A rogue OpenAI research agent bypassed blocks on the Australian government health-data portal, accessing non-public files in June.
Rongchai Wang
Jun 26, 2026 22:30
A report said Apple hardware VP Paul Meade, tied to Vision Pro and smart-glasses work, is expected to join OpenAI amid a leadership shakeup under new CEO John Ternus.
A report that Apple is losing a key leader tied to Vision Pro and smart-glasses hardware to OpenAI is feeding fresh attention on the AI talent race as Polymarket traders price the July model-leader outcome. In the Polymarket market “Which company has best AI model end of July?”, the top line still favors Anthropic at 85.5%.
A report said Apple is losing Paul Meade, a hardware vice president who led work tied to the Vision Pro headset and projects related to smart glasses, as the company faces shifting plans for wearables. The report described the Vision Pro as having become suddenly more expensive, while smart-glasses plans were said to be in the works. Meade is expected to join OpenAI and reunite with former Apple designers, including Jony Ive, Tang Tan, and Evans Hankey. The report also said OpenAI purchased the designers’ startup last year. The personnel move was framed as part of an executive shakeup under Apple’s new CEO, John Ternus.
On Polymarket’s multi-outcome board for “Which company has best AI model end of July?”, positioning remains heavily concentrated in Anthropic at 85.5% Yes / 14.5% No. Google is priced at 10.95% Yes / 89.05% No, while OpenAI trades at 3.3% Yes / 96.7% No, indicating a steep drop-off after the leader. Smaller outcomes sit near option-like tails, including xAI at 0.25% Yes / 99.75% No and several names at 0.15% Yes / 99.85% No. Matched volume stands at $2,133,993, and the leader’s edge implies traders see a wide gap rather than a tight three-way race.
Whether the implied lead compresses as July approaches will show up first in a rising Google or OpenAI Yes price and a falling Anthropic Yes price ahead of the July 31, 2026 resolution date.
Away from the model-leader board, Polymarket activity is also clustering in a mix of deadline-driven tech and geopolitical risk contracts. In “Which company has best AI model end of June?”, Anthropic leads at 99.05% on $21,944,398 in volume, while “GPT-5.6 released by…?” prices a July 31 outcome at 86.55% with $1,708,896 matched. On the geopolitics side, traders put “US forces enter Venezuela again by…?” at 95.65% for June 30 on $1,575,549, even as “Saudi Arabia bans US military aircraft by…?” sits at 0.2% for June 30 with $783,677 traded.
| Window | Change (pp) |
|---|---|
| 24h | -2.5 |
| 7d | -2.5 |
Top strike rungs
| Strike | Yes | No |
|---|---|---|
| Anthropic | 85.5% | 14.5% |
| 10.9% | 89.0% | |
| OpenAI | 3.3% | 96.7% |
| xAI | 0.2% | 99.8% |
+11 more strikes not shown
View market on platform
Image source: Shutterstock
Alvin Lang
Jun 17, 2026 00:04
On Monday, Anthropic staff met senior Trump administration officials after a Friday export restriction forced it to pull its newest model for non-U.S. users.
Trump administration officials met with Anthropic after a federally imposed export restriction prompted the startup to pull its latest model from the market, sharpening focus on the company’s near-term standing in the AI race. On Polymarket, traders nudged up the odds that Anthropic will be judged to have the best AI model by the end of June 2026.
Anthropic staff met Monday with senior Trump administration officials in their first in-person sessions since a federal export restriction pushed the company to pull its newest model from the market on Friday night. A senior White House official said it could take longer than a few days to reach an outcome that eases the Friday action, which barred Anthropic from allowing non-U.S. users to access its newest model due to potential security vulnerabilities, while also leaving open the possibility of a quicker resolution. The discussions followed hours-long weekend calls involving Anthropic co-founder Tom Brown, Commerce Secretary Howard Lutnick and National Cyber Director Sean Cairncross, with Anthropic’s public policy lead Sarah Heck also on those calls. Monday’s meetings, held by Commerce and Cairncross’ office, were described as more technical and led by staff, including Chris Fall, who heads Commerce’s Center for AI Standards and Innovation. Anthropic presented its cybersecurity safeguards and sent senior personnel including Frontier Red Team member Logan Graham, safeguards head Dave Orr and lead security researcher Nicholas Carlini, according to a person close to the company.
In Polymarket’s “Which company has best AI model end of June?” market, Anthropic led at 91.35% Yes versus 8.65% No, up from 90.75% in the prior reading, on $15,341,099 in volume. Google was priced at 5.0% Yes / 95.0% No, while OpenAI traded at 2.1% Yes / 97.9% No. Longer-tail outcomes were far below 1%, including xAI at 0.25% Yes / 99.75% No and Meta at 0.15% Yes / 99.85% No, underscoring a heavily one-sided market consensus.
Traders will be watching for any shift in pricing across the top three outcomes—Anthropic, Google, and OpenAI—alongside changes in volume and whether the gap between Anthropic and the field narrows ahead of the 2026-06-30 resolution date.
Beyond the AI leaderboards, Polymarket’s most-trafficked boards show traders concentrating risk on geopolitics and elections: “What Iranian demands will Trump agree to by June 30?” has Oil Sanction Relief at 98.8% on $3,953,820 in volume, while “Peru Election 2nd Round: Margin of Victory? (0.1% brackets)” prices Fujimori 0.2–0.3% at 94.95% on $2,266,692. In the U.K., “Will Rebecca Shepherd win the 2026 Makerfield by-election?” has No at 99.05% with $4,176,100 traded, and sports money remains active too, with “World Cup: Team to advance to Knockout Stages” putting Germany at 99.5% on $3,600,839.
| Window | Change (pp) |
|---|---|
| 24h | +11.6 |
| 7d | +11.6 |
Top strike rungs
| Strike | Yes | No |
|---|---|---|
| Anthropic | 91.3% | 8.7% |
| 5.0% | 95.0% | |
| OpenAI | 2.1% | 97.9% |
| xAI | 0.2% | 99.8% |
+11 more strikes not shown
View market on platform
Image source: Shutterstock
A sweeping new poll by Anthropic—the Claude maker that recently filed to go public—finds that Americans hold strikingly contradictory views on artificial intelligence: They’re eager for its potential to cure disease and improve lives, yet deeply anxious about the economic and social disruption it may bring, and profoundly skeptical of the companies building it.
The survey, which Anthropic is calling the “Public Record,” gathered responses from nearly 52,000 Americans in late 2025, making it one of the largest national polls on AI attitudes to date.
Job loss emerged as the dominant fear, with 64% of respondents expressing worry about AI-driven displacement—a concern that was the top-ranked fear among both Democrats and Republicans, and in every state surveyed. Notably, the anxiety was higher among more educated Americans, whose work overlaps more closely with what AI is increasingly being asked to do.
On the hopeful side, nearly half of Americans said curing diseases like cancer or Alzheimer’s ranked among their top three wishes for AI, followed by helping people with disabilities at 36%. Notably, hopes that AI might substitute for human connection—such as offering therapy or reducing loneliness—ranked lowest among the options presented.
Despite that optimism, trust in the industry itself is remarkably thin. Only 15% of respondents said they trust AI companies to make decisions about how the technology is developed and used—lower than the federal government, state and local government, and international bodies, and far below independent experts at 43%.
The public’s appetite for government oversight is broad and bipartisan. Over 70% of those surveyed said the government should play a role in regulating AI, with support running from 79% among Democrats to 68% among Republicans. Americans were most eager for government action on privacy, child safety, and corporate liability for harm.
Anthropic said it plans to repeat the survey regularly and expand it beyond the United States.
Start every day with the top news stories right now, plus original features, a podcast, videos and more.
CoreWeave just scored big. The cloud infrastructure company signed a multi-year deal with Anthropic to handle their AI workloads, and it’s pretty much cementing CoreWeave’s spot as the go-to provider for serious AI developers.
The partnership puts CoreWeave in an elite club – they now support nine out of ten major language model developers worldwide. That’s a wild achievement for a company that wasn’t even on most people’s radar just a few years ago. CEO Michael Intrator said the deal gives Anthropic access to CoreWeave’s powerful GPU clusters and optimized cloud solutions, which are basically essential for training those massive AI models that everyone’s talking about these days.
The agreement got finalized on April 10. Anthropic, the AI safety company founded by former OpenAI researchers, needed serious computational firepower for their model development work. CoreWeave delivered exactly that – specialized infrastructure that can handle the crazy demands of large-scale AI training and deployment.
But here’s the thing: neither company wants to talk money. The financial terms remain under wraps, and they’re not saying how long the contract runs either. Industry analysts are basically guessing at this point about what kind of revenue boost CoreWeave might see. Some estimate it could be worth hundreds of millions, but that’s just speculation.
Anthropic’s CEO Dario Amodei did say the partnership helps them build “safe and reliable AI systems,” which aligns with their whole mission around AI safety. For CoreWeave, it’s another validation that their infrastructure approach works for the biggest players in the game.
The timing makes sense too. AI companies are burning through compute resources like crazy, and building your own data centers takes forever. So they’re turning to specialists like CoreWeave who already have the hardware and know-how.
Not bad for a company that started as a cryptocurrency mining operation. CoreWeave pivoted hard into AI infrastructure and it’s paying off big time.
The Anthropic deal follows another major partnership CoreWeave announced in March 2026 with OpenAI. That earlier agreement showed CoreWeave could handle the technical demands of cutting-edge AI development. Now they’re proving it wasn’t just a one-off success.
According to Gartner, the global AI cloud services market should hit $13 billion by 2027. CoreWeave seems positioned to grab a decent chunk of that growth, especially with these high-profile partnerships under their belt. Analysts have drawn connections to XRP Eyes Target as Technical amid evolving conditions.
Their recent $200 million funding round suggests investors believe the hype. The money gives CoreWeave resources to keep scaling their operations and adding more GPU capacity. CTO Brian Venturo said they’re planning to enhance their GPU offerings to stay ahead of future AI demands.
Things are moving fast in this space. CoreWeave went from mining crypto to supporting the world’s most advanced AI companies in just a few years. That’s the kind of pivot that either makes you or breaks you.
The partnership raises questions about CoreWeave’s capacity planning. They’re not saying how much additional infrastructure they’ll need to support Anthropic’s workloads, or whether they’re planning major facility expansions.
Intrator hinted at more partnerships coming during a press conference on April 11, but didn’t name names. Given their track record with OpenAI and now Anthropic, other AI companies are probably taking meetings.
The lack of financial disclosure isn’t stopping investor interest. Tech sector shares saw a small bump after the announcement, and analysts think CoreWeave might be gearing up for another funding round soon.
Anthropic’s focus on AI safety could influence how CoreWeave operates too. The company is known for being extra careful about security and ethical considerations, which might set new standards for cloud infrastructure providers across the industry. Analysts have drawn connections to Bitcoin Needs Volume Surge to Hit amid evolving conditions.
For now, CoreWeave keeps adding major clients while competitors scramble to catch up. The company didn’t respond to requests for comment about their expansion plans or hiring targets to support the increased demand.
The broader cloud infrastructure landscape is experiencing massive shifts as AI demand explodes. Amazon Web Services and Microsoft Azure still dominate overall cloud services, but specialized providers like CoreWeave are carving out profitable niches. Google Cloud recently reported spending $13 billion on data center infrastructure in 2023 alone, mostly driven by AI workload requirements. Meanwhile, NVIDIA’s H100 chips – the gold standard for AI training – remain in such short supply that lead times stretch six months or more.
CoreWeave’s advantage lies in their GPU-first architecture, something traditional cloud providers are scrambling to match. Meta announced plans to acquire 350,000 H100 chips by year-end, while Tesla’s Dojo supercomputer project shows how tech giants are building custom solutions. But smaller AI companies can’t afford that route. Stability AI, Cohere, and Inflection AI have all turned to specialized infrastructure providers rather than building their own facilities. Industry sources suggest CoreWeave’s waiting list for premium GPU access now extends several months, creating a supply crunch that’s driving up prices across the sector.
Post Views: 1
CoreWeave signed a multi-year agreement to provide AI infrastructure services for Anthropic’s model development and deployment needs, making them the infrastructure provider for nine of the ten largest AI developers.
Neither company disclosed financial terms or contract duration, leaving analysts to speculate about the deal’s value and scope.