Iconiq’s State of Software in 2025: Forward Deployed Engineers Up 12x; 55% of GTM is Now in Post-Sales; 2025 Is Crushing 2024 For Funding

The software business has entered its “well, that escalated quickly” era.

ICONIQ’s State of Software in 2025 does not read like a routine benchmark deck. It reads like a field report from a market that has quietly swapped out its operating system. The old SaaS playbook still matters, but it no longer explains the whole movie. Efficient growth matters more than vanity growth. AI-native companies are moving faster with smaller teams. Go-to-market is becoming more technical. Post-sales is no longer the nice, polite department that shows up after the real work is done. In many cases, post-sales is the real work.

That is why three headlines from the report hit so hard. First, forward-deployed engineer hiring has surged 12x. Second, the center of gravity in modern GTM has shifted away from pure selling and toward implementation, adoption, expansion, and customer outcomes. Third, 2025 funding conditions, especially in AI, are materially stronger than 2024. Put those together and you get the clearest signal in the report: software is not merely adding AI features. It is rebuilding the org chart, the funnel, and the financing math around them.

And yes, that means your favorite 2019 SaaS org chart may now belong in a museum. Right next to “growth at all costs” and company all-hands with five separate swag announcements.

The Real Story in ICONIQ’s 2025 Report: The Software Playbook Has Been Rewritten

ICONIQ’s broader thesis is that software has moved from survival mode into redesign mode. The last two years forced companies to get leaner, care more about Rule of 40 math, and stop pretending every extra headcount request was a love letter from the future. In 2025, that discipline is still there, but it is now paired with renewed appetite for growth.

That combination matters. Companies are not simply cutting. They are reallocating. They are using AI to automate internal work, pushing more work into software-assisted workflows, and putting more talent closer to the customer value moment. That is why ARR per employee is improving, OpEx per employee looks more stable, and burn multiples are getting healthier. The point is not that software suddenly became cheap. It is that the best operators are getting more output from each dollar and each person.

ICONIQ’s data also suggests this is not a niche behavior among a few flashy AI startups with suspiciously energetic founders on podcasts. It is showing up across planning, hiring, GTM design, and internal tooling. AI adoption rose as a strategic priority. Internal AI use moved from experiment to operating assumption. And the companies moving earliest are starting to show better sales efficiency and better R&D productivity.

Why Forward-Deployed Engineers Jumped 12x

The forward-deployed engineer, or FDE, is the breakout character of this story.

For years, software companies hired sales engineers, solutions architects, customer success managers, implementation consultants, and support specialists. Those roles still exist. But AI products have created a new hybrid need: someone technical enough to shape the product inside the customer environment, customer-facing enough to understand messy real-world workflows, and fast enough to translate usage into product iteration.

That is where the FDE comes in. Think of the role as part engineer, part translator, part implementation quarterback, and part “let’s fix this before the renewal call gets awkward.” When a product needs deep workflow integration, fine-tuning, training, customization, or hands-on deployment guidance, the FDE becomes the bridge between product capability and customer outcomes.

This is why a 12x jump in hiring matters so much. It is not just a talent trend. It is evidence that AI software is sold differently. Traditional SaaS often sold a destination: buy the platform, roll it out, train people, and value will arrive on a reasonable but vague timeline. AI software is increasingly sold on immediate proof. It has to work in the customer’s data, environment, workflow, and political reality. That makes technical onboarding and continuous adaptation far more important than the old “closed-won and tossed over the fence” motion.

Why AI Products Need Customer-Facing Technical Talent

AI products are unusually sensitive to context. They live or die by data quality, permissions, process design, model behavior, user trust, and integration depth. A sleek product demo means very little if the system trips over a customer’s CRM mess, security policy, internal taxonomy, or six layers of procurement drama.

That is why customer-facing technical talent is becoming strategic, not supportive. The product is not fully “sold” until it is embedded, trusted, and producing measurable outcomes. In that world, technical implementation is revenue work.

Seen this way, the rise of FDEs is not surprising at all. It is the job title that appears when software stops being a static tool and starts acting more like a live system that learns, adapts, and must prove itself in production.

Why GTM Is Moving Into Post-Sales

The headline that “55% of GTM is now in post-sales” is a little dramatic, but the underlying idea is absolutely real: modern go-to-market is moving downstream.

The better interpretation is this: high-growth AI companies are shifting meaningful GTM weight away from pure front-end selling and toward post-sales, technical adoption, onboarding, implementation, and expansion motions. In older SaaS models, sales dominated the org chart because the product could often be standardized and expansion followed procurement. In AI-era software, value often appears only after deployment, tuning, and habit formation. That means the revenue engine no longer stops at signature.

Post-sales is where time-to-value is won or lost. It is where customers decide whether the workflow actually fits. It is where usage becomes habit, and habit becomes expansion. And in subscription software, especially AI-heavy enterprise software, expansion is not a side dish. It is the main course pretending to be a side dish.

The Old Funnel Sold Promise. The New Funnel Sells Outcomes.

That shift is also visible outside ICONIQ’s report. Customer success and renewal leaders have been saying for years that retention and expansion deserve more strategic attention, but AI finally gave that argument teeth. Gainsight has highlighted how much SaaS revenue now comes from expansion, while services and renewal research from TSIA points to growth coming from managed services, professional services, and ongoing customer value delivery. In plain English: once the product gets smarter, the customer journey gets less transactional and more operational.

McKinsey’s 2025 B2B work reinforces the same theme from another angle. More companies are actively implementing or evaluating generative AI in selling, but the practical gains show up across the whole deal cycle, not just in prospecting. AI improves discovery, recommendations, data processing, workflow orchestration, and sales productivity. That makes GTM a system, not a department.

So yes, sales still matters. A lot. But sales without activation increasingly looks like ordering a fancy gym membership and never entering the building. Technically impressive. Financially unhelpful.

2025 Funding Really Is Stronger Than 2024, With One Giant Asterisk

On funding, ICONIQ’s takeaway is directionally correct and important: 2025 has been much stronger than 2024 for AI capital formation. Private AI and machine learning funding in the first half of 2025 already surpassed the full-year total for 2024 in ICONIQ’s cited PitchBook-based view, and average deal size jumped sharply. Crunchbase also showed a much stronger venture backdrop in 2025, with Q1 up sharply year over year and the first half marking the strongest stretch for startup funding since the first half of 2022.

That is the bullish side. Here is the asterisk: the recovery is real, but it is not evenly distributed. Capital is concentrating around AI leaders, giant rounds, and companies that can show unusually clear adoption or growth. Carta’s 2025 market data tells a similar story: more money is flowing, but liquidity remains complicated and the market is still selective. Bessemer’s Cloud 100 benchmarks show valuation strength returning, especially where AI growth is visible, but that does not mean every software company suddenly got its 2021 swagger back.

So when people say 2025 is crushing 2024 for funding, the most accurate translation is: “If you are in or near the AI wave, the market looks dramatically better. If you are not, please continue to keep your spreadsheet open.”

What the Funding Rebound Actually Means

The funding rebound matters for more than investor mood. It changes operator behavior. Boards are more willing to fund growth if growth is paired with believable efficiency. Founders are more willing to spend if product pull is obvious. Recruiting gets easier when employees believe the category has momentum. Buyers become more comfortable betting on younger vendors when they see the ecosystem backing them.

In other words, funding does not just finance software companies. It changes the confidence level of the whole market around them.

Smaller Teams, Better Productivity, Sharper Economics

One of the most striking themes in the report is that the best companies are doing more with less. AI-native companies are reaching meaningful scale with dramatically leaner teams than traditional SaaS benchmarks. That does not mean headcount is irrelevant. It means headcount is being deployed differently.

ICONIQ’s data suggests ARR per FTE is improving while annualized OpEx per FTE remains comparatively stable. Burn multiples are improving, too. Internal AI adoption is no longer fringe behavior; most surveyed companies are actively experimenting with or adopting AI-powered internal tools. Offshore headcount is also rising, especially in engineering and support, which points to a second layer of efficiency: software leverage plus global labor leverage.

This is the part that many people miss. The new operating model is not “replace everyone with AI.” It is “redesign the company so humans sit closer to high-value judgment, customer context, and technical problem-solving.” That is why the winning org charts are not simply smaller. They are more concentrated around product, engineering, customer outcomes, and capital-efficient growth.

What Founders and GTM Leaders Should Do With This

1. Stop treating post-sales like overhead.

If your product needs integration, behavioral change, workflow redesign, or customer education to unlock value, post-sales is part of revenue architecture. Staff it like that.

2. Hire more technical closers, not just better talkers.

The line between product, sales engineering, implementation, and success is blurring. Build teams that can win trust in live customer environments, not just in polished demos.

3. Measure time-to-value as seriously as pipeline creation.

Pipeline without activation is just expensive optimism. In AI software, adoption velocity can be more important than the initial contract headline.

4. Treat AI efficiency as an operating discipline, not a marketing slogan.

The companies pulling away are using AI internally, improving R&D productivity, tightening costs, and pairing growth with accountability. “We’re exploring AI” is now the corporate equivalent of “I meant to go to the gym.” Nice sentiment. Limited outcome.

Field Notes From the Market: What This Feels Like in Real Life

Here is what this shift looks like when you step away from the charts and into the actual day-to-day rhythm of software companies in 2025.

A founder who used to think hiring meant adding three account executives now thinks first about one killer engineer who can sit with customers, fix onboarding friction, and turn messy use cases into repeatable product patterns. A CRO who once obsessed over top-of-funnel volume now spends more time with product and customer success because expansion depends on usage depth, not just contract size. A VP of Engineering is no longer asked only about roadmap velocity. They are asked whether AI tooling is actually improving output, whether developers are spending less time on repetitive work, and whether the team can support customers who want deeper customization without turning the company into a consulting shop.

Inside many companies, meetings sound different too. The best ones are no longer debating whether AI matters. That argument is over. The real debate is where to place the human effort around it. Which work should be automated? Which work needs expert review? Which customer segments justify high-touch deployment? Which integrations deserve an FDE versus a scaled onboarding path? Which renewals are really product problems wearing sales clothes?

There is also a psychological change. In the old SaaS model, growth and efficiency often felt like enemies living in the same apartment. One wanted to throw a party. The other wanted to read the electric bill. In 2025, the strongest operators are figuring out how to make them roommates. AI helps reduce repetitive work. Smaller teams can move faster. Better technical onboarding improves retention. Better retention improves expansion. Better expansion makes funding easier. The system starts to compound.

But the experience is not all sunshine and celebratory dashboards. There is real pressure underneath this transition. Customers are less patient. Investors are more selective. Token costs, infrastructure costs, and implementation complexity can still bite hard. Companies can overhire expensive technical talent just as easily as they once overhired SDRs. And not every AI feature deserves a parade. Some are just regular software wearing a futuristic hat.

That is why the most useful lesson from ICONIQ’s report is not “AI wins.” It is “the winners are reorganizing around customer outcomes faster than everyone else.” They are building tighter teams, shorter feedback loops, more technical GTM motions, and more disciplined capital plans. They are not assuming the old SaaS machine will magically work with a chatbot bolted onto the side. They are rebuilding the machine.

That is the lived experience of software in 2025: fewer ornamental roles, more outcome-focused roles; less faith in handoffs, more demand for cross-functional ownership; less appetite for empty pipeline theater, more pressure for real deployment and real usage. It is messier, more technical, more demanding, and, for the companies that adapt, much more powerful.

Conclusion

ICONIQ’s 2025 software outlook points to a market that is no longer waiting around for permission to change. Forward-deployed engineers are booming because the hardest part of modern software is not pitching intelligence. It is operationalizing it. GTM is shifting toward post-sales because adoption, retention, and expansion are where AI products prove their worth. Funding is stronger because investors can see real traction, even if the dollars are concentrated around the strongest narratives and operators.

The big lesson is simple: the next generation of software leaders will not be defined by who says “AI” the most times on earnings calls or landing pages. They will be defined by who redesigns product, GTM, and company structure around faster customer outcomes. That is the new playbook. And unlike a lot of old software advice, this one may actually age well.