John O'Shea at https://xitter.cf/t.co/DbtC5fCiSX.
Bengaluru, India
Joined May 2009
- Tweets1.2K
- Following689
- Followers744
- Likes1.1K
Reminds me of how Ethereum bag holders invested in DeFi apps. Fat protocol thesis...
I have the highest regard for you Gokul. But a crazy set of problems don't translate from AI native shops to the rest of the world. e.g. Accounting Ops are quite different for s/w companies and retail/mfg shops. One doesn't get you the other. [We have scars on both btw]
The New Kingmakers
It's become clear that revenue scale no longer matters - or is correlated - with raising a round. It's important but not sufficient.
Besides team caliber, what ACTUALLY matters to fundraising is how many top AI-native companies use your products, love your products, embrace your products and profess undying loyalty to you (i.e. say they won't replace it at the drop of a hat).
Why?
AI-native companies are the hardest customers to win. They have the engineers to build it themselves. They have the taste to know when something is mediocre. They try every new tool the week it launches and rip it out the week after. If one of them pays you and stays, your product survived the toughest evaluation there is. Investors know this.
Revenue can be bought. Pilots, services, discounts, a heroic enterprise sales team. ARR built from Fortune 500 pilots proves you can sell. The product might still be mediocre. Loyalty from a kingmaker can't be manufactured.
These companies are the leading indicator. Investors are underwriting where the market will be in 5 years, and the AI-native companies are already living there. Their workflows today are everyone else's workflows in 2029. If they picked you, you've been chosen by the buyers everyone else will copy.
The last reason is distribution. Kingmakers talk. Their engineers post what they use. Their founders compare notes. One of them begets the next three.
Who
There are probably ~50 AI-native companies that are perceived as kingmakers for startups. I've seen $50M+ rounds raised on the backs of just having one of these as a customer. (not pilot, but actual paying customer using the product as a core part of their infra and workflows).
Meanwhile, I see early-stage founders spending 12 months chasing CVS, Walmart, and Pfizer. Long procurement cycles, security reviews, a pilot that never converts. Even when it lands, the logo proves you can survive procurement. That's the wrong thing to prove in 2026.
If you're an early stage enterprise/B2B AI company, Instead of targeting CVS, Walmart, Pfizer, etc, go after the companies on Forbes' AI 50 list.
The new milestone is kingmaker logos. That's what your next round will be priced on.
Us accounting focused plebs will call this the internal audit function.
evals will be the primary moat for every enterprise long term.
Abridge CEO @ShivdevRao dropped an amazing article about this on how enterprises are approaching product evals (link in thread).
many AI product teams are already committing $10M+ annually to human expert evals, because you can't improve what you can't continuously measure.
our bet is that within 12 months, every Fortune 500 building AI products will have an 8-figure evaluation stack.
the very early days of the data exponential is now.
If I am building an AI App, my incentive is to spend the least in token cost on COGS not the most. And be generous with R&D spends [i.e. token budgets for teams]. Also, ARR is also a terrible metric - it forces one to underwrite the buyer's business' and hence become a quasi-VC.
Some solid Q&A on seat vs usage vs outcome pricing for AI. Clearly, no perfect answers but it def made me fine-tune my POV. Views:
> marginal cost of s/w = 0; marginal cost of AI is not. That alone implies you HAVE to find SOME way to correlate price to usage/COGS.
> "but AI labs all do tiered pricing" - actually all AI-lab rev is token metered which is why every tier has rate limits, caps, model gating, etc. Labs are actually the strongest case _for_ token/usage pricing.
> variable cost / task, but customers want predictability; agree - but true only for tasks that have median & mode usage within striking distance; in AI, power users can use 1000x median and seat pricing on variable COGS is a disaster. Needs to be _very_ carefully planned.
> I recall "runaway" usage as a huge issue when we launch bigQuery at GCP; knee-jerk reaction was to limit use; what made BQ a $10B biz was still usage pricing but paired with account budgets, threshold alerts, dry-run estimates. That's a better approach (but harder for long-horizon / unbounded tasks I'd admit)
> "why not do outcome based pricing" - imho, this _is_ usage/token based pricing, just metered at a higher order unit. It works best, again, when median vs mode are similar. e.g. say sierra charges 2$ / resolution, and resolution takes 2min (cost = 0.2$) or 2hrs ($12), the lose on the 2hr call (v low vol) and make money on the 2min call (v high vol). Unbounded agentic tasks need to be metered on usage.
My best guess is that the real uncapped promise & value in AI is on agentic tasks. There is this idea of agentic frontier where if you are more human-assist / co-pilot, you'd end up preferring seat-based but over time as you start solving agentic problems, which is the true value of AI lies, you'd end up above the frontier and shift to usage based. The more unbounded & agentic the problem, the higher the preference for usage pricing. Harvey may be sitting below the frontier today, but will eventually move above.
IT budgets of medium growth firms like BFSI are $1T+ and they don't operate on an ARR model - they hate variable costs. So much of this ARR + token spend is indistinguishable from crypto circle jerk
The same will happen to accounting. Everyone becomes an auditor/reviewer. And we get to real time close. ERPs become irrelevant along the way.
This is a likely model of how SAP and co gets screwed over by AI shops and data lakes. Since firms invest heavily in piping all data from sap into snowflake. ERPs _were_ the OG mothership which ran the enterprise. Not anymore.
Arun retweeted
By 2027 any IT services co without substantial same-day/next-day delivery offerings is ngmi.
Selling a stock or not buying one are also important investment decisions
PMs don’t just ship features. They kill them. Shipping isn’t the job. Shipping the right product is. A great PM doesn’t fall in love with the roadmap. They fall in love with the problem and have the guts to say: This isn’t solving it. This adds complexity. This doesn’t matter. Every feature, setting, UI, element should fight to exist.
At Nest, we had one rule:
If you can’t explain why it matters, it doesn’t ship.
You had to tell us the why.
The reason a real person would care.
That one rule killed dozens of features.
Infosys also recently tied up with Claude. These firms' distribution is immensely valuable for OpenAI/Anthropic to hit their near-term revenue goals.
OPENAI TEAMS WITH CONSULTING GIANTS TO EXPAND ENTERPRISE AI: CNBC
OpenAI has launched multi-year “Frontier Alliances” with Accenture, Boston Consulting Group, Capgemini, and McKinsey to deploy its enterprise AI platform, Frontier.
The partnerships aim to accelerate AI adoption by combining OpenAI’s technology with consulting expertise, helping companies integrate AI agents into workflows faster. Frontier connects organizational systems and data, enabling easier AI management and deployment.
OpenAI’s CFO says enterprises now account for ~40% of its business, expected to reach 50% by year-end. Consulting partners will build certified teams, offer strategic guidance, and work alongside OpenAI engineers to scale AI across clients.
From 2002 - 2009, newspapers stocks went down in a straight line bc the market foresaw the disruption of the internet yet earnings continued growing for 5 straight years. until they didn't.
