TOTVS: Where does TOTVS stand as token optimization accelerates?; AMX/TEF-BZ/ TIM: TEF-BZ reportedly in talks with SpaceX for D2D launch

Research evidence excerpt

TOTVS: Where does TOTVS stand as token optimization accelerates?; AMX/TEF-BZ/ TIM: TEF-BZ reportedly in talks with SpaceX for D2D launch

Global Research

22 July 2026ab

TMT Daily Equities

Emerging MarketsTOTVS: Where does TOTVS stand as token

optimization accelerates?; AMX/TEF-BZ/ Telecommunications

TIM:TEF-BZ reportedly in talks with SpaceX for Leonardo Olmos,AnalystCFA

D2D launch leonardo.olmos@ubs.com+55-11-9658 7533

Gustavo Farias

Associate Analyst

TMT Calendar: next main events: gustavo.farias@ubs.com

+55-11-2767 6154

(July 22) AMX conference Call

(July 23) Televisa and TelevisaUnivision earnings release

(July 27) TEF-BZ earnings release

(July 27) TIM earnings release

(July 29) TEF earnings release

(July 29) Intelbras earnings release

(July 29) Cognizant earnings release

(July 30) Grid Dynamics earnings release

TOTVS: As token optimization accelerates and cheaper models gain traction,

where does TOTVS stand?

What happened: UBS analyst Karl Keirstead published a report on cheaper open

models (Chinese or otherwise) and their impact on the AI trade, drawing on

feedback from the AI-native firms that presented at UBS's 5th Annual Private AI and

Software event in Menlo Park last week. Key takeaways were:

AI demand is strong, but cost control is now front of mind. AI-native firms

reported extraordinary revenue growth, a sign that enterprises are leaning

into AI well beyond coding. At the same time, controlling token spend has

become a real concern. We concluded last month that 60% of organizations

saw AI compute and token costs as a real issue, and that figure now feels

higher.

Model routing is now standard, and cheaper open models have arrived. The

race is no longer only about model performance, but increasingly about token

efficiency and cost.