SaaSpocalypse
If work gets done by models rather than people clicking screens, per-seat pricing is the wrong unit. The market is discounting mature SaaS multiples while paying up for the compute and infrastructure layer the agents actually consume.
How this story could end...
The endings this story genuinely supports, written from the same evidence as the story updates. The value ranges are illustrative examples of how the affected companies as a group could re-rate — not forecasts, targets or advice.
Mapping out the endings…
How the two sides have traded
The bull and bear baskets for this story, equal-weighted and indexed so the two lines start together.
Each scenario's basket — 3 months
Both sides on one chart — 10 stocks on the bull side, 10 stocks on the bear side. Neither side is emphasised.
Each line tracks the 10 companies most exposed to this story, so a broad basket of loosely related names does not flatten it. The full list below is unchanged.
Jul 20, 2026 – Oct 19, 2026 · 66 sessions
Equal-weighted performance of the 10 most exposed stocks over the last 3 months.
No data for PSTG
Most affected stocks
Bull case
Palantir can monetize enterprise agent deployment through its data ontology and AI orchestration platforms rather than relying primarily on per-seat licenses.
The company supplies dense, liquid-cooled AI servers that convert accelerating chip demand into deployable compute clusters.
Owns both the model distribution and the cloud it runs on.
Oracle's cloud infrastructure and database businesses benefit when enterprise agents generate intensive compute and data-query workloads.
Increasingly complex AI processors and HBM stacks require more intensive semiconductor testing, supporting demand for Advantest's test systems.
Agentic applications generate data-intensive, variable workloads that can expand consumption of MongoDB's developer database platform.
HPE sells AI systems, networking and hybrid-cloud infrastructure to enterprises building internal agent platforms.
Dell benefits from enterprise demand for AI servers, storage and private infrastructure used to run agents near proprietary data.
Its custom accelerators, optical connectivity and data-center networking silicon support the infrastructure required to scale agent workloads.
Optical transceivers and related photonics give Coherent direct exposure to rising bandwidth requirements inside AI data centers.
Bear case
Generative agents can directly substitute for many writing, design, coding and marketing projects purchased through Fiverr's freelance marketplace.
As models automate routine knowledge work, clients may require fewer freelance labor hours and shift spending from human talent toward model and compute providers.
AI-assisted design and engineering could let smaller teams produce more output, weakening growth in named-user software seats.
AI sales and marketing agents can replace portions of the human teams that underpin HubSpot's seat-based expansion economics.
Agentic accounting systems could automate reconciliations and close-management tasks, reducing the number of finance users needing dedicated workflow seats.
Zoom remains heavily tied to per-user collaboration subscriptions that could face seat compression as agents handle more meetings, summaries, and workflows.
Its per-employee human-capital-management economics are exposed if agents reduce administrative staffing and overall knowledge-worker growth.
Generative agents can compress the number of creative and marketing users needed, pressuring Adobe's predominantly subscription-per-user model.
Per-user content collaboration pricing is vulnerable if machine agents increasingly retrieve and process enterprise files without equivalent paid seats.
General-purpose AI agents can replace portions of traditional rules-based robotic process automation and pressure UiPath's robot and user licensing model.
Reporting behind this story
Looking at what moved this story…