TickerTain
TickerTain
NewsroomShortsPortfolioConvergence
NewsroomShortsPortfolioConvergence
←
▶ 38:09 · SaaS Business Models · AI spending crowds out traditional software budgets, leaving application layer stocks vulnerable despite historical stickiness
episode briefing
The Information

On the Ground at UBS’ Private AI, Software and Internet Conference | Menlo Park

2026-07-16 · 2 company · 7 thematic
sentiment
0 bull0 bear2 neu
speakers
carl kirstead

Leads UBS research coverage of software and AI; articulates a 'long infra data security, cautious apps' framework, analyzes hyperscaler vs semiconductor divergence, and tracks enterprise AI budget dynamics.

now playing · SaaS Business Models
AI Safety & Alignmenttailwindscore 7/10daniele megazzini
UBS Chief AI Officer emphasizes evals and mathematical proofs as critical for deploying AI in regulated financial services
Trust and reliability are paramount for AI in banking; translating human expertise into machine-readable evaluations and proving agent correctness mathematically are key research frontiers…
Enterprise AI Adoptiontailwindscore 7/10daniele megazzini
UBS adopts model garden strategy and mandatory AI learning hours to drive internal adoption and cost efficiency
Successful enterprise AI adoption requires routing tasks to the right model (frontier vs. small language models) via a model garden, and mandating hands-on learning time for employees to ov…
Semiconductorsmixedscore 8/10carl kirstead
Semiconductor stocks seen as easy longs on estimate revisions and memory inflation while hyperscalers face FCF pressure
Semiconductor stocks benefit from upward estimate revisions and DRAM/memory cost inflation driving revenue upside, whereas hyperscalers see free cash flow estimates decline as CapEx rises,…
Frontier AI Modelsriskscore 8/10carl kirstead
Frontier labs likely to vertically integrate into software, creating competitive overlap with incumbents like Microsoft and Salesforce
As model markets become more competitive with cheaper alternatives, frontier labs such as OpenAI and Anthropic will push up the stack into productivity and application software, directly ch…
AI Economics & Business Modelsmixedscore 7/10carl kirstead
Enterprise token cost overruns drive down-tiering to cheaper models, but Jevons paradox may eventually boost total AI consumption
Enterprises are throttling back on expensive frontier model usage as token costs exceed budgets, shifting to smaller or open-source models; this near-term headwind for frontier labs and inf…
SaaS Business Modelsheadwindscore 9/10carl kirstead
AI spending crowds out traditional software budgets, leaving application layer stocks vulnerable despite historical stickiness
Rising enterprise AI spend is crowding out traditional software and IT services budgets, and investors now price in the threat that improving AI models will let customers build custom alter…
Robotics & Physical AItailwindscore 7/10dustin peterson
Locus Robotics CFO sees massive long-term opportunity in warehouse robotics but warns real-world data scarcity keeps the sector fragmented for years
Physical AI requires real-world deployment data to handle edge cases, creating a chicken-and-egg problem that favors companies with deployed fleets; a single 'robotics brain' is decades awa…