Two weeks of artificial intelligence news rarely arrive with this much consequence packed in. Between June 7 and June 20, 2026, the United States government forced Anthropic to pull its two most capable models offline worldwide, OpenAI bought a German cloud startup to keep its coding agents running around the clock, and three rival AI lab chief executives sat at the same G7 table in France for the first time in history. Layered underneath those headlines sat a steady churn of funding rounds, a fresh cybersecurity exploit hitting half the coding-agent market, and a healthcare research milestone that quietly closed gaps doctors had missed for years.
What makes this stretch notable is not any single announcement but how the threads connect. Export controls on frontier models, a scramble for persistent agent infrastructure, and an enterprise sales race through Seoul all point to an industry where geopolitics, infrastructure, and commercial deployment are colliding in real time. Investors, developers, and policy watchers each have a reason to pay close attention to what happened over the past two weeks.
The following roundup walks through the most significant, independently verifiable AI stories from this window, drawn from official company newsrooms, congressional and regulatory filings, and reporting from established outlets including CNBC, Bloomberg, and The Hacker News.
Key headlines from June 7 to 20, 2026:
- US export control directive forces Anthropic to suspend Claude Fable 5 and Claude Mythos 5 worldwide
- OpenAI acquires cloud execution startup Ona to power long-running Codex agents
- Gemini 3.5 Pro remains unshipped past its promised June launch window, frustrating developers
- Sam Altman, Dario Amodei, and Demis Hassabis appear together at the G7 summit in France
- Anthropic opens its Seoul office and signs a wave of Korean enterprise partnerships
- “Agentjacking” exploit compromises Claude Code, Cursor, and Codex through a Sentry MCP flaw
- Sixfold raises a $30 million Series B for its AI insurance underwriting platform
- OpenAI’s GPT-Rosalind model assists in 18 new diagnoses of previously unsolved rare disease cases
- US tech layoffs hit roughly 1,100 jobs per day in 2026, with AI cited as a recurring rationale
- OpenAI and Molecule.one demonstrate a near-autonomous AI chemist improving drug-relevant chemistry
- AlphaSense closes a $350 million round for AI-driven market intelligence software
US Export Control Directive Forces Anthropic to Pull Fable 5 and Mythos 5 Worldwide
On the evening of June 12, 2026, Anthropic received a letter from the US government, citing national security authorities, ordering it to suspend all access to Claude Fable 5 and Claude Mythos 5 for any foreign national, whether inside or outside the United States, including the company’s own foreign national employees. Anthropic received the directive at 5:21 pm Eastern Time, and the letter did not provide specific details of its national security concern. Because Anthropic has no practical way to distinguish foreign nationals from domestic users across cloud platforms in real time, it disabled both models entirely for every customer within hours. Reporting from CNBC confirmed the order came directly from Commerce Secretary Howard Lutnick to Anthropic chief executive Dario Amodei.
The two models had only launched days earlier, on June 9, making this the first known instance of a US government forcing a complete worldwide shutdown of a publicly deployed frontier AI model. Anthropic’s understanding was that the government had become aware of a method of bypassing, or jailbreaking, Fable 5, and the company reviewed a demonstration of the technique tied to previously known, minor vulnerabilities. Crucially, every other Claude model, including Opus 4.8, remained fully available throughout, limiting the operational blast radius even as it raised the legal one.
The episode lands at the center of an emerging debate over what industry analysts now call hardware and model sovereignty, the idea that organizations should not build mission-critical workflows entirely on top of cloud-hosted models that a government can recall without warning. Enterprise security teams in finance, healthcare, and critical infrastructure are already discussing multi-provider routing and self-hosted open-weight fallbacks as insurance against a repeat episode. For OpenAI and Google, the suspension is a quiet opening to court enterprise customers nervous about putting all their eggs in one regulatory basket, even as both companies brace for the precedent to eventually reach their own frontier releases.
Source: Anthropic | https://www.anthropic.com/news/fable-mythos-access
OpenAI Acquires Ona to Give Codex Agents a Persistent Cloud Home
OpenAI announced on June 11, 2026, that it would acquire Ona, the cloud execution and orchestration startup formerly known as Gitpod, folding its technology directly into the Codex coding agent ecosystem. The deal brings Ona’s secure cloud execution and orchestration technology into OpenAI’s rapidly expanding Codex ecosystem, giving agents a sandboxed environment that survives after a developer closes their laptop. More than 5 million people now use Codex each week to research, analyze, build, and automate their work, a 400 percent increase from earlier this year, and OpenAI said the most valuable Codex tasks increasingly unfold over hours or days rather than minutes.
Financial terms were not disclosed, and the transaction remains subject to regulatory approval, according to Bloomberg’s initial report. Ona, founded in Kiel, Germany, in 2020, says it has helped roughly two million developers work inside secure, reproducible cloud environments. Founder and chief executive Johannes Landgraf will join OpenAI’s Codex group along with his full team once the deal closes, writing on LinkedIn that the sale felt less like an ending and more like the company’s mission getting bigger.
The acquisition is the latest in a string of OpenAI purchases aimed at owning the deployment layer beneath its models, following the cybersecurity startup Promptfoo in March and health tech company Torch for roughly 60 million dollars in January. The timing is not incidental. Anthropic’s Claude Code has been a significant driver of its enterprise growth, and both companies have filed confidential paperwork for a public listing. Winning the coding agent category increasingly depends not just on model quality but on whether an agent can run unattended in a compliant, auditable cloud environment, exactly the gap Ona was built to close.
Source: OpenAI | https://openai.com/index/openai-to-acquire-ona/
Gemini 3.5 Pro Misses Its June Launch Window, Testing Developer Patience
Google’s most anticipated model of the year still has not shipped to the public as of June 19, 2026, despite chief executive Sundar Pichai’s promise at Google I/O on May 19 that Gemini 3.5 Pro would arrive “next month.” Gemini 3.5 Pro, unveiled at Google I/O on May 19, was slated for general availability in June 2026, a launch Google teased but had not yet delivered as of early June. Pichai’s exact framing on stage drew an audible groan from developers expecting the launch that day, and that reaction has aged into a running joke across developer forums as the month winds down.
Confirmed specifications include a 2-million token context window and a Deep Think reasoning mode, with expected pricing in the range of 15 dollars per million input tokens and 60 dollars per million output tokens, positioning it close to Claude Opus 4.7 and below GPT-5.5 Pro. The model has been available only through a limited Vertex AI enterprise preview, with no access yet through the consumer Gemini app or public Google AI Studio. By contrast, Gemini 3.5 Flash, the faster and cheaper sibling model, shipped on schedule at I/O and reportedly outperforms the prior generation’s Pro tier on coding and agentic benchmarks while running roughly four times faster than competing frontier models.
The delay carries real competitive weight. Developers who built switching plans around a June Gemini 3.5 Pro release are now stuck holding off decisions or defaulting to Claude or GPT-5.5 for top-tier reasoning workloads, a dynamic that benefits Anthropic and OpenAI by default rather than design. Prediction markets tracking the release, including Polymarket, have shifted their odds toward a launch in the final ten days of June, and Google has yet to issue a formal update explaining the slip.
Source: TechTimes | https://www.techtimes.com/articles/317919/20260606/google-gemini-35-pro-nears-june-launch-2-million-token-context-deep-think-reasoning.htm
Three Rival AI Lab CEOs Share a Table at the G7 Summit in France
Sam Altman of OpenAI, Dario Amodei of Anthropic, and Demis Hassabis of Google DeepMind all attended the G7 summit in Evian-les-Bains, France, the week of June 15 through 17, 2026, marking the first time the three rival frontier lab chief executives appeared before world leaders together. CEOs, including OpenAI’s Sam Altman, Anthropic’s Dario Amodei, and Google DeepMind’s Demis Hassabis, alongside around a dozen other tech leaders, took part in a working lunch meeting at the summit on June 17, seated alongside President Trump and other heads of state.
Analysts framed the joint appearance as a signal of how much geopolitical weight frontier AI companies now carry. Commentators told CNBC that the companies’ seats at the table reflect the growing geopolitical influence of tech bosses, with frontier AI risk, infrastructure buildout, and sovereignty all on the formal agenda. The summit took place just days after the Fable 5 and Mythos 5 export suspension, giving Amodei’s presence an especially pointed backdrop. Brookings Institution visiting fellow Cameron Kerry told CNBC that the release of capable cyber-focused models, including Mythos and GPT-5.5 Cyber, marked an inflection point in AI development that has unsettled businesses and governments over digital security exposure.
The convergence of lab leadership and head-of-state diplomacy underscores a shift that started quietly over the past two years and is now explicit: frontier AI policy is being negotiated directly between company founders and national leaders, not solely through traditional regulatory agencies. For competing labs without a seat at that table, including well-funded challengers like Mistral, xAI, and Chinese frontier players, the optics of exclusion may matter as much as any specific policy outcome discussed in the room.
Source: CNBC | https://www.cnbc.com/2026/06/17/g7-trump-ai-tech-leaders-openai-anthropic-google.html
Anthropic Opens Seoul Office Amid Its Own Export Control Storm
Anthropic formally opened its Seoul office on June 17, 2026, its third location in the Asia-Pacific region after Tokyo and Bengaluru, alongside a sweeping set of enterprise and research partnerships across Korea’s AI ecosystem. KiYoung Choi, former General Manager of Snowflake Korea, leads the office as Representative Director. New enterprise deployments include NAVER rolling out Claude Code across its entire engineering organization, Samsung SDS deploying Claude Cowork and Claude Code across Samsung Electronics, LG CNS bringing Claude to LG Group, Nexon using Claude Code for live-service game development, and Hanwha Solutions running Claude through AWS Bedrock with in-region data controls.
The launch carried an awkward undertone given its timing. The office opening coincided directly with the US export controls that barred foreign nationals from accessing Fable 5 and Mythos 5, and Anthropic confirmed that access work for South Korean companies and institutions linked to a cybersecurity project, including the Korea Internet and Security Agency, Samsung Electronics, SK Hynix, and SK Telecom, had been halted as a direct consequence. Anthropic’s head of international, Chris Chaoui, told reporters at the briefing that he expected the export restrictions to be resolved within days, though as of this writing, they remain in force.
Korean enterprise revenue tied to Claude has been climbing fast. Company officials at the briefing put Anthropic’s overall revenue at 47 billion dollars as of mid-June, up from 9 billion dollars at the end of 2025, an extraordinary trajectory even by 2026’s inflated AI growth standards. The Seoul push, layered directly on top of an active export dispute touching Korean partners, illustrates a structural tension every US frontier lab will increasingly face: international enterprise expansion now runs squarely through Washington’s evolving national security calculus, not just local sales execution.
Source: Anthropic | https://www.anthropic.com/news/seoul-office-partnerships-korean-ai-ecosystem
“Agentjacking” Exploit Hijacks Claude Code, Cursor, and Codex Through a Single Fake Bug Report
Security researchers at Tenet Security disclosed a novel attack class called Agentjacking on June 9, 2026, demonstrating how a single forged Sentry error report can hijack popular AI coding agents into executing attacker-controlled commands. During a validation period that ended on June 17, 2026, Tenet researchers identified 2,388 organizations with exposed Sentry DSNs, and the technique worked in tested environments using popular AI coding tools, including Claude Code, Cursor, and OpenAI Codex, across Windows, macOS, and automated cloud pipelines.
The mechanics are unsettling in their simplicity. An attacker needs only a project’s public, write-only Sentry DSN credential, discoverable from browser JavaScript or a basic GitHub code search, to post a crafted fake error event. When a developer later asks their coding agent to fix unresolved Sentry issues, the agent retrieves the injected event through the Model Context Protocol and treats the embedded instructions as legitimate diagnostic guidance rather than untrusted external content. The impact is severe, with environment variables including AWS keys, GitHub tokens, Sentry auth tokens, git credentials, private repository URLs, and developer identity silently exfiltrated to the attacker’s server, and controlled validation waves saw over 100 organizations have AI coding agents act on injected errors, yielding an 85 percent exploitation success rate.
What makes Agentjacking distinct from conventional supply chain attacks is that it bypasses essentially every standard security control. The technique bypasses EDR, WAF, IAM controls, VPN, Cloudflare, and firewalls entirely because every action in the attack chain is technically authorized, since the developer genuinely asked their own agent to investigate a real-looking error. Sentry has acknowledged the report and added filtering for the specific proof-of-concept payload, but has characterized the deeper architectural issue as not fully fixable at the ingestion layer, pointing instead to model-side middleware as the correct long-term mitigation. For any organization running MCP-connected coding agents in production, this represents the clearest evidence yet that prompt injection has moved from academic concern to an active, scaled exploitation vector with real credential theft attached.
Source: The Hacker News | https://thehackernews.com/2026/06/agentjacking-attack-tricks-ai-coding.html
Sixfold Closes $30 Million Series B to Scale AI Underwriting
Insurance technology startup Sixfold raised a 30 million dollar Series B round on June 17, 2026, to expand its generative AI platform for end-to-end insurance risk analysis. Sixfold has unveiled its AI Underwriter, a tool for evaluating insurance submissions, with the product launch backed by the new funding round aimed at building this platform further. The underlying technology was developed over three years and has been tested by major insurance carriers before this funding announcement.
The round lands at a moment when AI-native vertical software is increasingly outcompeting general-purpose model wrappers for enterprise dollars. Insurance underwriting, with its dense, document-heavy workflows and high tolerance for structured automation, has become one of the more durable beachheads for applied generative AI outside of coding and customer support. Sixfold’s pitch centers on streamlining a process that traditionally consumes days of manual analyst time per commercial policy submission, replacing it with AI-driven document synthesis and risk scoring.
The deal also reflects a broader funding pattern visible across this two-week window. Late-stage AI infrastructure megarounds, from Anthropic’s reported 65 billion dollar raise to SpaceX’s record-breaking IPO, continue to dominate headline figures, but mid-market Series B rounds focused on a specific, defensible vertical use case are quietly proving that AI venture appetite has not narrowed to only foundation model labs. For insurance incumbents watching from the sidelines, Sixfold’s traction with major carriers signals that AI underwriting tools are moving past pilot programs into genuine production deployment.
Source: Fundz | https://app.fundz.net/fundings/sixfold-funding-round-series-b-5a77a7
OpenAI’s GPT-Rosalind Helps Diagnose 18 Rare Disease Cases That Stumped Specialists for Years
A study published June 18, 2026, in NEJM AI demonstrated how an OpenAI reasoning model helped researchers identify diagnostic leads for rare genetic conditions that had evaded years of expert review. Following expert review, additional testing, and clinical confirmation, physicians established diagnoses in 18 cases, an additional diagnostic yield of 4.8 percent after earlier analysis by specialists. In the study, OpenAI o3 Deep Research helped researchers identify leads that were later assessed through established clinical processes, with the model surfacing evidence-linked candidate explanations rather than making any diagnosis itself.
The scale of the underlying problem is significant. Even with genomic sequencing, many people with rare diseases never receive a clear genetic diagnosis, and roughly half remain undiagnosed after extensive testing and specialist review. Researchers reanalyzed 376 previously unsolved cases, with the model producing evidence-linked hypotheses for specialists to review and, where appropriate, investigate through additional testing and confirm in a clinical laboratory rather than acting autonomously on patients.
OpenAI paired this announcement with an updated GPT-Rosalind model released June 18, specifically tuned for life sciences research. GPT-Rosalind achieves industry-leading performance in medicinal chemistry, evaluating multimodal chemical structure understanding, structure-activity relationship prediction, drug potency and toxicity prediction, and retrosynthesis, outperforming GPT-5.5 on the company’s internal MedChemBench while using 7.2 percent fewer tokens. The dual announcement positions OpenAI squarely against Google DeepMind’s AlphaFold legacy and Microsoft’s healthcare AI investments in the race to own the clinical and translational research layer of AI, a market where regulatory caution and clinical validation timelines, not raw model capability, increasingly determine who wins enterprise trust.
Source: OpenAI | https://openai.com/index/diagnose-rare-childhood-diseases/
OpenAI and Molecule.one Demonstrate a Near-Autonomous AI Chemist
OpenAI published research on June 18, 2026, showing how an AI system built on GPT-5.4, developed in partnership with Polish drug discovery startup Molecule.one, improved the yield of a historically difficult pharmaceutical chemistry reaction with minimal human intervention. OpenAI and Molecule.one show how a near-autonomous AI chemist using GPT-5.4 improved a key drug-making reaction, advancing medicinal chemistry research. The specific chemistry involved Chan-Lam coupling, a method chemists use to build pharmaceutically relevant molecules, where one difficult version involving primary sulfonamides has historically produced low yields, limiting how useful it can be in medicinal chemistry.
The research arrives alongside OpenAI’s introduction of LifeSciBench, a new expert-authored evaluation benchmark designed to test how AI systems perform on realistic life sciences research tasks rather than narrow recall questions. The benchmark distinguishes between models that can merely retrieve known chemistry facts and those capable of reasoning through unfamiliar experimental scenarios using evidence presented in the moment, a distinction increasingly relevant as pharmaceutical companies evaluate which AI tools can be trusted inside actual wet lab workflows.
For the drug discovery industry, the practical significance lies less in any single reaction yield improvement and more in the demonstration itself. A frontier lab showing a model that can propose, reason about, and meaningfully improve a real synthetic chemistry pathway, validated by an external partner with actual laboratory infrastructure, moves the conversation about AI-driven drug discovery from theoretical to operational. Expect competitive responses from Google DeepMind’s Isomorphic Labs and well-funded AI chemistry startups eager to demonstrate similar autonomous experimental capability before OpenAI cements first-mover credibility in this niche.
Source: OpenAI | https://openai.com/index/accelerating-science-gpt-5/
US Tech Layoffs Average Over 1,100 Jobs Per Day, With AI Cited as the Default Explanation
Tech industry layoffs accelerated sharply through the first half of June 2026, with companies increasingly citing artificial intelligence as the rationale, even as independent data casts doubt on how much of the cutting is genuinely AI-driven. As of June 14, 2026, 247 layoff events had displaced 183,966 workers across the tech, finance, and healthcare sectors, an average of 1,115 jobs lost every working day, nearly double the 564 per day pace recorded in 2025. The month of May alone accounted for roughly 40,000 cuts, the highest single month in two years.
Notable individual cuts during this window included ServiceNow, which laid off hundreds of employees on June 12 as the enterprise software company continues to increase its use of artificial intelligence, and Salesforce, which laid off 86 employees across its Mulesoft and Marketing Cloud teams on June 13. Yet survey data complicates the narrative that AI is the primary driver. A Gallup dataset released this week found that when laid-off workers were asked, in their own words, why they lost their jobs, only 1 percent mentioned AI or automation as the primary reason, with organizational restructuring, budget cutting, and economic conditions cited far more often.
The gap between corporate messaging and worker-reported experience matters for how both investors and policymakers read the labor data ahead. A May 2026 Gartner study cited in reporting found that companies cutting the most jobs showed no measurable improvement in financial returns, suggesting AI is functioning as a convenient cover for layoffs that organizations would likely have made regardless. California has already begun responding at the policy level, with Governor Newsom’s executive order directing a review of worker displacement protections, a model other states are likely to study closely as the “AI washing” debate intensifies through the second half of 2026.
Source: TechTimes | https://www.techtimes.com/articles/318466/20260616/tech-layoffs-hit-1115-day-2026-companies-cite-ai-cuts-fail-boost-returns.htm
AlphaSense Closes $350 Million Round for AI Market Intelligence
AlphaSense, an AI-enabled market intelligence and workflow orchestration platform, closed a 350 million dollar funding round in the first week of June 2026, valuing the New York-based company at 7.5 billion dollars. The round was led by Vitruvian Partners, Accenture Ventures, J.P. Morgan Asset Management, D.E. Shaw Ventures, and Pinegrove Opportunity Partners, a roster that signals strong institutional confidence in AI-native research and due diligence tooling among financial services buyers.
AlphaSense has carved out a niche serving investment banks, hedge funds, and corporate strategy teams that need to synthesize earnings calls, regulatory filings, expert network transcripts, and broker research at speed. The platform’s AI layer searches and summarizes across these fragmented document types in ways that traditional keyword-based financial research tools could not match, a use case that has proven durable even as general-purpose chatbots commoditize simpler research tasks.
The participation of strategic investors, including Accenture Ventures and J.P. Morgan Asset Management, rather than purely financial sponsors, suggests AlphaSense is positioning for deeper enterprise distribution partnerships rather than a near-term exit. As frontier labs push further into vertical-specific tools of their own, including OpenAI’s enterprise financial offerings and Anthropic’s banking partnerships with firms like TCS and DXC, independent platforms like AlphaSense face growing pressure to demonstrate that domain-specific data integration and workflow depth can outcompete general-purpose model access alone.
Source: Crunchbase News | https://news.crunchbase.com/venture/biggest-funding-rounds-june-5-2026/
Quick Comparison: Major AI Developments, June 7 to 20, 2026
| Date | Company | Development | Category |
|---|---|---|---|
| June 9 | Tenet Security | Agentjacking exploit disclosed against Claude Code, Cursor, Codex | Security/Research |
| June 11 | OpenAI | Acquires Ona to power persistent Codex cloud agents | Funding/Acquisition |
| June 12 | Anthropic | Forced to suspend Fable 5 and Mythos 5 worldwide via export directive | Regulation/Policy |
| June 17 | Sixfold | Raises $30M Series B for AI insurance underwriting | Funding |
| June 17 | Anthropic | Opens Seoul office, signs Korean enterprise partnerships | Product/Expansion |
| June 17 | G7 Summit | Altman, Amodei, and Hassabis appear jointly before world leaders | Policy |
| June 18 | OpenAI | GPT-Rosalind aids 18 new rare disease diagnoses (NEJM AI) | Healthcare AI |
| June 18 | OpenAI/Molecule.one | Near-autonomous AI chemist improves drug-relevant reaction | Research |
| Early June | AlphaSense | Closes $350M round at $7.5B valuation | Funding |
Closing Section
The two weeks spanning June 7 to 20, 2026, made clear that AI’s center of gravity has shifted from pure model releases to the messier business of deployment, governance, and trust.
Anthropic’s forced shutdown of Fable 5 and Mythos 5 showed how quickly national security policy can override commercial roadmaps, while the Agentjacking exploit proved that the agentic coding tools enterprises now depend on remain dangerously exposed to basic prompt injection.
Watch for Gemini 3.5 Pro’s eventual launch, the resolution of Anthropic’s Korean export dispute, and whether Agentjacking forces a broader rethink of how AI agents validate the tools they trust, as the stories most likely to carry directly into the final week of June.
1. Why did Anthropic shut down Fable 5 and Mythos 5?
The US government issued an export control directive on June 12, 2026, citing national security authorities, requiring Anthropic to suspend access to both models for foreign nationals worldwide. Because Anthropic could not filter users by nationality in real time, it disabled both models for all customers globally, while every other Claude model remained available.
2. Is Gemini 3.5 Pro released yet?
As of June 19, 2026, Gemini 3.5 Pro had not launched publicly despite Google promising general availability in June. It remains in limited preview for select Vertex AI enterprise customers, with no confirmed public release date.
3. What is the Agentjacking attack, and which AI tools does it affect?
Agentjacking is a security exploit disclosed by Tenet Security that injects malicious instructions into fake Sentry error reports, tricking AI coding agents into executing attacker commands. It successfully compromised Claude Code, Cursor, and OpenAI Codex in controlled testing with an 85 percent success rate.
4. Why did OpenAI acquire Ona?
OpenAI bought Ona, formerly known as Gitpod, to give its Codex coding agent persistent cloud environments that keep running after a developer closes their laptop, enabling tasks that take hours or days rather than minutes.
5. Is AI actually causing the rise in tech layoffs in 2026?
The evidence is mixed. While companies frequently cite AI as a rationale for layoffs, a Gallup survey found only 1 percent of laid-off workers themselves identified AI as the primary cause, with budget cuts and restructuring cited far more often.
6. What happened at the G7 summit involving AI company CEOs?
Sam Altman, Dario Amodei, and Demis Hassabis all attended the G7 summit in France from June 15 to 17, 2026, marking the first joint appearance of the three rival frontier lab CEOs before world leaders, with frontier AI risk and sovereignty on the formal agenda.
7. Why did Anthropic open an office in Seoul during an export control dispute?
Anthropic opened its Seoul office on June 17, 2026, to deepen Korean enterprise partnerships even as the same week’s export controls had halted access for some Korean institutional partners, illustrating the tension between global expansion and US regulatory restrictions.
8. What is GPT-Rosalind, and what does it do?
GPT-Rosalind is OpenAI’s reasoning model built specifically for life sciences research, including medicinal chemistry and genomics. An updated version released June 18, 2026, helped researchers identify diagnostic leads in 18 previously unsolved rare disease cases.
9. How much funding did AI startups raise in this period?
Notable rounds included Sixfold’s $30 million Series B for AI insurance underwriting and AlphaSense’s $350 million round at a $7.5 billion valuation for AI-driven market intelligence software, alongside larger infrastructure deals reported throughout June.
10. Will Fable 5 and Mythos 5 access be restored?
Anthropic executives expressed optimism that the export restrictions would be resolved quickly, with one executive suggesting a resolution within days, though as of late June, the suspension remained in effect with no official restoration date confirmed.
