ANTHROPIC Anthropic PBC : Bullish and Bearish Analyst Opinions

Anthropic's reported $30B ARR uses GAAP-compliant accounting that books 100% of cloud partner resale revenue and marks the ~80% partner share as a marketing expense, unlike OpenAI's method. This makes Anthropic's topline revenue appear artificially inflated and not directly comparable to peers, misleading investors and observers about its true scale and efficiency. AVOID making investment or competitive assessments based on the headline ARR number, as it presents a distorted view of the company's underlying business. Increased scrutiny leads to accounting standard changes or the market corrects its valuation perception.

Anthropic is launching a controlled release of its advanced "Mythos" AI model to 11 partner companies first due to cybersecurity risks, as it can effectively find software bugs. The speaker states Anthropic is "a little ahead" of competitors like OpenAI and Google but "not overwhelmingly ahead" and lacks a "permanent moat." The model's capabilities present both significant economic value and security risks. The staggered release is a defensive move, but the intense competitive race pressures all players to prioritize speed over safety. WATCH due to its technological edge and the strategic implications of its controlled release, balanced against intense competition and rising safety concerns in the AI arms race. Competitors (OpenAI's "Spud," Google's Gemini) quickly catch up or leapfrog, negating the temporary advantage. Safety corners cut in the race lead to a major reputational or operational failure.

The speaker states Anthropic's new "Mythos" model is so powerful at hacking software it's essentially a "cyber weapon of mass destruction," leading the company to withhold public release and work only with a consortium of critical partners. This capability creates an existential, two-tier dynamic in the AI race and national security. The speaker infers this forces a conversation about nationalization and covert government use, comparing it to the Manhattan Project. The situation demands close monitoring (WATCH) because it represents a pivotal, high-stakes moment for the company, the AI industry, and geopolitics, with unpredictable outcomes for valuation and strategy. The core claims about the model's danger could be overstated for IPO narrative purposes. An open-source model could achieve parity faster than expected, undermining the strategic advantage.

The speaker states Anthropic has taken a 73% market dominance in first-time enterprise usage and leads in retail user growth (~1M signups/day), flipping the previous dynamic where OpenAI was dominant. Anthropic's success in enterprise and coding (Claude Opus 4.6) and its cohesive "super app" experience have made it OpenAI's primary competitor, puncturing the perception that "ChatGPT was AI." The direction is WATCH because Anthropic represents the key competitive threat and benchmark for OpenAI. Its current momentum and product strengths make it a critical factor in evaluating the AI landscape, especially ahead of its own potential IPO. OpenAI's Spud model is a major leap that negates Anthropic's advantages, or Anthropic fails to expand beyond its text/chat strengths into multimodal AI.

Chamath ordered the IPO urgency: SpaceX first, then OpenAI and Anthropic must "file quickly, get out and just get the money." He believes trillions in new market cap will cause tech sector P/E multiples to converge downward toward non-tech P/Es. A flood of giant IPOs will compete for finite capital. The first issuers (SpaceX) will be consumed eagerly, but later ones risk poor reception as investor appetite fills and the market digests the AGI/ASI risk to all other software moats. Timing is critical. These three companies have the most urgent need for capital and the strongest stories. Delaying increases the risk of being caught in a capital crunch and a sector-wide multiple compression. The IPO window closes suddenly due to macro events (e.g., Iran war escalation) or a recession. AGI proof points fail to materialize, causing loss of investor faith in the narrative.

The speaker detailed how AI inference for leading models (OpenAI's GPT, Anthropic's Claude) is massively loss-making under current subscription plans, with power users burning thousands of dollars in compute on $200/month plans. He explicitly compared the funding frenzy and unsustainabile economics to the dot-com bubble. The fundamental business model is broken because competitive pressure forces models to burn exponentially more tokens for useful outputs, eroding hardware efficiency gains. The path to profitability via per-token pricing would crater demand and is untested. WATCH because the setup for a major sector dislocation is clear, but the timing of a bust is uncertain (akin to the NASDAQ doubling after 1998 before crashing in 2000). The upcoming IPOs of OpenAI and Anthropic could be pivotal events. A breakthrough in inference efficiency or a massive, sustained subsidy from vendors/governments could prolong the unsustainable model, deferring the reckoning.

Shares of OpenAI have "fallen out of favor" on the private secondary market, with investors pivoting to Anthropic. Demand is shifting because Anthropic has been the "big disrupter this year," and its valuation is less than half that of OpenAI, with anticipation it may catch up. This indicates a loss of momentum and relative overvaluation for OpenAI in private markets, while Anthropic is gaining investor favor due to its products (like Cowork) and perceived valuation gap. AVOID OpenAI (in private secondary markets) due to loss of investor demand and high relative valuation. WATCH Anthropic for its momentum and potential valuation catch-up. OpenAI regains product momentum or Anthropic's execution falters.

When forced to choose between Anthropic and OpenAI for an all-in investment, Salen said "Anthropic," citing that "all the founders I know are using Claude Code" and it has "captured lightning in a bottle." He perceives strong product traction (Claude Code) among a key user base (founders) and sees it as having momentum, despite acknowledging both companies have issues. LONG on Anthropic based on current product-market fit and developer momentum in a forced-choice scenario against its primary competitor. Rapidly shifting competitive landscape in AI; execution or regulatory challenges specific to Anthropic.

The speaker reports Anthropic could target an IPO as soon as October 2025, putting it in a competitive "race" with SpaceX and OpenAI to go public. A crowded pipeline of major tech/AI IPOs could split investor attention and capital, making timing and market windows critical for success. WATCH because its path to public markets is developing and contingent on both company readiness and the broader market's appetite for high-profile, high-valuation tech listings, especially if SpaceX lists first. Deteriorating market conditions or a poorly received SpaceX IPO could delay Anthropic's plans significantly (potentially into 2026/2027).

Anthropic is considering going public as soon as October, boosted by a court win blocking a Trump administration ban on government use of its AI tool. IPO could provide exposure to a leading AI company with a focus on enterprise and safety, with expected high demand from investors. WATCH for IPO timing and valuation, as it represents a significant event in the AI sector with potential for market impact. Market conditions or regulatory changes could delay or affect the IPO success.

The speaker states Anthropic cited a potential loss "in the billions" from being cut out by the Pentagon due to a "supply chain risk designation." Even if Anthropic wins its case to remove the designation, it will have to "claw its way back" into agencies that are actively removing its technology, competing against entrenched rivals like Google and OpenAI. Anthropic faces significant, immediate revenue risk and a costly, uphill battle to recover its government business, putting it at a competitive disadvantage. A favorable legal ruling or a new administration could swiftly reinstate Anthropic as a preferred vendor.

Vestager explicitly stated the need for multiple AI sources (European, U.S., Chinese, Indian) to avoid lock-in, and warned that if Anthropic and OpenAI become dominant like "the internet's," competition will be lost. Over-concentration in AI by a few private companies could lead to reduced innovation, higher costs, and increased regulatory scrutiny, making investments in these companies riskier. AVOID because the speaker implies that reliance on these players is detrimental, with a push for diversification and competition that may limit their growth or attract antitrust actions. Emergence of strong competitors, successful regulatory interventions, or technological breakthroughs that decentralize AI could break the thesis.

The Pentagon has explicitly "declared Anthropic a supply chain risk." Being labeled a "supply chain risk" by the US Department of Defense is a severe reputational and commercial blow for a US-based AI company. This could hamper future government contracts. Amazon (AMZN) and Google (GOOG) are major backers of Anthropic. AVOID Anthropic exposure. The designation could be reversed or apply only to specific classified networks.

It is now possible to plug advanced LLMs (like Claude) directly into robotics hardware (like Unitree dogs) to create autonomous agents. The market is underpricing the speed at which AI moves from "screens" to "physical reality." This convergence will drive value to both the model providers (Anthropic) and the hardware manufacturers. Long the AI/Robotics intersection. Safety failures leading to strict regulation on autonomous physical agents.

The Trump administration "tore up" the guardrails and is banning Anthropic from federal use due to their refusal to support lethal applications. While Anthropic is a private company, this signals a massive reduction in their Total Addressable Market (TAM). Losing the US Defense Department—the world's largest customer for such technologies—because of ethical "red lines" puts them at a severe revenue disadvantage compared to competitors willing to comply with DoD requests. AVOID. The company is effectively locking itself out of the public sector revenue stream under the current administration. Enterprise adoption (private sector) could outweigh government contract losses.

SpaceX's timeline would "open the door for the company to beat out OpenAI and Anthropic to going public." First-mover advantage in the "Private Mega-Tech" asset class is critical. If SpaceX secures the bulk of available growth capital first, subsequent IPOs like OpenAI or Anthropic may face investor fatigue or tighter wallets. WATCH. This puts pressure on these firms to either accelerate their timelines or wait for capital reserves to replenish post-SpaceX. Investor appetite for AI/Space tech is insatiable, allowing all three to list successfully.

"I really hope that this is not a corporate against corporate playing out in government contracts... implying that there's another company that may not be so happy with the drop in getting their this kind of advantage." Hochstein alludes to a proxy war happening inside the Pentagon between major AI firms (he names Anthropic, implying the "other company" is a rival like Microsoft/OpenAI). This suggests that government contracts are becoming the new battleground for Big Tech, and political influence is determining winners and losers in AI procurement. WATCH the "Pentagon AI" space. Volatility is expected as corporate lobbying interferes with procurement. Regulatory capture could lock out superior technology in favor of politically connected firms.

"He's [Anthropic CEO] kind of caught in a bit of a lose lose situation. Even he annoys the fence [Defense] again, or he loses the backing of his own employees." Anthropic (backed by AMZN/GOOGL) is structurally handicapped in the defense sector ($800B+ budget) due to its employee base's ideology. This limits their Total Addressable Market compared to a competitor willing to do "dirty work." The ethical stance is a cap on potential government revenue upside. They may pivot if funding dries up, but that triggers the talent drain risk.

"The Anthropic story is roiling... Katrina Manson sort of come on." Anthropic is a key competitor in the AI space (backed by Amazon and Google). If a story is "roiling" the company, it implies significant volatility or governance issues similar to previous AI drama, which could impact the broader AI trade and major tech backers. Watch for details on the specific controversy or development to assess contagion risk to the broader AI theme. If the news is contained to the specific company, broader sector impact may be minimal.

Sadler states explicitly that Anthropic "really has no standing" to put limitations on the Pentagon and notes the "core issue" is the threat of the Defense Production Act (DPA) to "compel or to basically take over the intellectual property." If the government invokes the DPA, Anthropic loses control over its product deployment and potentially its IP rights. This destroys the company's "ethical moat" and introduces severe regulatory/sovereign risk to its business model (and by extension, its major backers like Amazon and Google, though not explicitly named). The speaker believes the government holds all the leverage; fighting the DoD is a losing battle for a tech vendor. The DPA might not be renewed or the government may choose not to exercise such extreme powers to avoid spooking the broader tech sector.

Anthropic is refusing the Pentagon's demand to remove AI guardrails. The DOD is threatening to label them a "supply chain risk." If Anthropic is blacklisted or sidelined by the DOD, the government's massive AI budget must go somewhere. It will flow to "patriotic" AI firms that are willing to comply with DOD requirements and lack restrictive "moral" guardrails on military application. Palantir (`PLTR`) is the logical alternative for defense-grade AI implementation. LONG PLTR (Second-order beneficiary of Anthropic's refusal). Anthropic may capitulate at the last minute to save the contract, or the DOD may soften its stance.

The Pentagon is "suggesting that despite threats to cut off ties with anthropic and in fact designate the company as a supply chain risk, he says the company is not changing its position." A "supply chain risk" designation by the DoD is toxic. It not only kills government revenue but can cause enterprise clients (banks, healthcare) to pause contracts due to perceived security or regulatory risks associated with the vendor. Avoid exposure to Anthropic (private shares or future IPO) until the regulatory cloud clears. Anthropic may reverse course; the "ethical" stance may attract significant consumer/commercial goodwill that outweighs the loss of defense revenue.

Anthropic faces a deadline tomorrow to agree to Pentagon usage terms; the Pentagon refuses to let companies dictate "mass surveillance" policies. If Anthropic balks, they lose access to massive government contracts. This would benefit competitors like Microsoft (OpenAI) or Google who may be more willing to accommodate defense requirements. WATCH the deadline outcome. A refusal by Anthropic is a bullish signal for MSFT/GOOG (competitors) who will absorb that government spend. The Pentagon caves to Anthropic's safety pledge, validating their restrictive model.

While OpenAI's growth stalls, Anthropic's Claude has been "steadily climbing." Claude Code hit a $1 billion run rate in just six months. The market is rewarding "Agents" (software that operates software) over "Chatbots." Anthropic is successfully positioning itself as the leader in this new "Agent Era," taking momentum from incumbents. LONG (Thematic/Private). OpenAI or Google successfully counter-launch superior agentic tools.

"Anthropic's Claude Code... hit $1B run rate in just six months." Conversely, "Chinese models are showing us that this can be done a lot more efficiently." Momentum has shifted away from the incumbent (OpenAI) toward Anthropic (performance) and Chinese models (efficiency). As the "Agent War" begins, capital and user growth are flowing to these newer, more specialized or efficient players rather than the legacy chatbot winner. Long the emerging winners of the Agent cycle. (Note: Anthropic is private; look for proxies or private market access. Chinese AI offers an efficiency hedge). Regulatory crackdowns on Chinese tech or API restrictions.

Zoom recorded a "$532 million pretax gain" on its Anthropic stake in Q4. When asked about selling, Yuan said: "In AI? I think we just started and huge opportunity ahead." Yuan has insider access to the utility and growth trajectory of top-tier LLMs. His refusal to take half a billion dollars in profit signals extreme confidence that the valuation of foundational model providers (and the sector at large) is nowhere near a top. This reads through positively for the broader AI infrastructure and model layer. LONG. Continued exposure to foundational AI models and infrastructure. AI valuation bubble bursting or regulatory crackdowns on model providers.

The Pentagon is demanding full access to Anthropic's AI models; Anthropic is refusing on safety grounds. The US may use the Defense Production Act to force compliance. This represents a major regulatory/sovereign risk for private AI labs. If the government seizes access or dictates terms, it alters the valuation models and operational independence of these firms (and their backers like AMZN/GOOG). WATCH for regulatory precedent that could spill over to other AI Model Providers. N/A (Private asset, monitoring for sector impact).

Secretary of War Pete Hegseth threatened to label Anthropic a "supply chain risk" if they do not grant full military access by Friday. If labeled a supply chain risk, "no company that has a deal or a contract with the US government can use Claude." This would effectively blacklist Anthropic from the enterprise market, as most major companies have some government adjacency. Conversely, if they comply, they lose their "ethical" brand moat but gain the "trillion dollar defense budget." WATCH. This is a binary outcome for the company's commercial viability. Total blacklisting by the US government leading to business failure.

"The US is using AI to kill people... Pete Hegseth demands full military access to Anthropic... Capex commits are as good as gold." The US government has declared AI a military necessity (using it for operations like the Maduro capture). They are forcing labs to cooperate via the Defense Production Act. This effectively backstops AI CapEx with the US defense budget. Palantir (intelligence layer) and Microsoft (OpenAI proxy) are the immediate beneficiaries of this "state-sponsored" AI push. LONG. AI is now a sovereign arms race; government contracts provide a floor for valuation. Regulatory backlash or "War Claude" (Anthropic) refusing to cooperate, leading to nationalization or legal battles.

"A defense official telling Axios that they need Anthropic. And the problem for these guys is they are that good... The US military thinks Cloud [Claude] is the best model in the world." While Anthropic is private, this is a massive endorsement of their tech stack over competitors (like OpenAI/Microsoft or Google). For investors monitoring the private markets or holding stakes in Anthropic's backers (Amazon/Google), this confirms the asset's strategic value is increasing despite the safety controversy. WATCH. Monitor for IPO signals or impacts on its major public backers. Loss of the $200M contract if they refuse to comply fully; alienation of safety-focused talent.