Jury trial
Court order requiring OpenAI to stop abusing Apple trade secrets
Destroy all Apple proprietary materials held by OpenAI
Redesign upcoming OpenAI hardware products to ensure they do not contain Apple technology
Event Timeline
Condensing the 41-page legal document:
I am Chang Liu
Apple Senior Systems Electrical Engineer
Worked on the iPhone team for 8 years
January 2026: Left Apple to join OpenAI
Apple requested return of computer
I ignored it
Haha this is my computer now
Within hours of leaving
Messaged friend Yu-Ting "Alyssa" Peng still at Apple:
Access Apple confidential information using her account
Within a few weeks, borrowed her Apple work computer
February 9: Attempted to access Apple's network storage
That is a cloud repository for confidential engineering files
Authentication had a bug, surprisingly still works!
Messaged Peng: "Haha, I found I can still access [network storage], too funny"
While developing hardware for OpenAI
Downloaded dozens of confidential files
Including a collection of technical documents over 1000 pages
Including MLB (Main Logic Board) manufacturing and test reports
Sent Peng links to Apple proprietary folders
Directed her to look at specific project data
Taught her how to copy files "to avoid being discovered by the security team"
Told her which Apple confidential materials to study before the OpenAI interview
Warned her that another person "messed up" when answering Tang Tan's questions about Apple secret projects
Told her to "download some materials" to look at in advance
Told her: Switch to LINE messaging software, don't let people see these conversations
She got the OpenAI offer, left Apple on April 16
Meanwhile every message remained on the Apple-issued work computer
July 10: Apple Inc. sues Chang Liu
He is the first on the defendant list, before OpenAI and Tang Tan
Stakes
OpenAI is really struggling to keep up with Anthropic, and many of its operational methods are being exposed. Consecutive product failures, project delays, legal troubles, strategic mistakes, and partnership breakdowns all add extra risk to OpenAI—while it is preparing for a heavy-weight IPO sometime next year together with SpaceX and Anthropic. At that time, the company's momentum may have stalled or even declined relative to competitors (Meta, Google, Anthropic, SpaceXAI, Alibaba, ByteDance, etc.). In terms of perception, this completely overshadows the spotlight of GPT-5.6.
Even if you are bearish on OpenAI (like me), watching this scene is painful. After all, the US government mainly funds OpenAI and Anthropic (more than all other companies and projects), making them the future AI leaders. But these are not strategic mistakes—this AI device project is one of OpenAI's major moonshots, yet it is destroyed by anti-competitive misconduct. They are digging their own grave! They might not even get the qualification to participate, because cheating was once an option on their table. Signs of dishonest culture within OpenAI existed long ago, but this legal battle is hard to deny.
What Happens Next
OpenAI will likely have to redesign their wearable devices and AI hardware, which will consume a lot of time and money. Meanwhile, competitors' AI glasses will hit the market in 2027, and other companies including Apple have aggressive product lines.
Apple sues OpenAI for stealing trade secrets, accusing the AI startup and its head of hardware of participating in an organized action to steal information about upcoming products.
In May 2025, OpenAI announced the acquisition of io for $6.5 billion. But their secret hardware project has been delayed, and now with the lawsuit, they might miss the critical window of 2027—when various cool AI wearable devices will be released, including several AI glasses products.
Tang Tan (a 25-year Apple veteran)—also mentioned in the complaint—is a former Apple executive who participated in iPhone and Apple Watch projects, currently serves as OpenAI's Chief Hardware Officer, and is also a co-founder of io. This is very bad for OpenAI's ethics and leadership image. This is not the first lawsuit Tan faces. A less well-known tech startup iyO Inc. once sued Ive and OpenAI CEO Sam Altman for trademark infringement, because the names sounded similar and both parties had past interactions. The startup also sued one of its former employees for allegedly leaking confidential drawings of iyO's unreleased products, and later added accusations of trade secret theft against Tan in the lawsuit.
Apple seeks damages, injunctions, and court orders forcing OpenAI to stop using its trade secrets. And OpenAI looks directionless on products, with lawsuits pouring in from all sides. OpenAI keeps releasing "new" features and products for existing functions, which most companies would ignore at this time.
OpenAI is in Chaos
This lawsuit really opens the door for traditional giants to win the AI wearable and device race. How much chance did OpenAI have to compete against companies like Apple from the start?
OpenAI's Internal Drama Has No End
Before OpenAI acquired io, the OpenAI Startup Fund had already secured 23% of the company's shares for $1.5 billion at the end of 2024, seizing the initiative. At that time, the OpenAI Startup Fund was just an alias for Sam Altman's personal investment. So OpenAI's AI wearable project is actually Sam Altman attempting to compete against companies like Apple. This is not a bet I am willing to make. OpenAI acquired io for over $6 billion, bringing in 55 hardware and manufacturing experts. OpenAI has about 400 former Apple employees.
OpenAI messed up relationships with these two key partners, Microsoft and Apple, which is really serious for the company's future. If you are constantly losing ground on executives, features, products, and partners—this points to structural problems within your company.
Rumored Products
OpenAI's flagship AI wearable device is their "Sweetpea" AI earphones. Sweetpea is designed as a screenless, ambient ChatGPT interaction interface, now probably also including an app unified with Codex. According to so-called leaks, unlike traditional in-ear headphones (like Apple AirPods), Sweetpea adopts a unique behind-the-ear design.
As for the design, I'm not sure if this is accurate, it is claimed that the device consists of a metal egg-shaped charging case, with two pill-shaped modules built-in, hanging comfortably behind the ear, similar to high-end modern hearing aids or open sports headphones.
Considering Meta's AI glasses are very popular, plus Apple's own AI wearable devices are coming, OpenAI competing on hardware was already difficult. Meta itself took many years to make a usable hardware product. OpenAI is traditionally an LLM research lab and has been unsuccessful so far in launching products, apps, and browsers. And these products are frankly much easier than hardware, let alone AI hardware, let alone a whole set of multiple different AI devices. Their ambition is obviously unrealistic.
Leaked supply chain reports show that OpenAI has requested manufacturing partner Foxconn to prepare a production line to produce at least 5 different AI devices by the end of 2028. We are talking about an investment of at least $10 billion. By the time the products hit the market, even ChatGPT's most fervent fans might not use them much. Now who knows when they can hit the market, legal issues haven't been resolved yet.
Jony Ive, Apple's former Chief Design Officer, started collaborating with OpenAI in 2023. Now nearly four years have passed, there is not even a shadow of product announcement, and it may have to be completely redesigned. But 2027 is a critical time point to compete with companies like Apple, Meta, Google, etc. in the AI device and wearable field. For a non-profit company like OpenAI that burns large amounts of cash every quarter, time is money. Trust and credibility are future sales.
Cloud Providers' Backlog is Mostly Anthropic + OpenAI
If OpenAI struggles to compete and is unworthy of a nearly $1 trillion valuation, it will ultimately hurt all cloud providers. Following the trend of companies like Google and Meta, industry consolidation is likely to continue hurting OpenAI. Anthropic's operating profit will reach about $1 billion when it IPOs in October, and ARR may exceed $100 billion by 2027.
OpenAI's path to profitability and survival in such an ecosystem looks weaker and more uncertain every quarter.
OpenAI's Pattern of Misconduct Continues to Expand
If OpenAI wants to IPO, it must abandon Sam Altman. There is no other way to save this company. When Meta aggressively poached people from OpenAI (quite successfully) in June/July 2025, OpenAI was also cruel to Apple's AI team. But it seems these practices have continued until now.
OpenAI's credibility and the arrival of the consumer hardware moonshot seem to be in serious danger.
OpenAI and ChatGPT May Have No Bright Future
According to the lawsuit, io seems to have a habit of trademark and intellectual property theft. Apple has reasons to be a quite secretive company, and OpenAI's disregard for this may be obvious to any rational jury.
What Does Alex Know?
How many times can we give Mr. Sam Altman the green light? Does Palantir know about this lawsuit? Palantir CEO Alex Karp said last week that foundation model companies are stealing your intellectual property, which is why you need data sovereignty. If Apple's accusations are true, OpenAI stole trade secrets from a company worth $4 trillion... You can understand why enterprises might be hesitant to hand over data to them. You can understand why OpenAI hasn't made much progress in enterprise AI or B2B. ChatGPT has a reputation for being an addictive product. With Meta launching new models, and ChatGPT being forced to do advertising, you have to imagine OpenAI's direct competitors are not Anthropic, but Meta, SpaceX, and to a lesser extent Google.
I described Sam Altman as the "Sam Bankman-Fried of AI", and this statement is starting to become increasingly accurate. The problem is that too many parts of the ecosystem need to rely on OpenAI to flourish. OpenAI proposed handing over 5% equity to the Trump government, which is obviously a bad signal for its long-term competitive sustainability and integrity. But when so many parts of the AI bubble are based on circular financing, this makes OpenAI an important part of the gang dynamics within it.
Usually companies with a "loser mentality" like this cannot survive the "consolidation phase" of the new technology wave we are rapidly approaching. Judging from how popular outspoken OpenAI critics like Ed Zitron have become, this is not a good sign for OpenAI's financial situation. This makes everything even more unsettling. But honestly, if the US had rule of law, this company wouldn't even be allowed a chance to go public. But too much capital has already been invested, and cloud hyperscalers like Azure and AWS are overly reliant on their business. Meanwhile, no matter what they do, they will lose significant market share to competitors like Google and Anthropic in 2026, ARR will slow down significantly, and 2027 will likely be worse.
Between hollow promises of AGI and massive talent drain to startups, this is also the history of AI. One lawsuit after another. Sora's failure was embarrassing enough, but when the browser Atlas was taken down, most people didn't even notice and forgot what it was for. In the past eighteen months, Google has taken more B2C market share from OpenAI every month, and Anthropic has taken more B2B market share. When Meta's Musespark 1.1 gets more attention than the release of GPT-5.6, you know OpenAI is becoming like a relic of 2022. Back then, you really felt you had to listen to what Sam Altman said, or that ChatGPT was a real existing product. A few years later, we are clearer now.
AI is Pushing Older Employees Directly Out of the Labor Market
In the past year and a half, we have heard a lot about the impact of AI on new college graduates, entry-level positions, and career ladders in a tighter labor market. But what about the other end?
Generative AI may be accelerating age discrimination in the workplace, and I'm not sure if the media has reported enough on this.
Labor lawyers have noticed that more and more companies are using "lack of AI adaptability" as a seemingly neutral euphemism to lay off older employees.
More and more data shows that the relative increase in the number of older workers leaving jobs is higher, especially turning to unemployment status.
- (Source: Futurism) In a latest study at the Boston College Center for Retirement Research, Economics Professor Geoffrey Sanzenbacher carefully studied labor data to map the impact of AI on older workers in the US.
- A brief titled "Are Older Workers' Careers Being Shortened by AI?" released by the Boston College Center for Retirement Research in June 2026 was written by Geoffrey T. Sanzenbacher, exploring whether the rise of artificial intelligence is affecting the employment longevity of late-career workers.
- They concluded that overall, in jobs with higher AI exposure, there has been a relatively large increase in the turnover rate of older workers since the launch of ChatGPT.
- The significant increase in the number of departures threatens older workers in exposed positions; if they cannot find new jobs, their careers will be shortened.
I'm afraid generative AI is starting to squeeze the labor market from both ends: reduced entry-level hiring for young graduates + faster departure of professionals nearing retirement. This may erode capitalism and exacerbate the K-shaped US economy, where a serious affordability crisis is occurring.
Sanzenbacher observed that after the release of ChatGPT, the proportion of workers aged 55 and above leaving white-collar jobs surged.
In other words, more and more older white-collar workers are being fired or forced into early retirement.
The adverse effects of AI on older workers may far exceed what mainstream media acknowledges.
Generative AI is compressing, merging, and changing technical roles, and the rapid progress and growing complexity of AI in recent years have prompted education training providers and employers to rethink workplace roles, thinking about what constitutes being ready for work. But also think about what leaving the labor market means, when to retire, and how the cost efficiency paid for adopting AI reduces opportunities for workers in the last quarter of their careers.
I don't like the direction this might go: this late-career AI pressure is quite artificial.
Imagine, some challenges AI brings to older workers include: negative stereotypes about their willingness to learn new skills and comfort with technology, limited opportunities to obtain employer-provided or related training, and lack of digital access and basic digital skills—the basic ability to utilize technology, such as browsing the web, as pointed out by the Center for Retirement Research.
A labor market squeezed from both ends, just to be able to afford AI infrastructure and fast agentic AI, this does not sound good for society or the health of US consumers and workers.
AI and Older Workers
Similar to the patterns observed with entry-level labor, this seems to particularly affect older white-collar workers.
Workers over 55 have exposure to generative AI in their professions as high as middle-aged workers. Older workers tend to use AI less than younger people, and some view it more as a threat; this normal adoption gap exists.
There seems to be a key gap in research on how generative AI affects older workers.
Great Squeeze: The Double Dilemma of Overhyping Immature Technology on Demographics
I think Geoff is right, his bottom line is: when many policy experts urge extending working life, AI may be pushing some older workers in the opposite direction. If this tool is real, we need to ensure it promotes our well-being and financial health, not the opposite. As a large amount of talent leaves the labor market, some white-collar workers accept early retirement packages, and generative AI de-skills the rest of us, there will be real talent costs in some fairly critical industries.
Historically, turnover rates for workers in high-exposure jobs (usually white-collar and less physically demanding) were much lower, and career longevity was much longer than workers in low-exposure, physical jobs. There is more reason to believe this narrative has been completely reversed. Of course, we are only 3.5 years into this generative AI experiment, and in a tighter US labor market, there are many factors to consider.
Given the uncertain ROI of generative AI, token usage, and even agents, it is a pity to see many young and older workers struggling in a less meritocratic labor market, a market that succumbs to some unpleasant aspects of monopoly capitalism. This large-scale data center rollout and the listing of these AI companies may actually not be in the best interest of the US. This argument is starting to become more prominent, and the impact on real people shows disturbing signs.
Is K-shaped AI labor the future? Sounds extremely ageist.
"If entry-level hiring has slowed to a crawl, and professionals of retirement age are leaving in large numbers, how should ordinary workers navigate a labor market squeezed from both ends?" - Futurism
These are some random observations and stories I have been following.
Is an AI Sovereign Wealth Fund Feasible?
A June survey of 1,690 US adults by Verasight Research found that 69% of Americans now support "forcing" AI companies to transfer 50% of their stock to a public sovereign wealth fund, because the tech industry laid off over 140,000 people this year, and AI is considered the main reason for layoffs.
Read the press release
Of course this is related to Bernie Sanders. Senator Bernie Sanders proposed the US AI Sovereign Wealth Fund Act in June. The legislative proposal explicitly requires the largest US AI companies to give up 50% ownership to the public. Sanders positioned the bill as a safeguard against wealth concentration, believing that the economic upside of artificial intelligence cannot be limited to Silicon Valley boardrooms.
OpenAI offering 5% equity to the Trump government does not seem to be an appropriate response to this debate.
Are Europe and UK Dead in Tech and AI?
Not so. The venture capital boom experienced in the US in AI in recent years has also spread to Europe and the UK.
So far in 2026, UK startups have raised $17 billion.
These are the largest funding rounds so far:
- Isomorphic Labs – $2.1 billion
- Nscale – $2 billion
- Wayve – $1.2 billion
- Ineffable Intelligence – $1.1 billion
- Recursive – $650 million
- ElevenLabs – $500 million
- CuspAI – $400 million
- Oxford Quantum Circuits (OQC) – $350 million
- PhysicsX – $300 million
- CellCentric – $220 million
- OLIX – $220 million
- Fractile – $220 million
Although Europe doesn't have many winners in the AI field yet, it does have some interesting startups.
- Mistral (mostly Meta spin-off)
- Isomorphic Labs (Google spin-off)
- NEURA Robotics (Germany)
- Nscale (London, UK, new cloud provider, comparable to Crusoe, Nebius, etc.)
- Helsing (Germany, just raised $1.8 billion)
OpenAI Concentrates Power in Brockman's Hands
There is a very unusual financial relationship between Brockman and Sam Altman. After Fidji Simo officially stepped down, OpenAI President Greg Brockman will continue to oversee the company's product business.
Brockman will be responsible for OpenAI's:
- ChatGPT product business
- Business development team
- Enterprise team
- Compute resource projects
It is rare for one person to hold such great power. Evidence presented in court shows that as early as 2017, Sam Altman quietly gave Brockman a certain percentage of ownership in his personal family office (worth about $10 million at the time). Now the value of this money is far beyond the original.
Brockman disclosed that he holds personal shares in several important startups strongly supported by Altman, including nuclear fusion energy company Helion Energy and AI chip manufacturer Cerebras.
The disclosure about Cerebras raised concerns because during the period when OpenAI was negotiating billions of dollars contracts to purchase its hardware, Brockman was holding shares in the chip manufacturer.
In addition, Brockman's equity in OpenAI's business division is now worth nearly $30 billion, which forms a core argument: Brockman is economically motivated to support Altman in shifting OpenAI from its original non-profit charter to commercialization. These and other economic incentives mean Brockman is essentially bound to Sam Altman's leadership and business tactics (circular financing and private deals). OpenAI's core operations are basically full of conflicts of interest.
Oracle Faces Significant Risks Due to OpenAI
S&P Global Ratings downgraded Oracle's long-term credit rating from BBB to BBB- (one step away from junk status), and pointed out that OpenAI is a key credit risk. The stock has fallen 28% so far in 2026. $ORCL
One more downgrade, and Oracle will become a junk-grade company for the first time. S&P listed OpenAI as a "key credit risk" in the report.
OpenAI accounts for about half of Oracle's remaining performance obligations. This is that famous $638 billion backlog. This is extremely dangerous for Oracle: because unlike hyperscale cloud providers like Microsoft, AWS, or Google, Oracle lacks large-scale own internal AI workloads or extensive consumer platforms to absorb excess infrastructure capacity. If OpenAI slows down consumption, shifts direction, or faces financial difficulties, Oracle will be highly exposed to risk.
OpenAI is a private company. It has no credit rating and has never been profitable. It may never be profitable, or only achieve profitability in 2029 or later. And a rating agency has just written the risk of this company's failure directly into the credit profile of a public company held by global pension funds and bond funds. This is no longer speculation.
But what is the real driver of the downgrade? The downgrade is because Oracle increased its capital expenditure for fiscal year 2027 from an expected $60 billion to a staggering $90 billion to $95 billion, expanding its free cash flow deficit to an expected negative $42 billion.
If OpenAI does not continue to grow or even slows down, Oracle and SoftBank will be in trouble.
Oracle's data center leases are for 15 to 19 years. Its cloud customer contracts are about 5 years. Oracle disclosed these two numbers in its own annual report. So Oracle signed a 19-year obligation to serve a 5-year commitment made by a customer that has never profited a dollar. This is quite unreliable. OpenAI's cash burn situation is echoing in other institutional pipelines. All the lawsuits don't help. All other behaviors evading rule of law and conflicts of interest don't help either.
Actually, if OpenAI cannot pay, Oracle will be left holding data center leases that may not be exitable, and may have to sublease to others on worse terms. By then, compute demand and dynamics may have changed. Oracle and SoftBank have assumed absurd levels of risk to support OpenAI. Oracle spent about $55.7 billion on capital projects in fiscal year 2026, with negative free cash flow of $23.7 billion.
S&P now expects the cash flow deficit for fiscal year 2027 to expand to about negative $42 billion.
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
