
according to the latest reports, ai startup giant anthropic has committed to paying google a staggering $200 billion over the next five years for cloud services and in-house chip computing power. this figure underscores the extraordinarily high computational threshold currently facing the ai industry—industry estimates indicate that this deal accounts for more than 40% of google’s recently disclosed total future revenue commitments.
across the entire industry, orders from the two leading firms, anthropic and openai, now make up more than half of the pending order backlog at major u.s. cloud service providers, with the total value approaching $1 trillion. however, the intricate financial structure behind this “deal of the century” has also sparked widespread market skepticism. google and amazon are not only anthropic’s service providers but also its core investors.
financial reports show that alphabet’s first-quarter profits surged, largely due to the revaluation of its equity stake in anthropic. this dual role—as both landlord and shareholder—has been jokingly described by some industry insiders as a “magic trick” of the capital markets, raising suspicions of circular transactions that could lead to a disconnect between corporate valuations and actual returns. current cost projections further highlight the brutal nature of the ai arms race: by 2026, openai’s server expenditures could reach $45 billion, while anthropic is expected to spend another $20 billion. although major tech giants’ capital spending in the ai sector now rivals government-level investments, whether this capital-intensive, high-investment business model can ultimately translate into sustainable profits remains a key question for investors.
with anthropic’s valuation in the secondary market surpassing the $1 trillion mark, the bubble game in the ai industry has entered a feverish phase. whether this competition, fueled by cutting-edge computing power and massive capital, truly represents the inevitable path toward artificial general intelligence—or is merely another illusion driven by valuation hype—will ultimately depend on time.