Recent OpenAI revenue revisions to around $50 billion annualized, below prior expectations, have amplified trader concerns over AI valuations amid heavy infrastructure debt and rising rates. Ray Dalio and others have flagged classic bubble signals from leveraged capex by hyperscalers and labs, while cheaper Chinese open-weight models like Qwen and GLM have cut U.S. market share from 70% to 30% through 90% lower token pricing. Microsoft’s new Maia 300 chip and 2027 multi-gigawatt plans, plus Meta’s Muse Wonder frontier releases, highlight ongoing capability gains and in-house shifts, but analysts note potential 2027 revenue shortfalls against trillions in contracted compute obligations. Competitive pricing pressure and earnings growth versus multiples remain key swing factors for any burst timeline.