AI Startups Will Thrive In Adversity
'Platform or die' and business model alignment will make AI spend robust
Software Synthesis analyses the evolution of software companies in the age of AI - from how they're built and scaled, to how they go to market and create enduring value. You can reach me at akash@earlybird.com.
If you’re a founder building in AI, join me on the 7th of May in London as I unpack how to build enterprise-native AI with Snowflake over breakfast at their office - sign up here.
The prospect of a global trade war and the unwinding of globalisation has left the public markets reeling. Businesses were bracing themselves for tariffs, but the magnitude and breadth was worse than expected.
Despite a paring back of some of the losses after the 90-day pause was announced, continued uncertainty is hurting business confidence. Capex allocation becomes particularly challenging in periods of volatile policy; capex investments that would take years to implement may prove unnecessary if the tariffs are just a negotiation for fairer trade deals.
The rationalisation of software spend is inevitable, too.
Morgan Stanley’s CIO survey indicated a measurable decline in IT spend forecasts for those surveyed after 02/20.
Prior to 02/20, 2025 budget growth expectations were in line with the long-term historical average of 4.1%. After 02/20, this declined materially to 3.2%.
As earnings season is now behind us, we can discern some worrying signs - what's more concerning than the number of companies that beat consensus estimates is the steep decline in forward guidance.
Companies with high international exposure are most vulnerable, but technology will acutely feel the second-order effects on sales cycles and pipeline conversion.
Platform or die
The set of companies most insulated from this impact are platforms/compound startups with multi-product suites. This would include names like Intuit (SMB back-office), Hubspot (front-office), Palo Alto Networks (cybersecurity), Microsoft (productivity) as just a few examples of companies that have had strong multi-product penetration in their install base (which normally results in high gross and net retention).
AI or die is a sound principal to follow in this new paradigm, but becoming a compound company may become just as important as macro headwinds threaten to constrain the enterprise value created by AI.
Infrastructure software companies have historically seen higher attach rates on new SKUs than application software, but the silos that may have previously inhibited app software cross-sell may fade away in the age of the data lakehouse.
Going too broad too soon runs counter to conventional startup advice to build with a narrow scope and find PMF, but as with all advice there’s nuance when it comes to platformisation. Expanding rapidly into adjacent workflows/agentic capabilities that you have a natural right to win will be important in the coming macro climate - think about the SKUs that should be unified into one platform to derive synergies. Looking at the horizontal AI agent platforms like Ema, Writer, Sana, Glean, and Dust, they all claim to serve multiple departments, which will be key to becoming the equivalent of the system of record for the AI age. Sequencing matters, though, and new SKUs where the buyer is the same is the path of least resistance.
Business model alignment
A lot has been said about the business model innovation of AI startups. Given that agents complete work at a comparable quality to that of a human, outcome-based pricing was supposed to have seen much higher uptake by now, so we were told.
In hindsight, we were just too early when it came to the timeline for agents to become production-ready. However, even when agents become ubiquitous, we probably won't see outcome-based pricing become the dominant pricing model.
As Dharmesh Shah suggested, the notion of results-as-a-service doesn’t work outside of a handful of domains where QA of outputs is objective. For most domains, the quality of the output is subjective, which is why a better alignment of value created and value capture is work-as-a-service.
The adoption of hybrid pricing models in software has been a growing secular trend for several years now, which is likely to continue being the case as the new cohorts of AI apps scale. Across each category of software there are AI companies that either sell work or significantly enhance the productivity of existing workers - over time, we may get to a point where the latter transforms into the former, but it’s hard to say when that will be.
In the interim, though, any component of consumption-based pricing represents a much better alignment of value with buyers who are tightening their belts. Take the below example of customer support, where growing consumption of AI software drives significant savings.
There’s ample evidence of giant software companies with large instal bases on seat-based contracts beginning to embrace consumption-based pricing.
As a macro climate tightens, it's important for founders to remember that the greatest companies are usually built in periods of adversity. This may well represent a unique moment where AI startups are well-placed to capitalise on these headwinds. By building compound startups faster and with business models that align with value creation, founders may well outperform through this turbulence.
After all, AI/ML budgets are the most durable budgets through these times. The real question is, will it be the incumbents or startups who will capture those budgets?
Additional Reads and Content
Larry Fink’s 2025 Annual Chairman’s Letter to Investors
The Visions of Neil Mehta by Colossus
Scarcity and Abundance in 2025 by Alex Danco
The case against conversational interfaces by Julian Lehr
Reflexive AI usage is now a baseline expectation at Shopify
Jobs
Companies in my network are actively looking for talent:
An AI startup founded by repeat unicorn founders and researchers from Meta/Google building 3D foundation models is looking for a 3D Research Engineer, Research Scientist and ML Training / Inference Infrastructure Engineer (London or Munich).
A seed-stage vertical AI company building in accounting is looking for Founding Product Engineers (London).
A seed-stage AI code review startup is looking for a Founding Python Engineer (London).
Reach out to me at akash@earlybird.com if you or someone you know is a fit! If you're looking for your next role, click below to share your details and join the Talent Network.
Not for Everyone. But maybe for you and your patrons?
Dear Akash,
I hope this finds you in a rare pocket of stillness.
We hold deep respect for what you've built—and for how.
We’ve just opened the door to something we’ve been quietly handcrafting for years.
Not for mass markets. Not for scale. But for memory and reflection.
Not designed to perform. Designed to endure.
It’s called The Silent Treasury.
A sanctuary where truth, judgment, and consciousness are kept like firewood—dry, sacred, and meant for long winters.
Where trust, vision, patience, and stewardship are treated as capital—more rare, perhaps, than liquidity itself.
The two inaugural pieces speak to a quiet truth we've long engaged with:
1. Why we quietly crave for signals from rare, niche sanctuaries—especially when judgment must be clear.
2. Why many modern investment ecosystems (PE, VC, Hedge, ALT, SPAC, rollups) fracture before they root.
These are not short, nor designed for virality.
They are multi-sensory, slow experiences—built to last.
If this speaks to something you've always felt but rarely seen expressed,
perhaps these works belong in your world.
Both publication links are enclosed, should you choose to enter.
https://tinyurl.com/The-Silent-Treasury-1
https://tinyurl.com/The-Silent-Treasury-2
Warmly,
The Silent Treasury
Sanctuary for strategy, judgment, and elevated consciousness