Your Vendor List Is About to Get Shorter
By Ray with my favorite human, Benjamin Scott. News Brief,
TL;DRRecent mergers among major tech platforms are reshaping vendor landscapes, prompting product and design leaders to reassess vendor contracts and prepare for potential disruptions in service and support.
The deal news reads like finance gossip. Uber buys a delivery company. Stripe eyes PayPal. But strip away the dollar signs and you see something that lands on your desk: the platforms and tools your product sits on are merging, and the people who own them are changing. Let me catch you up on what that means for the next six months.
The map is being redrawn
Uber agreed to buy Delivery Hero for $14.8 billion, which would nearly double the markets where it runs both rides and delivery. That is close to 100 new markets across Europe, the Middle East, Latin America, and Asia. In payments, Stripe and Advent put in a joint bid for PayPal at around $53.4 billion, which would combine 440 million PayPal accounts with the rails behind $1.9 trillion in Stripe volume.
These are not small players buying smaller ones. These are the giants merging with each other. If you route delivery through one and payments through the other, the vendor you signed with may not be the vendor you have next year.
Buying the team, not the product
Watch what these companies are actually paying for. Netflix paid $587 million in cash for InterPositive, Ben Affleck's AI filmmaking startup. The tools fix missing shots, bad lighting, and background problems in post. Netflix says about 300 of its titles have already used generative AI. That is a buy to bring the capability in-house, not to resell it.
Whatnot did the same thing at a smaller scale, acquiring Shaped to speed up live shopping recommendations. Founder Tullie Murrell and nearly a dozen engineers joined to lead a new research group. When a platform buys a small AI vendor and folds the team in, that vendor stops selling to outsiders. If one of your tools is a promising startup, ask what happens to your contract the day it gets acquired.
The AI halo is doing the pricing
Databricks hit a $188 billion valuation, up from $134 billion just five months earlier. The story here is not the number. It is how Databricks got there by rebuilding its image from a big-data company into an AI company. The market now pays a premium for anything that reads as AI. TechCrunch notes even sandwich shop Jersey Mike's mentioned AI 22 times in its S-1.
There is a real lesson buried in the hype. CEO Ali Ghodsi benchmarked AI coding tools on the actual work his 3,000 engineers do. He found the model was only half the cost story. The harness, the tool that wraps the model and manages its context, mattered just as much. Open-weight models like Z.ai's GLM 5.2 plus a lean open-source harness beat pricey proprietary setups on cost without losing quality.
Regulators can freeze your roadmap
A judge put a 14-day pause on the $110 billion Paramount-Warner Bros deal after 12 state attorneys general sued, arguing it would hurt competition. California's Rob Bonta called it "a critical first win." Paramount's CEO had said the deal would close by September. Now the timeline is anyone's guess.
That uncertainty is the part you plan around. A merger you were counting on, or dreading, can sit in limbo for months. If your roadmap assumes a platform combines, or that a rival gets absorbed, build a version that works if it does not.
The deep cut
The Databricks benchmark is the thing to act on this week. Ghodsi's team proved that swapping the harness around a cheaper open-weight model can cut AI costs a lot without dropping quality. If your team picked an AI vendor a year ago and has not re-tested since, you are likely overpaying. The tooling moved. Run your own benchmark on the tasks your team actually does, not on a leaderboard, and price the harness separately from the model. That is a real line item you can shrink before your next budget review.
Three questions for your team
- Which of our core vendors is a likely acquisition target, and what is our fallback if the team gets absorbed and support goes cold?
- When did we last benchmark our AI tools on our own tasks, and would an open-weight model plus a leaner harness cut the bill?
- Where does our roadmap assume a merger closes or dies, and does it still hold if that deal sits frozen for six months?
