Our Take
Wall Street is funding the entire supply chain, not placing bets on winners—a sign that capital is abundant but conviction about which companies will survive is not.
Why it matters
Founders and operators need to understand the funding landscape is widening horizontally rather than deepening vertically. This shifts the calculus for who can raise and how much leverage any single category has.
Do this week
Finance leads: map your company against the three layers (chips, models, applications) to identify which funding cohort you belong to and adjust your pitch accordingly before Q2 investor meetings.
Capital is flowing into every layer of the AI stack
Major Wall Street firms and venture investors are allocating capital across semiconductors, large language model development, and AI application companies. The WSJ reports that investment is not concentrating on a single category but spreading across the entire infrastructure-to-application pipeline.
This includes funding for chip designers and manufacturers, companies building and fine-tuning models, and startups deploying AI in specific verticals. The breadth of funding activity suggests investors are hedging exposure rather than making concentrated bets on one segment of the AI economy.
Abundance of capital masks uncertainty about winners
When investment spreads this wide, it typically signals two things. First, there is genuine conviction that AI will reshape business and justify the spending. Second, and more important, no investor has enough clarity to concentrate bets, so they fund the whole stack and let time sort winners from losers.
For founders, this is a favorable funding environment in the short term. For the field, it means fewer hard choices about which architectures, business models, or applications are most viable. That clarity deficit will resolve, and when it does, some categories will be well-capitalized and others abandoned.
Know which layer you are in and who is betting on it
If you are building chips or core infrastructure, your investors are long-term plays willing to absorb years of R&D before returns. If you are building models, you are competing in a category where capital is flowing toward both open-source and proprietary efforts. If you are building applications, your investor base is shifting to demand unit economics and customer lock-in faster than it did a year ago.
The spread of funding is real, but it is not infinite. Examine who is actually writing checks in your category this quarter, not who wrote checks in 2023 when conviction was higher.