ReportGem ReportGem 中文

REAL-TIME GLOBAL RESEARCH

GREED & fear: The mother of all cycles

Published: 2026-06-25Institution: JefferiesPages: 21Original language: EnglishEvidence page: 1

Research evidence excerpt

GREED & fear: The mother of all cycles

25 June 2026

The mother of all cycles

Verbier

The past week has seen another DeepSeek moment. GREED & fear refers to the launch on 13 June of GLM-

5.2 by Hong Kong-listed Z.ai, formerly known as Zhipu AI. GREED & fear is no expert but GREED & fear hears

from more informed sources that this new model is almost equal to Anthropic as a competitor for the

corporate market and is just one quarter of the cost in terms of cost per token. This comes at a time when the

reaction against tokenmaxxing, previously discussed here (see GREED & fear – The joys and costs of

tokenmaxxing, 28 May 2026), is likely to lead to a slowdown in Anthropic’s uptil now remarkable revenue

growth ahead of its planned IPO. Anthropic’s annualised run-rate revenue has surged from US$9bn at the end

of 2025 to US$47bn in May (see Exhibit 1). The challenge posed to the likes of OpenAI, which has already lost

out to Anthropic in the corporate market, is all the greater, most particularly as it is also hoping to IPO.

Exhibit 1: Anthropic annualised revenue run-rate

50 (US$bn) 47

45 Anthropic annualised revenue run-rate

35 30

19 20

14 15

9 10 7

5 1 2 3 4 5

Jan-25 Mar-25 May-25 Jul-25 Aug-25 Oct-25 Dec-25 Feb-26 Mar-26 Apr-26 May-26

Source: Anthropic

The arrival of GLM comes at a time when the reaction against tokenmaxxing has already led in recent weeks

to growing use of cheaper Chinese models. Top Chinese AI models processed 21.37tn tokens on the global

aggregator platform OpenRouter in the week ended 21 June, up from 4.37tn in late April, compared with 5.76tn

tokens for the top US models, based on the weekly usage of the top nine models on OpenRouter (see Exhibit

2). This is a reminder of what already seemed to be the base case, namely that large language models will

The English excerpt is extracted automatically from the cited source page and may contain layout or recognition errors. It is never batch translated.

Open report viewer