Wall Street spent this week arguing about oil, Treasury yields near multi-decade highs, and which AI headline deserved the microphone. The S&P 500 closed Friday near 7,743. The Nasdaq held near record territory. Indexes can look calm while the real fight is underneath.
My take is simple. Three seats. No targets. No calendar guesses. One winner with real scarcity, one loser still fighting the wrong product story, and one dark horse most investors still treat like a side story.
The winner: Micron (MU)
If you only own the idea of artificial intelligence, you own a slogan. If you own the bottleneck every accelerator needs, you own the factory floor.
Micron Technology (MU) is my winner because high-bandwidth memory stopped behaving like a commodity cycle and started behaving like contracted infrastructure. The company has pointed to multi-year customer deals locking supply, customer cash committed to secure chips, and HBM capacity that cannot meet full demand. Memory prices rose hard as AI racks pulled capacity away from phones and PCs. That is ugly for gadget makers and wonderful for the people selling the scarce stack next to the GPU.
I like this story when the product is physical and prepaid. Software can reprice overnight. A sold-out memory line with deposits and long agreements is harder to wish away. Micron already ran a monster re-rating this year, so nobody gets a free entry gift. What would change my mind is a real demand walk-back from the big AI buyers, or new capacity that shows up earlier than the industry admits. Until then, the winner seat is still the chip every AI box has to buy before the model can run.
The loser: HubSpot (HUBS)
Software did not die. Seat software got put on trial.
HubSpot (HUBS) is my loser face for the SaaSpocalypse that still hangs over the sector. The fear is not complicated. If agentic AI can run marketing workflows, draft campaigns, and stitch customer data without a human clicking through a dozen tabs, the old per-seat subscription machine looks soft. HubSpot has been one of the harder-hit names in that tape, with a brutal slide from last year’s highs while investors argue whether AI replaces the tool or sits on top of it.
Plenty of these companies still grow revenue. That is what makes the seat painful. The market stopped paying for the old multiple on the old story. Buybacks and AI feature decks help the press release. They do not restore the belief that every knowledge worker needs another license forever. Salesforce (CRM) is the giant version of the same fight. HubSpot is cleaner here because the damage is obvious and the product sits in the workflow lane agents are hunting.
What would flip the seat is proof that AI agents raise seat demand instead of capping it. Until the numbers force that rewrite, seat software remains the loser theme inside a market that still loves AI.
The dark horse: GE Vernova (GEV)
Everybody can spell Nvidia (NVDA). Fewer people want to talk about turbines, transformers, and the multi-year wait for juice.
GE Vernova (GEV) is my dark horse because the next AI constraint is electricity and grid gear, not another press release about model parameters. Data centers do not run on vibes. They run on power that has to be generated, moved, and stabilized. Vernova sells the heavy equipment side of that problem, and the order book for gas generation and electrification has been quiet proof that utilities and hyperscalers are racing physical limits.
This seat is a dark horse for investors who still think the only AI stocks are chip designers and chat apps. Power equipment is slower and less glamorous. It is also the part of the buildout you cannot fake with a software update. Risks are real: project delays, policy fights, valuation after a big run, and any pause in data-center spending. What would break the thesis is AI campuses that stop signing load.
I still like the under-counted angle. Memory won the scarcity trade first. Software is paying for fear of replacement. Power is the seat people mention last and need first.
Bottom Line
Right now, my market map is simple. Micron is the winner on prepaid memory scarcity. HubSpot is the loser face of seat software under agent pressure. GE Vernova is the dark horse because the AI boom eventually collides with turbines, wires, and watts.
Comments
Log in or sign up to join the conversation.