Another AI lab is burning through cash as Tencent earnings rise
THE SO WHAT
Labs burning cash while platforms like Tencent throw off earnings is the classic R&D–distribution split — value accrues to whoever owns customers and channels. If you’re not a hyperscaler or a mega-platform, assume foundation models trend toward being a cost of goods sold, not a margin engine, and build your economics accordingly.
READ THE SOURCE
MORE FROM THE WIRE
Tech & InnovationLessons from Ukraine on keeping up morale
Sustaining morale under prolonged stress is an operational capability, not a soft skill. Early signal for founders and execs: treat psychological resilience like you treat runway and infra capacity — something you design for, measure, and reinforce, especially as AI cycles compress expectations and volatility.
Tech & InnovationThe UK's on-device scanning plan is a threat to enterprise environments
Government-mandated on-device scanning collapses the boundary between personal and corporate devices — and with it, clean threat models. CISOs with UK exposure need to scenario-plan now for regulated client software with privileged access on endpoints and what that means for data residency, key management, and zero-trust assumptions.
Tech & InnovationAI capex could hit $1.6 trillion next year, says fund manager who sees more echoes of 1998 than the dot-com bust
If AI capex really pushes toward $1.6T, the constraint for operators won’t be ‘is there compute’ but ‘who controls the financed capacity and on what terms’. Treat hyperscaler AI buildout like you’d treat a new utility monopoly — map your dependency, pricing power exposure, and where you can realistically negotiate or diversify.
.jpg)
Tech & InnovationYou’re Thinking About Online Trends All Wrong
If virality is a poor proxy for durable behavior, AI-accelerated trend cycles are even noisier. Treat social and cultural “signals” — especially those amplified by AI content — as hypotheses to test against your own data, not as roadmaps for product or brand decisions.