Good Enough AI Already Won the Cost War

The AI cost war ran for about two years, and it's already over. "Good enough" models won it. Enterprise blended token costs fell from $18.40 to $6.07 per million tokens between January and April 2026, drawn from 2.4 billion API calls in AI.cc's 2026 AI API Infrastructure Report. Open-weight models captured 38% of that enterprise volume, up from 11% a year earlier. DeepSeek's V4-Flash matches a paid flagship's intelligence score at a fraction of its listed price, according to an August 2026 model-pricing comparison. Even Microsoft's own AI division has leaned its public strategy toward "good enough" capability priced low, ahead of chasing the top of any benchmark chart, per an industry analysis published in July 2026.

1. 1Token cost: $18.40 to $6.07 per million tokens, Jan–Apr 2026 (AI.cc).
2. 2Open-weight share: 11% to 38% of enterprise token volume in one year (AI.cc).

"Good enough" is doing what the label promises, mostly. Models that needed a six-figure contract two years ago now run for lunch money, and what's left over sits somewhere between irrelevant and still the whole game for the hardest slice of tasks, which is a less satisfying answer than either side of that argument wants to give you. Set that aside for a second, because it isn't the interesting part.

Affordable Didn't Make It Manageable

A few weeks ago I wrote about a friend of mine at a European marketing agency's China office. Her agency's French headquarters signed a partnership with Anthropic and handed her office a token budget, generous on paper, tight in practice. She burned through it fast enough that she started weighing a personal account just to cover the shortfall. I didn't resolve that story then, and I'm not resolving it now, because it was never about her, or her employer, or which vendor happened to sign the contract. It's one visible version of something that's everywhere: roughly 47% of GenAI users access their tools through personal, unmanaged accounts, bypassing whatever controls their employer thought it had, according to Netskope's 2026 telemetry.

Her office had a deal. By the time it mattered, price wasn't the thing deciding whether the budget held.

The Model Was Never the Interesting Part

The model is infrastructure. The system around it is the product. I made that argument building The Seven Dimensions of Enterprise AI, and "good enough" models priced this low make it truer than ever. Once every serious option clears the same floor of capability at close to the same low price, the model stops being a decision worth making. Nobody wins by picking the model anymore, the way nobody wins a coin flip by calling heads with more conviction. The competition moved to what gets built around that choice, and it moved before most companies noticed the game had changed.

There's No Single Fix, and That's Not a Cop-Out

Ask five people what a company should do about this and several of the five answers will be right, for different companies, at the same time, which is an uncomfortable thing to hear when you wanted one clean recommendation. Deployment support is a real answer: it's the same instinct behind the forward-deployed-engineer hiring surge I covered a few weeks ago, postings up more than 800% year over year by early 2026, per SkillScouter. Visibility tooling is a real answer too. An entire product category now exists to detect this kind of unmanaged usage, built because nobody could see it happening until it already had. Training helps somewhere. Governance helps somewhere else. None of them alone closes the distance, and a company that picks one and calls the problem solved has chosen a symptom to treat and is hoping it picked the right one.

Autonomous driving runs into the same shape of problem at a different scale. A perception agent, a prediction agent, a planning agent, and a control agent each own one piece of a moving car, and none of them drives it alone. Get the handoff between them wrong by fifty milliseconds and the car brakes for a pedestrian who already crossed. The car works because four narrower jobs each get done right, in relation to the others.

That's the case for building across all seven dimensions of a system. One lever never covers enough ground, and that's why the framework runs seven entries deep. Data Architecture, Domain Logic, Orchestration, Models & Agents, Governance & Containment, Human Judgment, Calibration: each one covers a different way a system, or a company's use of one, can fail, and no downstream dimension compensates for a decision made poorly further up.

TikSense is where I've built this. Two examples, chosen because they sit on different fronts. Neither alone would have been enough.

The first is a business call made before any model runs. A routing step decides whether a creator counts as proven, emerging, or aspirational, meaning commercial intent without an actual sales record behind it. A person decided what counts as a track record, encoded it once, and it shapes every downstream analysis before the analysis starts, the least glamorous afternoon of the whole build and also one of the only decisions that's never needed revisiting since.

The second is an authority boundary, decided before anything runs. TikSense runs two tiers of agents. The first tier classifies and scores; none of those agents can resolve a conflict between their own outputs. A second, senior tier activates only on three defined conditions: two tracks disagreeing past a threshold, confidence dropping below floor on a high-potential creator, or a data point in what's scraped that falls outside every track's scoring domain. It resolves what it can itself. A human only sees what that tier can't authorize, which by design stays rare. That's the Agentic Authority Escalation Model at work on live creator data, past the diagram stage.

Neither of those is the whole fix, on purpose. The system is where all of this lives, and a system doesn't get built one afternoon and one lever at a time.

Somewhere between a low-cost model and an outcome a seller can use, someone has to own the whole picture. Patching each piece on its own doesn't add up to one. Most companies haven't decided that's one job yet, and most haven't decided whose job it would be either. The AI cost war is over. "Good enough" models won it, and it turns out that was never the expensive part.

This lives wherever you already are, Substack, X, or LinkedIn, drawn from building TikSense.