What does the future hold for the AI economy? Read the latest research from Lak Ananth, global managing partner at N47, and Jonathan Goldberg, founder of D2D Advisory.
The premise is simple: Ubiquitous, on-demand AI is in view, but we won’t get there with the tech stack we have today.
The technology works. The demand is real. The capital is flowing. The harder question is what must change in the economics, architecture, and value chain for Intelligence on Tap to become a sustainable, long-term reality.
The founding parallel
Every transformative technology reaches a moment when its initial infrastructure cannot support its ultimate ambition as the initial tech stack simply won’t scale. Google Search could not have existed on the enterprise technology stack of its era: proprietary servers, commercial databases licensed at $47,500 per processor, and networking equipment with 65% gross margins.
Google’s founders didn’t question whether search was real. They reinvented the stack: Commodity hardware designed to fail, software that assumed failure and routed around it, and warehouse-scale architectures optimized for aggregate cost. The result was a 20-fold cost advantage that made ad-supported search viable.
The opportunity ahead
AI sits at an analogous moment. AI queries cost 1 cent to 10 cents to serve versus 0.2 cents to 0.3 cents for search while generating comparable revenue per interaction. The dominant chip supplier extracts 73% to 75% gross margins. Hardware becomes obsolete in two to four years but is depreciated over five to six.
These are engineering problems, not existential ones.
We took a hard-nosed look at every layer of the AI cost structure, tested whether current economics match the value being created and mapped the specific shifts required to make Intelligence on Tap a durable reality.
Here’s what we found in our research:
1. Value will accrue where it always has: stack reinvention.
The AI industry’s $5 to $9 trillion infrastructure buildout through 2030 is not a bubble. It is the same kind of capital cycle that built the electric grid, the telephone network, and the cloud. The returns will go to the ones who reinvent the stack.
2. Companies that work through the AI-cost stack methodically will win in the end.
The companies that understand exactly where their dollars are going are the ones positioned to compress costs where compression is possible and to avoid spending capital chasing leverage that does not exist.
The cost stack, when decomposed, looks like Google’s did: multiple discrete layers, each solvable, none solved all at once. The companies that worked through those layers methodically—Google with servers, then networking, then silicon—built the most valuable infrastructure businesses in history. Intelligence on Tap requires working through the same layers in AI.
3. Both the vertical architecture model and the distributed stack will survive. They won’t both win.
Large AI labs today are building vertically integrated systems around their models and offering them along with deployment expertise, similar to the “walled garden” model of early enterprise software providers. Distributed alternatives based on low-cost, open-source projects are also beginning to thrive as we saw recently with the rapid growth of OpenClaw.
Both architectures will coexist for a long time. The vertical model will retain strong positions in regulated industries, large enterprise deployments, and use cases in which integration simplicity justifies a premium. The distributed stack will expand fastest at the developer layer, in cost-sensitive markets, and in any application where data sovereignty is nonnegotiable.
The historical pattern across comparable transitions is that distributed architectures eventually capture the larger share of long-run value by expanding the total market. The pipes carry more water when they are cheap and open. The question for every investor and builder in this space is: Are you positioned for the architecture that maximizes margin today or the one that maximizes volume tomorrow?
4. The token is the wrong unit of value.
The AI industry is caught in a trap of its own construction. It built pricing around the token because the token is what the hardware counts. It is a convenient unit of measurement that has almost no relationship to the unit of value created.
The path out runs through two layers: 1) Routing solves the supply side—matching each query to the cheapest model that can answer it. 2) Segmentation solves the demand side—matching each customer to a price that reflects what the answer is worth.
The routing layer is being built now by companies such as Martian and Unify.ai. The segmentation layer—the infrastructure that prices intelligence by what it is worth to the buyer—is where the next decade of AI value will be captured.
5. Founders must build for a flexible future.
The AI infrastructure buildout underway is real and necessary. The question for capital allocators is not whether to invest—it is whether the infrastructure being built today is optimized for the architecture of the next three years or the next 10.
History is consistent on this point: The companies that assumed their dominant architecture was permanent did not lead the transition that followed. The ones that built for flexibility captured the value.
The winning bet is not the best engine. It is the most flexible chassis.
6. Infrastructure buildouts enable application-layer dominance.
The AI transition to commodity economics is not yet complete. The transition to commodity-level pricing is likely to be completed by 2027 to 2028, and then the value stack will invert.
Chips, models, and inference compute will become competitive inputs rather than sources of monopoly margin. The prize will shift to the application layer: The companies that used cheap, reliable, widely available AI to build durable positions in industries that are only beginning to adopt it.
This is not an argument against infrastructure investment. No application-layer value can materialize without the buildout that precedes it. The investment question is not only whether to fund AI infrastructure but also whether the portfolio is positioned for the value inversion that follows.
The future of Intelligence on Tap
Intelligence on Tap will be a reality when a doctor in a rural clinic, a small business owner without an IT department, and a student in an underfunded school district all have access to the same quality of intelligence that a Fortune 500 company deploys today.
The technology to deliver that is already in development. The economics are moving in the right direction at an unprecedented pace. The value question is who builds what runs on top of it.
Every technology cycle asks the same question. The answer, across railroads and semiconductors and the internet, has been consistent. The question for every investor in this space is the one that mattered in 1880, and in 1990, and in 2000: Are you funding the pipes or the water?
Read the full report at N47.com.
Note: This content was created by N47.