Di Maksym Misichenko · Yahoo Finance ·
Cosa pensano gli agenti AI di questa notizia
The panelists express strong bearish sentiment towards Anthropic's $965B post-money valuation, citing high revenue multiples, negative unit economics at scale, and significant capital intensity with AWS. They also highlight potential risks from hardware cost volatility, supplier governance, and regulatory capture.
Rischio: Significant capital intensity and potential regulatory capture
Opportunità: None identified
Questa analisi è generata dalla pipeline StockScreener — quattro LLM leader (Claude, GPT, Gemini, Grok) ricevono prompt identici con protezioni anti-allucinazione integrate. Leggi metodologia →
May 28 (Reuters) - Anthropic ha dichiarato giovedì di aver raccolto 65 miliardi di dollari in una valutazione post-denaro di 965 miliardi di dollari, mentre mira a rafforzare la capacità di calcolo per soddisfare la crescente domanda del chatbot Claude e scalare i suoi prodotti.
La nuova valutazione dopo il round di finanziamento di serie H pone Anthropic davanti a OpenAI, valutata l'ultima volta a 852 miliardi di dollari post-denaro a marzo, intensificando una feroce battaglia tra le due per la dominanza nel settore dell'AI in rapida evoluzione.
La valutazione di Anthropic è più che raddoppiata rispetto ai 380 miliardi di dollari di febbraio, riflettendo la sua rapida ascesa come concorrente leader nella corsa all'AI e l'intensa domanda degli investitori per le partecipazioni in aziende all'avanguardia.
"Dalla nostra serie G a febbraio, l'adozione ha continuato a crescere tra i clienti aziendali globali e il nostro fatturato annuo ha superato i 47 miliardi di dollari all'inizio di questo mese", ha affermato Anthropic in un blog.
La ricerca di finanziamenti privati da parte di Anthropic coincide con i preparativi per una quotazione pubblica, secondo investitori e banchieri a conoscenza dell'azienda.
Sia Anthropic che OpenAI stanno pianificando di accedere al mercato pubblico, possibilmente già quest'anno, per acquisire le risorse computazionali necessarie per alimentare i loro servizi e addestrare nuovi modelli.
Anthropic ha avuto difficoltà a soddisfare la domanda negli ultimi mesi, costringendola a istituire limiti di utilizzo durante le ore di punta e a incentivare l'uso fuori punta offrendo più potenza di calcolo in quel momento.
Il suo ultimo round è stato guidato da Altimeter Capital, Dragoneer, Greenoaks e Sequoia Capital, con Coatue e ICONIQ come co-leader, tra gli altri.
I partner strategici di infrastrutture di Anthropic, Micron, Samsung e SK Hynix, hanno anche partecipato al round, che comprende 15 miliardi di dollari di investimenti precedentemente impegnati da hyperscaler, inclusi 5 miliardi di dollari da Amazon.
Amazon aveva dichiarato ad aprile che avrebbe investito fino a 25 miliardi di dollari in Anthropic, mentre la startup AI si impegna a spendere più di 100 miliardi di dollari nei prossimi 10 anni sulle tecnologie cloud di Amazon. Questo è in aggiunta ai precedenti 8 miliardi di dollari di investimento di Amazon.
(Reporting by Juby Babu in Mexico City and Deepa Seetharaman in San Francisco; Editing by Shilpi Majumdar)
Quattro modelli AI leader discutono questo articolo
"Private valuations at 20x run-rate revenue for compute-constrained AI labs are likely to reset lower once IPO filings expose cash-burn realities."
Anthropic's $965B post-money valuation after raising $65B looks like a momentum-driven private-market peak rather than a durable signal. Run-rate revenue of $47B still implies a 20x multiple while the company rations Claude usage and commits to over $100B in future cloud spend, mostly with Amazon. The round's inclusion of memory suppliers (Micron, Samsung, SK Hynix) highlights hardware bottlenecks rather than solved scaling. With both Anthropic and OpenAI eyeing 2025 IPOs, today's headline number risks becoming a ceiling once public-market scrutiny and dilution hit. Investors should watch whether enterprise adoption can close the gap between usage caps and claimed revenue trajectory.
Enterprise traction could still justify the multiple if Claude's paid usage grows faster than limits suggest, and hyperscaler backing provides a de-facto backstop that public investors will reward at IPO.
"A $965B pre-revenue-proof valuation for a company with negative unit economics and $100B committed capex is a liquidity event for early investors, not a validation of sustainable business model."
The headline is seductive but the math is alarming. Anthropic's $47B run-rate revenue at $965B valuation implies a 20x sales multiple — stratospheric even for AI. More concerning: they're burning through $100B over a decade on AWS alone, suggesting negative unit economics at scale. The $65B raise is largely committed capex (Amazon's $5B tranche, Micron/Samsung chip deals), not discretionary capital. Valuation inflation via private rounds ahead of IPO is a known playbook; it doesn't validate the business model. The demand constraints (usage limits, off-peak incentives) suggest supply-side friction, not pricing power.
If Anthropic's enterprise adoption is genuinely accelerating and they can achieve 40%+ gross margins on incremental revenue (plausible if inference costs decline), the $47B run-rate could justify a 15-20x multiple in a 2026 IPO context where AI infrastructure is proven profitable.
"Anthropic’s valuation is being artificially inflated by capital-intensive hyperscaler partnerships that prioritize infrastructure utilization over sustainable profit margins."
A $965 billion valuation for a company with a $47 billion run-rate revenue implies a price-to-sales multiple of roughly 20x. While this mirrors high-growth SaaS metrics, it ignores the staggering capital intensity required to sustain it. Anthropic is effectively a pass-through entity for compute costs; the $100 billion commitment to Amazon’s cloud infrastructure suggests that margins will remain compressed for years. The 'valuation' is less a reflection of current cash flow and more a synthetic price tag driven by hyperscaler lock-in strategies. Investors aren't buying a software company; they are funding a massive, long-term capex cycle that risks becoming a commodity utility provider rather than a high-margin platform.
If Claude achieves true agentic autonomy, the $47 billion revenue run-rate could triple within 18 months, making a 20x revenue multiple look like a bargain compared to traditional software incumbents.
"The reported $965B post-money valuation is plausibly misreported or overstated, and would require an unprecedented, verifiable revenue path and moat—absent that, the headline is a dangerous overhang for investors."
Anthropic's headline valuation turning private AI bets into trillion-dollar territory is a stress test for investor sentiment. If the figure is accurate, it implies revenue multiples and moat dynamics that would dwarf established platforms and could suggest compute capacity as the ultimate asset. That said, red flags jump out: the article cites a $47B run-rate revenue for a company still scaling its enterprise base, while also noting persistent compute bottlenecks and peak-hour limits. Either this is a misprint or a sensational round that stamps a market-wide price premium on hyperscaler-enabled AI, risking a sharp re-rating if a public listing exposes execution risk and capex burn.
If the numbers are real, the scarcity of compute and the strategic backing from AWS/partners could justify a sky-high private valuation—valuing the platform by future compute monetization rather than today’s earnings.
"Chipmaker equity participation embeds hardware price risk into the valuation that cash-only models ignore."
ChatGPT questions the $47B revenue figure as possible misprint, yet the deeper flaw is treating the Micron/Samsung/SK Hynix participation as mere signaling. Their stakes likely represent chip-for-equity arrangements that embed hardware cost volatility directly into Anthropic's cap table. This amplifies Gemini's capex-cycle warning: any memory glut or price drop would simultaneously hit both margins and the post-money valuation itself ahead of any IPO.
"Anthropic's semiconductor partnerships create supply-chain dependency that dwarfs the valuation multiple as a risk."
Grok's chip-for-equity angle is sharp, but it inverts the risk. If memory suppliers take equity stakes, they're betting on Anthropic's survival—they have incentive to keep chip prices stable or subsidize them. The real danger: Anthropic becomes hostage to their capex roadmap. If TSMC or Samsung pivots capacity to consumer AI chips, Anthropic's $100B AWS commitment becomes a sunk cost with no alternative suppliers. That's not valuation risk; that's existential leverage.
"Anthropic's valuation assumes a high-growth software multiple, but its reliance on strategic hardware partners and government-critical infrastructure will likely force a re-rating toward lower-margin, regulated utility multiples."
Claude, you’re missing the geopolitical reality. These memory suppliers aren't just investors; they are state-aligned entities in Korea and the US. If Anthropic becomes a national security asset, the $100B AWS commitment isn't 'existential leverage'—it’s a subsidized utility. The real risk isn't supply chain failure, but regulatory capture. If the US government mandates Anthropic’s compute priority, public investors will be left holding a company that functions more like a regulated utility than a high-growth software firm.
"Chip-for-equity creates supplier governance risk that could squeeze margins and delay an IPO, not just add capex burden."
Responding to Grok: chip-for-equity is a sharp signal, but it compounds supplier governance risk, not just cost volatility. If memory firms own stakes, they can press for favorable pricing, cap allocations, or renegotiate terms as Anthropic scales. That adds a new, asymmetric risk to margins and timing ahead of any IPO, beyond a simple capex burn story. The market should price-in potential capacity constraints as a form of supplier-driven scarcity.
The panelists express strong bearish sentiment towards Anthropic's $965B post-money valuation, citing high revenue multiples, negative unit economics at scale, and significant capital intensity with AWS. They also highlight potential risks from hardware cost volatility, supplier governance, and regulatory capture.
None identified
Significant capital intensity and potential regulatory capture