Anthropic PreStocks (ANTHROPIC) on Solana

Prediction market traders on Polymarket are giving 77% odds that Anthropic will extend free-tier access to Claude Fable 5 past the current July 19 deadline, with over $57,000 in volume traded on the contract. Anthropic has already pushed the cutoff back three times—from an initial June 22 date to July 7, then July 12, and now July 19—with each extension announced hours before expiration.

Subscribers on Pro, Max, Team, and eligible Enterprise plans currently access Fable 5 at no extra cost for up to 50% of their weekly usage limits. Once the promotional window closes, the model will shift to prepaid usage credits priced at $10 per million input tokens and $50 per million output tokens, marking what observers are calling the first usage-based billing model adopted by a frontier AI lab for consumer subscriptions. A separate Polymarket contract puts a 68% chance on Anthropic announcing permanent paid-plan access, while intensifying competition from OpenAI's GPT-5.6 Sol adds pressure to Anthropic's pricing decisions.

JPMorgan Chase CEO Jamie Dimon said at Senator Dave McCormick's Pennsylvania Defense and Innovation Summit on July 15 that the risks posed by Anthropic's Mythos AI model are a "real issue" receiving active attention from U.S. government officials. Dimon compared granting broad access to the model to "giving ballistic missiles to individuals," underscoring concerns about misuse by adversarial military or intelligence actors.

Mythos, released by Anthropic in April 2026 to a select group of organizations including JPMorgan, is valued for its ability to identify cybersecurity vulnerabilities and help financial institutions detect and remediate security weaknesses at speed. In June 2026, U.S. authorities ordered Anthropic to restrict Mythos and Fable 5 access for foreign nationals on national security grounds; that restriction was subsequently lifted after Anthropic put enhanced safeguards in place. Dimon's public remarks signal that scrutiny over frontier AI model distribution remains high even as access controls evolve.

Anthropic is advertising 32 specialized positions aimed at preventing its AI systems from being weaponized for catastrophic harm, with roles targeting chemical and explosives misuse, nuclear weapons development, cybercrime, financial fraud, and radiological threats. Salaries for the positions run into the mid- to upper-$200,000s. The company says hiring domain experts in sensitive fields allows it to "stress-test our systems and bolster our defenses before a model ever goes live," reflecting a policy of adversarial red-teaming before any model deployment.

The push underscores Anthropic's safety-first founding thesis. CEO Dario Amodei has publicly warned that bad actors could exploit AI to enable biological or other mass-casualty attacks, and Anthropic has previously broken with the U.S. Defense Department over concerns about surveillance and autonomous weapons applications. The company already employs hundreds of people in dedicated safety functions, and the new wave of hires signals that those commitments are scaling alongside its models — a distinction the company is clearly seeking to maintain as it approaches a potential IPO.

Elon Musk's recent declaration on X that Anthropic is the "clear leader in AI" — praising its Mythos and Fable model families, predicting an imminent Mythos 2 release, and pledging that SpaceX's computing infrastructure would never be weaponized against the company — looks different in light of the underlying business relationship. According to analysis from Yahoo Finance, Anthropic leases essentially the entire output of xAI's Colossus 1 data center, roughly 300 megawatts of capacity, for approximately $1.25 billion per month through 2029, a contract worth around $40 billion in total. That arrangement makes Anthropic among the largest customers Musk's AI venture has, giving him clear financial incentive to keep the relationship cordial and publicly visible.

The timing amplifies that reading. Both Anthropic and OpenAI filed confidential IPO paperwork in June 2026, with Anthropic targeting an October Nasdaq debut and OpenAI expected to follow in 2027. Musk is simultaneously locked in litigation against OpenAI, and analysts note that elevating Anthropic as the benchmark competitor serves a dual purpose: it positions xAI's biggest client for a stronger public-market valuation while casting OpenAI as the laggard in the race. For Anthropic, the endorsement from a direct competitor — however strategically motivated — provides a notable credibility signal ahead of what could be a $1 trillion listing.

Anthropic has announced a strategic partnership with LTM, the technology services arm of the Larsen & Toubro Group, aimed at scaling enterprise adoption of Claude, Claude Code, and Claude Cowork across banking, financial services, hi-tech, consumer, and industrial sectors. The deal is structured around three delivery pillars: LTM's BlueVerse AI Fabric, which integrates Claude into software engineering and agent orchestration workflows; an AI1000 talent program that will certify thousands of engineers as Claude architects; and a dedicated Claude Center of Excellence for governance, reference architectures, and compliance tooling. Anthropic's Chris Ciauri highlighted LTM's combination of delivery expertise, trained personnel, and established client relationships as key to embedding Claude into core enterprise systems.

The partnership is a clear expression of Anthropic's partner-led go-to-market playbook — rather than building a large direct sales force, Anthropic is co-opting systems integrators with deep vertical relationships to carry Claude into large-scale deployments. LTM's internal adoption of Claude to benchmark productivity gains across the software development lifecycle is also notable, giving the partnership a built-in proof point that can be sold directly to enterprise clients.

Elon Musk publicly reversed his long-standing criticism of Anthropic on July 10, 2026, stating "I was clearly wrong about Anthropic. They are obviously currently the leader in AI." The admission marks a sharp turn from a sustained campaign of attacks spanning late 2025 through early 2026, during which Musk accused Anthropic of stealing training data "at massive scale," described Claude models as "misanthropic and evil," claimed the company "hates Western Civilization," and predicted that winning was "never in the set of possible outcomes" for the firm. Musk also praised Anthropic's Claude Mythos and Fable models as superior to competing systems and pledged fair treatment despite competitive rivalry.

The reversal comes amid a significant commercial relationship: Anthropic pays approximately $1.25 billion monthly for access to SpaceX's Colossus 1 supercomputer facility through 2029. When questioned about whether that infrastructure dependency gave him leverage to harm a rival, Musk rejected the idea publicly, citing Tesla's open-sourced patents and SpaceX's track record of fair competitor treatment as evidence of his broader business philosophy.

Anthropic is taking over an entire 16-story office building at 330 Hudson Street in Manhattan, expanding from a smaller existing office at 155 Sixth Avenue, with desk capacity for roughly 1,700 people. The company plans to more than double its New York headcount — from under 500 employees at the start of 2026 to over 1,000 by year-end — and is actively hiring across research, engineering, policy, sales, and operations.

The move is part of Anthropic's broader $50 billion U.S. AI infrastructure commitment and reflects a deliberate bet on New York as a commercial AI hub. Chief commercial officer Paul Smith cited the city's concentration of financial institutions, media companies, and cultural organizations as central to Anthropic's go-to-market strategy, positioning the company as a technology partner to those industries rather than merely a model provider.

Anthropic appointed former Federal Reserve Chair Ben Bernanke to its Long-Term Benefit Trust (LTBT) on July 9, 2026, making him the fourth member of the independent oversight body. The LTBT was established in 2023 to hold Anthropic to its public benefit mission; trustees carry no equity stake and receive no share of profits, and the trust holds authority to appoint members to Anthropic's board while advising leadership on AI risks and societal impact.

Bernanke, who led the Fed from 2006 to 2014 through the 2008 financial crisis and received the Nobel Prize in Economic Sciences in 2022, brings a track record of managing systemic risk at scale. Anthropic's President noted that "Ben's career has run from studying disruptive economic moments to steering the world's largest economy," signaling that the company is seeking economic credibility alongside its existing policy and security expertise on the trust. Bernanke said the potential of AI is "enormous" and that he is "honored to have this opportunity."

China's National Vulnerability Database (NVDB), operated by the country's industry ministry, issued a security advisory alleging that Anthropic's Claude Code AI coding assistant contains a built-in monitoring mechanism capable of transmitting users' geographic location and identity-related identifiers to remote servers without consent. The warning covers versions 2.1.91 through 2.1.196, and the NVDB urged organizations and individuals to immediately inspect affected systems, remove flagged versions or upgrade to the latest release, and impose tighter restrictions on external network access for development tools.

Alibaba responded by prohibiting employees from using Claude Code for all work purposes, effective July 10, after security researchers alleged the tool contained hidden code designed to detect whether users were located in China or affiliated with Chinese AI labs. Anthropic has acknowledged the mechanism, describing it as an experiment launched in March aimed at preventing account abuse and protecting against model distillation, and said the code was removed on July 1. The episode has sharpened the ongoing technology rivalry between the U.S. and China, with Alibaba directing staff to switch to its own Qoder alternative.

Anthropic filed confidentially for an IPO on June 1, 2026, and analysts now suggest the company could seek a $1 trillion valuation when it goes public — potentially as soon as fall 2026. That would mark a step up from its last private market valuation of $965 billion, set during a June 2026 funding round, as the company's annualized revenue approaches $50 billion, up from a roughly $4 billion run rate in mid-2025.

Two publicly traded companies already hold significant stakes ahead of any listing. Amazon has invested $8 billion in Anthropic through convertible notes — portions of which have converted to nonvoting preferred stock — giving it an estimated mid-to-high teens percentage of the company, a stake now worth an estimated $135–160 billion; Amazon recognized $16.8 billion in pre-tax gains from its Anthropic investments in Q1 2026 alone and has committed to invest up to $20 billion more. Alphabet holds approximately 14% of Anthropic (contractually capped at 15%), valued at roughly $135 billion at the last private round, and reported $28.7 billion in equity securities gains in Q1 2026, with a commitment of up to $40 billion in additional funding.