Copyright Investing.com
Alphabet (NASDAQ: ) shares jumped 3.5% after hours on news of a potential "high tens of billions" cloud computing deal with AI startup Anthropic. But the real story isn’t the headline number. It’s who’s capturing the lion’s share of AI’s explosive growth. When Anthropic and Google Cloud confirmed they’re negotiating a massive infrastructure deal valued in the high tens of billions of dollars, Wall Street saw it as validation of Google’s AI strategy. And they’re right, but not for the reasons most investors think. The announcement sent Alphabet stock surging while Amazon shares slipped 2%, reflecting concerns about Anthropic’s multi-cloud diversification. Yet beneath the surface lies a more profound shift in how value is distributed in the AI economy, with implications for every investor trying to navigate this transformative technology. The Economics That Matter There’s one important point forgotten: cloud providers are capturing 16-32% of AI startup revenue before those companies even approach profitability. Morgan Stanley’s analysis of Amazon’s (NASDAQ: ) similar partnership with Anthropic reveals AWS will generate $1.28 billion from the company in 2025, climbing to nearly $3 billion in 2026 and $5.6 billion by 2027. This is infrastructure revenue that scales regardless of whether Anthropic itself makes money. Google Cloud’s operating margin stands at about 20.7%, meaning this deal could generate billions in high-margin revenue while Anthropic continues burning through capital. The startup projects $26 billion in revenue by 2026 but remains deeply unprofitable. Anthropic’s $183 billion valuation on roughly $5 billion in annualized revenue represents a 36x revenue multiple (extraordinary even by AI standards), but the company won’t see positive cash flow for years. It’s a clear asymmetry: Google monetizes Anthropic’s growth immediately through infrastructure spending, while the startup bears all the risk of achieving profitability in an intensely competitive market. The Multi-Cloud Power Play What makes this deal especially strategic is Anthropic’s deliberate positioning across multiple cloud providers. The company maintains substantial relationships with both Google ($3 billion invested, 14% stake) and Amazon ($8 billion invested), a balancing act that’s more sophisticated than simple vendor diversification. This multi-cloud strategy serves three critical purposes. First, it provides negotiating leverage. With credible alternatives, Anthropic can extract better pricing and terms from each provider. Second, it addresses capacity constraints in a market where every major cloud provider faces GPU shortages and massive infrastructure backlogs. AWS alone reports $195 billion in committed but undelivered capacity. Third, and perhaps most strategically, multi-cloud positioning makes Anthropic a neutral player that enterprises can adopt without strengthening their competitors. A bank hesitant about deepening ties with Amazon can use Claude through Google Cloud; a retailer wary of Google can access the same AI through AWS Bedrock. In enterprise AI adoption (where 80% of organizations now use multi-cloud strategies), this optionality is increasingly valuable. Strategic AI partnerships like the Anthropic deal are critical for Google to narrow the market share gap with AWS, while pressuring Amazon’s slower growth trajectory Cloud Wars 2.0: The Infrastructure Battleground The Anthropic deal shows the real competition in cloud computing. While AWS still leads with 30% market share and $30.9 billion in Q2 revenue, its 17.5% growth rate lags behind Azure’s 39% and Google Cloud’s 32%. Google is aggressively deploying strategic partnerships to close the gap, recently securing a $10 billion, six-year deal with Meta and expanding collaborations with Oracle and ServiceNow. What’s particularly notable is the margin profile. AWS generates a 32.9% operating margin, while Google Cloud achieved $2.8 billion in operating income (20.7% margin) in Q2 2025, up 133% year-over-year. These margins suggest substantial pricing power, even as compute costs theoretically decline. There’s one counterintuitive reality: Why have falling GPU rental prices (down over 70% since 2021 for some chip types) not translated to proportional cost savings for AI companies? Because the total infrastructure stack (networking, storage, data egress, specialized services) creates lock-in effects that preserve cloud provider economics. The tens of billions Anthropic commits to Google represents not just compute access but the entire integrated infrastructure ecosystem. Valuation Reality Check Anthropic’s $183 billion valuation also demands scrutiny. Traditional SaaS companies trade at 2.5-7x revenue, while AI SaaS companies command 20-30x multiples. Top-tier AI infrastructure startups like Anthropic see multiples of 40-50x, with outliers approaching 100-200x. Nothing irrational, but a calculated bet on network effects and data moats. AI models improve with usage, creating compounding advantages. However, investors should note the disconnect between valuations and current economics. OpenAI, valued at around $500 billion, may generate only $13 billion in 2025 revenue while burning billions annually. The bet is that once these models achieve sufficient scale, margins will expand dramatically. But that inflection point remains theoretical, and the path to profitability requires continued massive infrastructure spending, flowing directly to cloud providers. The Enterprise Adoption Inflection Anthropic’s explosive growth (from $1 billion in early 2025 to over $5 billion by August) is driven almost entirely by enterprise adoption. The company now serves 300,000 business customers (up from under 1,000 two years ago), with enterprise representing 80% of revenue. Companies like Norway’s sovereign wealth fund achieved 20% productivity gains; Novo Nordisk (NYSE: ) reduced clinical documentation time by 99.9%. More than marginal improvements, they’re transformational outcomes justifying continued investment. However, 74% of companies remain in the pilot phase with AI, and only 5% have integrated AI tools across operations. The parallel scaling of Anthropic’s projected revenue (from $5 billion in 2025 to $26 billion in 2026) against AWS’s captured share, rising from $1.28 billion in 2025 to $5.6 billion by 2027. Cloud providers maintain 16-32% of AI startup economics, benefiting from growth without bearing development risks, a pattern likely mirrored in the Google deal What Investors Should Watch The Anthropic-Google deal offers clear investment implications: Cloud infrastructure providers capture disproportionate value. With high-margin revenue representing 16-32% of AI startup economics, companies like Alphabet, Amazon, and Microsoft Corporation (NASDAQ: ) monetize AI growth regardless of which models win. Custom silicon matters. Google’s TPU advantage allows 40% cost reductions versus Nvidia-dependent competitors, strategically significant as AI workloads scale. Enterprise metrics trump model benchmarks. Anthropic’s growth from 1,000 to 300,000 business customers represents genuine commercial traction. Focus on customer count, average revenue per account, and retention rather than model performance scores. Regulatory constraints create unexpected value. Antitrust limitations on big tech AI acquisitions increase independent players’ leverage while forcing cloud providers to compete via partnerships, favoring both hyperscalers and top-tier AI independents. The tens of billions flowing from Anthropic to Google is much more than just infrastructure spending. It’s a transfer of wealth from venture-backed AI companies to established cloud providers with the capital and scale to monetize this transition. In the AI revolution, the arms dealers are capturing more certain value than the prospectors.
