Software’s Re-Founding Moment: What Separates Transformation from Theatre

Every major software company in the world is asking the same question right now: do we bolt AI onto what we have, or do we rebuild from scratch?

The term industry has settled on for this choice is “re-founding.” Airtable, Handshake, Opendoor, Block, MoneyGram, and Atlassian have all used it publicly in the last twelve months. The word carries weight because it borrows from the gravity of creation. It says: this is not a feature release or an incremental upgrade. This is existential for the company.

The context for this vocabulary is notable. Over 90,000 tech workers have been laid off in the first 97 days of 2026 alone, according to TrueUp, at a pace of 933 people per day. Oracle executed the largest single-day layoff in tech history in March 2026, cutting 30,000 employees to fund a $2.1 Bn AI infrastructure restructuring. Block reduced its workforce by 40%. 

Atlassian cut 1,600 employees while its cloud revenue grew 26% year-on-year.

Meta is planning up to 15,000 additional cuts while committing $115-$135 Bn in AI capital expenditure for 2026. All this after massive cuts over 2025, with over 124,000 employees laid off across 271+ tech companies

These are not small adjustments. But do they represent the largest reallocation of corporate resources from human capital to AI infrastructure in history? 

The question worth asking is whether this reallocation will produce genuine transformation or whether the language of re-founding is simply providing a more palatable frame for what would have been called layoffs in any other era.

The honest answer is that it depends entirely on who is doing it and how.

Re-Founding Software Companies Is Not Easy

The challenge for incumbents is that building an AI-native company is fundamentally different across nearly every organisational vector. A framework from Redpoint Ventures lays this out with uncomfortable clarity.

In traditional software, you hire executives who have “been there, done that.” In AI-native companies, you need first-principles thinkers, because genuinely nothing has been done before from this layer at this scale. Product development is inverted: traditional SaaS is customer-led, built to spec. 

AI-native teams must deeply understand what the models can do and reason forward from capability, not just backwards from requirements. The entire orientation shifts from demand-led to possibility-led.

The economics change too. Traditional SaaS runs at 75-80% gross margins with near-zero marginal cost per user. AI-native companies operate at 50-60% blended margins, with real inference costs that scale with usage. Engineering moves from deterministic to probabilistic systems, which requires entirely different quality assurance, testing, and reliability mental models. 

The sales motion shifts from packaged products sold as-is to field deployment engineering (FDE), where everything begins as a proof of concept on premise with heavy customer education and context immersion. Pricing may move from seat-based predictable ARR to consumption or outcome-based models that remain genuinely experimental.

This is precisely why bolting an AI feature onto a traditional SaaS product and calling it re-founding may be more of performance theatre rather than permanent transformation. Real re-founding means accepting that your org chart, your margin structure, your sales motion, and your product development process all need to change at the same time.

There is a deeper structural point here, and it runs counter to the prevailing doomer narrative. When AI makes the unit cost of code production fall, companies build more software, not less. This is the Jevons Paradox at work. 

Citadel Securities data, sourced from Indeed, shows software engineer job postings up approximately 11% year-on-year even as overall Indeed job postings are down 5.2%. TrueUp reports that software engineering roles have doubled since their low point in mid-2023, with over 67,000 positions currently open across 9,000 tracked companies. 

Demand for software is increasing. What is changing is the type of software being built, the teams building it, and the business models sustaining it.

The Spectrum Of Execution

The most successful re-founder in modern corporate history remains Satya Nadella. When he took over Microsoft in 2014, the company had missed the mobile wave, was losing cloud market share to AWS, and was culturally fragmented by internal silos and title-ranking. 

Microsoft was replaying a game it had already won while the growth moved elsewhere.

Nadella’s re-founding worked because it operated across every axis simultaneously. He exited unwinnable markets, shutting down the Nokia mobile business rather than continuing to invest in a losing position.

He bet on the next platform before it was consensus, making Azure rather than Windows the company’s centre of gravity. He restructured the culture from “know-it-all” to “learn-it-all.” He also made the multi-billion dollar OpenAI partnership happen before most companies had an AI strategy, later hiring Mustafa Suleyman to run Microsoft AI.

The result: Microsoft’s market capitalisation grew from roughly $300 Bn to over $3 Tn in a decade, with revenue reaching $281.7 Bn in fiscal year 2025.

In January 2026, Nadella launched a personal blog declaring this year “a pivotal year for AI,” arguing the industry must evolve from “models to systems.” Even he acknowledges the next phase will be messy. But the Nadella playbook offers a clear benchmark: preserve what makes the company great (developer ecosystem, enterprise relationships, platform DNA) and discard what holds it back (Windows-centricity, internal competition, hardware distractions). Re-founding is not erasing the past. It is refining it.

Among today’s re-founders, the quality of execution varies enormously.

Handshake is perhaps the most quoted current example. The company validated its new direction before announcing it, growing an AI business from zero to $100 Mn in annualised revenue in eight months (self declared) by pivoting from career platform to an expert network for AI post-training. 

Its AI division expanded from 15 to 150 employees. The five-day return-to-office mandate and cultural reset were part of a coherent package. CEO Garrett Lord was direct: “Winners and losers are being defined right now.” Handshake proved the business case first, then announced the re-founding. That sequencing seems to matter.

Airtable represents a conviction-driven bet. CEO Howie Liu chose “re-founding” deliberately over “relaunch” or “transformation” because, as he put it, “the stakes feel the same.” The company launched Superagent in January 2026, its first standalone product in thirteen years. 

It acquired DeepSky for multi-agent coordination capabilities. It made AI the default across every pricing tier, including the free one. The valuation has declined roughly 66% from its 2021 peak to approximately $4 Bn on secondary markets, but Liu says the company is cash-flow positive with half its raised capital still in the bank.

 The jury is still out, but the depth of product commitment seems real.

Opendoor illustrates the risk of using re-founding language without the operational foundation. New CEO Kaz Nejatian declared that “we are re-founding Opendoor as a software and AI company.” But the quarterly results told a different story: revenue down to $915 Mn from $1.57 Bn in the prior quarter, a wider-than-expected loss, and contribution margins collapsing to 2.2%. The stock fell 23% in after-hours trading. Re-founding requires credibility. Credibility requires results, or at minimum, a visible path to them.

The pattern across these examples is instructive. Handshake validated first, then announced. Airtable committed with genuine product depth. Opendoor reached for the narrative before the numbers supported it. 

The word is the same in each case. The substance varies enormously.

The Talent Paradox

For every company considering a re-founding, the talent implications deserve serious attention. The data from recent academic research is nuanced and important.

A Harvard study by Hosseini and Lichtinger, published in 2025, tracked 62 Mn workers across 285,000 US firms and found that junior employment at companies actively adopting generative AI declined by 7.7-10% within six quarters of adoption. Senior employment at those same firms continued to rise. The researchers call this “seniority-biased technological change.” 

The mechanism is important: companies are not firing junior employees. They are simply not hiring them. The entry-level pipeline is quietly closing.

A Stanford Digital Economy Lab study by Brynjolfsson, Chandar, and Chen, analysing payroll data from ADP, found that employment for software developers aged 22 to 25 dropped nearly 20% from its late-2022 peak through mid-2025.

Developers over 26 remained stable or grew. The two age cohorts tracked in lockstep until ChatGPT launched in November 2022. Then they diverged sharply.

MIT’s Project Iceberg, developed with Oak Ridge National Laboratory, estimates that AI can technically perform tasks equivalent to 11.7% of the U.S. workforce, representing approximately $1.2 Tn in wages.

But current adoption is concentrated in just 2.2% of the workforce. The gap between what AI can technically do and what organisations have actually deployed is where the next five years will play out.

The Vanguard 2026 economic forecast offers a crucial counterpoint. It found that the approximately 100 occupations most exposed to AI automation are actually outperforming the rest of the labour market in both job growth and real wage increases.

 The professionals who master AI tools are becoming more valuable, not more replaceable.

The re-founding implication is clear. Companies that re-found well will create more technical roles at higher skill levels and higher leverage. The Jevons Paradox rewards genuine transformation. But companies that hollow out their junior pipelines in the name of efficiency are eating their own seeds. Every senior engineer was once a junior engineer.

Microsoft’s Azure CTO Mark Russinovich and VP Scott Hanselman recently published a paper in Communications of the ACM warning that agentic AI is creating an economic incentive to stop hiring junior developers, and that organisations acting on this incentive are gutting the talent pipeline that produces senior engineers. They propose a “preceptorship model,” borrowed from medical education, pairing seniors with juniors at structured ratios. 

This is the kind of institutional thinking that separates re-founding from mere cost-cutting.

What Real Re-Founding Requires

Re-founding is the hardest strategic manoeuvre a mature company can attempt. It is also, for the companies that get it right, the most valuable. 

The difficulty is real, but it is worth spelling out what it actually takes, because the companies that approach it with clarity will have an enormous advantage over those that treat it as a communications exercise.

First, it requires principal alignment across the entire organisation. The CEO, the board, the product leadership, and the frontline teams all need to understand that this is not a “use AI more” initiative. 

It is a structural reset of how the company creates and delivers value. When Airtable’s Liu chose the word “founding” over “transformation,” he was making a point about the level of commitment required. A transformation can be managed incrementally. A founding demands total conviction and buy-in.

Second, it forces leaders to make hard choices about what to stop doing. Nadella shut down Nokia mobile. He did not try to run it alongside Azure and hoped for the best. The leaders who try to re-found without making hard choices about what to exit will end up with bloated hybrid organisations that are good at neither the old model nor the new one. Preserving everything is not a strategy. It is avoidance.

Third, the process and people dimensions cannot be ignored. This is where most re-foundings will fall short. Vague “use AI” mandates obscure the real work: retraining teams, redesigning workflows from first principles, and rebuilding product development culture around probabilistic rather than deterministic systems.

AI-native startups are growing two to three times faster than top-quartile traditional SaaS companies, according to ICONIQ Capital’s State of Software 2025 report. Some have reached $30 million in ARR in just twenty months, roughly five times faster than conventional SaaS trajectories. All incumbents cannot match that velocity just by layering AI features onto existing workflows. They can only match it by rebuilding the workflows themselves.

Fourth, and perhaps most importantly, re-founding is a one-time card. As Handshake’s CMO Katherine Kelly put it: “If you’re really going to make people believe it at your company, you have to put a lot of effort into it, because you only get this one chance.” 

Employees, investors, and customers will give a company one window to make this transition. A second attempt will be received as evidence of failure, not ambition. The credibility cost of a failed re-founding is far higher than the credibility cost of never having attempted one.

Stanford HAI’s 2026 prediction captures the moment precisely: “The era of AI evangelism is giving way to an era of AI evaluation.” The question is no longer whether AI can do something, but how well, at what cost, and for whom. Companies that re-found with that evaluative rigour will earn the right to the word. Companies that use it as narrative cover will not.

India’s Re-Founding Advantage

India’s software and SaaS ecosystem faces the re-founding question from a fundamentally different position than Silicon Valley.

Indian SaaS companies have spent the last decade building for global markets with structurally leaner cost bases. India’s 1,800-plus Global Capability Centres and its share of the global STEM workforce mean the talent infrastructure for AI-native product development already exists at scale.

Indian SaaS companies that grew up building for global enterprise customers already understand the field deployment model that AI-native sales requires, and this has been a core strength of the Indian IT services enterprises from inception.

The risk is complacency. Indian SaaS companies that assume their cost advantage alone will protect them are misjudging the nature of the shift. AI does not just compress labour costs. It changes the nature of the product itself, the sales motion, the margin structure, and the customer relationship.

An Indian SaaS company with excellent engineering talent but a traditional product development process will be outrun by an AI-native competitor with one-tenth the team, wherever that competitor happens to sit.

The opportunity, however, is significant. For Indian founders and technology leaders, this is a re-founding moment for the ecosystem itself. The companies that embrace AI-native product development, pricing experimentation, and workflow redesign will define the next generation of Indian technology companies building for the world. 

India’s structural advantages in talent, cost, and digital infrastructure make it one of the best-positioned ecosystems globally for this shift. But the advantage must be actively exercised. It will not hold on its own.

The Only Test That Matters

In the end, the word “re-founding” is only as meaningful as its outcome.

Does the company come out the other side with more relevant products, better unit economics, a clearer growth trajectory, and a reason to be used repeatedly by its customers? That is the only reason to endure the pain of a re-founding.

Not narrative positioning. Not investor signalling. Not the runway extension that comes from cutting headcount.

The founders who understand this will build the next generation of category-defining companies. Some will be first-time founders building from zero. Some will be the rare re-founders who earn the title by rebuilding from within. 

The difference will not be in the language they use. It will be in the products they ship, the teams they build, and the customers who stay.

The post Software’s Re-Founding Moment: What Separates Transformation from Theatre appeared first on Inc42 Media.