TOPSHOT – Banners of US synthetic intelligence firm OpenAI fly close to an set up depicting the legendary “Computer virus” constructed solely out of microelectronic circuit boards and different pc parts, exterior on the campus of Tel Aviv College in Tel Aviv on June 5, 2023. (Picture by JACK GUEZ / AFP) (Picture by JACK GUEZ/AFP by way of Getty Photos)
AFP by way of Getty Photos
Constructing software program has by no means been cheaper. Ninety-two percent of U.S. builders now use AI coding instruments day by day, and 46% of all new code is AI-generated. What that abundance has not touched is distribution; the price of getting a purchaser to belief, combine, and standardize in your product, which has at all times been the tougher drawback in enterprise software program.
Aaron Levie, co-founder and CEO of Field, described it on X: “The plurality of prices in most enterprise software program corporations is definitely on GTM, as a result of at scale most enterprise software program classes are robust to interrupt into and wish a heavy quantity of consultative promoting and help for implementation and integration of options.” AI, he argued, has not lowered that want. In lots of instances it has elevated it.
The enterprise capital market has guess closely on the construct aspect. Over $1 billion flowed into AI coding platforms in 2025 alone, and Replit raised $400 million at a $9 billion valuation in March 2026, tripling its valuation in six months. That capital displays a real perception: growth velocity has elevated. The query buyers are slower to reply is what occurs to the remainder of the unit economics when code turns into a commodity.
The reply is seen in SaaS benchmarks. GTM expenditures run 40% or more of income at high-growth cloud software program corporations, in response to a BCG survey of greater than 90 executives at mid-sized B2B software program companies. Enterprise gross sales cycles average six months or longer, and buyer acquisition prices for enterprise offers run 5 to 10 instances increased than the B2B SaaS median. Cheaper code technology doesn’t transfer any of these numbers.
Levie framed the dynamic as a conservation regulation: “In case you make one factor cheaper and extra ample (growth of software program) then the brand new drawback of discoverability and market differentiation (GTM) turns into the toughest half.” The provision aspect has by no means been extra liquid. App Retailer submissions jumped 84% 12 months over 12 months in Q1 2026, to 235,800 new apps in a single quarter. Apple expanded assessment instances from 24-48 hours to 7-30 days to manage. Extra software program means extra noise for each enterprise purchaser navigating a call.
That purchaser drawback compounds in regulated industries. Levie famous in a March 2026 roundup of conversations with greater than 20 enterprise AI and IT leaders that implementation timelines haven’t compressed with AI functionality positive factors. Enterprises nonetheless require safety critiques, integration with present information infrastructure, authorized sign-off, and alter administration earlier than any software program will get right into a manufacturing workflow. None of that has an AI shortcut.
The crowding on the utility layer additionally modifications what indicators patrons can use for differentiation. When 60% of new code is AI-generated and any founder can ship a working prototype in a weekend, product options are now not a dependable filter. Belief, distribution, reference clients, and the depth of implementation help develop into the differentiating variables; every of which requires time and relationships, not compute.
Karri Carlson, a product govt who replied to Levie’s put up, flagged a downstream danger that VCs haven’t totally priced: what occurs when AI brokers make software program procurement choices? Machines, she argued, will solely worth belief and style if these qualities could be quantified and in contrast. That situation will not be imminent, however the class of founders constructing for an agentic procurement future, the place discoverability requires machine-readable indicators fairly than relationship-driven ones, has an edge right now.
For buyers allocating to the applying layer, the implication is that GTM infrastructure is underpriced relative to dev tooling. Corporations which have cracked low-cost distribution; by means of PLG (product-led progress), embedded integrations, or community-driven adoption, carry sturdy benefits that AI can not replicate on the similar charge it replicates options. The moat in enterprise software program was by no means the code. It was at all times the belief required to run on another person’s infrastructure, and that value construction is, if something, going up.