AI Startup Soft Landing: M&A vs Shutting Down

If your AI startup is not working the way you hoped — the fundraise is not coming together, growth has stalled, runway is shrinking — shutting down is not the only option. A soft landing acquisition can be a significantly better outcome for founders, employees, and investors alike. This article explains what a soft landing is, when it makes sense, and how to approach one if you are in that situation.

There is no judgment here. Most AI startups do not make it. Getting to a soft landing instead of a disorderly wind-down is a meaningful achievement.

What a Soft Landing Is

A soft landing is a structured sale of a company that is distressed, stalling, or running out of runway — typically before things get to the point of default or forced liquidation. The term covers a range of structures, but the two most common are:

Acqui-hire: The acquirer’s primary interest is the team. The company’s technology, IP, and assets may or may not transfer in full, but the core value being acquired is the people — particularly engineers, researchers, and technical leads. Employment packages for key team members are the main consideration, with a smaller component going to satisfy investor preferences or cover remaining liabilities.

Asset sale: The acquirer buys specific assets — the model, the codebase, the training data, the customer contracts, or some combination — while the legal entity winds down separately. This structure is common when the acquirer wants the IP but not the team, or when the team does not want to join the acquirer.

Both structures are “soft” relative to the alternative: a complete shutdown with zero recovery, employee layoffs, and a scramble to return (or write off) investor capital.

The Spectrum: From Tough But Viable to Three Weeks Left

Not all soft landings start from the same place. The circumstances shape the options available to you.

12+ months of runway, growth stalled. You still have time and leverage. A soft landing at this stage is proactive rather than reactive — you are choosing to find a good home for the company rather than continuing to burn capital on a trajectory that is not working. At this point, you can run a proper, competitive process: prepare materials, build a target list, create tension between multiple potential acquirers. You are more likely to get meaningful consideration for IP, a better team package, and some return for investors.

6–12 months of runway, no clear path forward. You need to move now. This is the most common starting point for AI startup soft landings. You have enough time to run a process if you start immediately, but not enough time to be slow or unfocused. Outreach needs to be targeted — not a mass blast, but a prioritised list of the five to ten acquirers most likely to value your team and technology. This is the stage where an advisor pays for itself.

Less than 6 months of runway. Your options narrow significantly. Potential acquirers can see the clock and know they have leverage. That said, many soft landings do close quickly — particularly acqui-hires, where the acquirer’s main risk is that key engineers leave before signing. If you can credibly hold the team together through a fast close, you can still get a reasonable outcome. Be direct with potential acquirers: the honest framing usually works better than trying to manufacture urgency that sophisticated buyers will see through anyway.

Three weeks of runway. This is crisis mode, not soft landing mode. You are now looking at emergency conversations with the most likely single acquirer, not a process. The focus shifts to protecting the team — making sure people get paid and land somewhere soft — rather than maximising consideration.

Why AI Is Different

AI startup soft landings are more common than in other sectors because the asset profile is different. In most tech startups, if the product does not achieve product-market fit, there is limited residual value. The technology may be generic, the codebase mediocre, the customer relationships thin.

AI startups are different on two dimensions:

Team value is high even when product fails. A small team of ML engineers, researchers, or applied AI specialists who have been working together for two years is genuinely scarce and genuinely valuable. The talent market for experienced AI engineers remains competitive, and an acquirer can often justify paying $1–2 million per head simply to avoid the cost and time of hiring and ramping equivalent talent from scratch. This dynamic drives acqui-hire activity across the AI sector regardless of whether the product succeeded.

Acquirers prefer live companies over corpses. Once you shut down — once the employees have dispersed, the engineers have taken jobs elsewhere, and the codebase has gone stale — the residual value of the IP drops sharply. Acquirers who might have been willing to pay for the team in an acqui-hire have no interest in picking through the assets of a dissolved entity. The soft landing option disappears when the company formally winds down.

This is why timing matters more in AI than in other sectors. The window between “things are bad” and “the team has scattered” can be measured in months.

Real Examples from the AI Sector

The AI acqui-hire market has been active across the value spectrum in 2024 and 2025. At the high end: Microsoft’s $650 million arrangement with Inflection AI structured primarily around hiring co-founder Mustafa Suleiman and the technical team; Google’s approximately $2.7 billion transaction for Character.AI structured around acquiring founders Noam Shazeer and Daniel De Freitas. These are extraordinary cases involving world-class research talent.

More instructive for most founders are the smaller transactions: OpenAI’s acquisition of personal finance AI startup Hiro Finance, a small team acqui-hire where the company shut down operations and the team joined OpenAI. Meta’s December 2025 acquisition of Dreamer, an agentic AI startup, followed a similar pattern — team and technology absorbed, company wound down.

These smaller soft landings rarely receive press coverage. That does not mean they are not happening. Per-CB Insights data, acqui-hire activity hit multi-year highs in Q1 and Q2 of 2025, driven specifically by big tech and well-capitalised AI platform companies absorbing teams from startups that did not raise their next round.

Proactive vs Reactive: Why Earlier Is Better

The distinction between a proactive soft landing and a reactive one is the single most important concept in this article.

A proactive soft landing is initiated by founders while the company still has meaningful runway, the team is intact, and there is no public perception of distress. The founder controls the narrative: “We have built something valuable, we have evaluated our path to independent scale, and we believe the mission is better served by joining a larger platform.” Potential acquirers see a company with options, not a company in trouble. Process discipline is maintained. Founders have time to evaluate multiple conversations and choose the best fit.

A reactive soft landing happens when the situation forces the founder’s hand. Runway is critical, morale is low, and the conversation with potential acquirers starts from a position of visible distress. Sophisticated acquirers — and all the good ones are sophisticated — adjust their offer accordingly. The team packages shrink, the IP consideration disappears, and the timeline compresses in ways that favour the buyer.

The practical implication: start earlier than feels necessary. If you are reading this article because things are getting hard, that is the right time to start — not when you have two months left.

What You Can Still Get in a Soft Landing

Going into a soft landing with clear expectations protects everyone involved. Here is what is typically available:

Team employment packages. In an acqui-hire, the primary consideration flows to employees as employment offers, signing bonuses, and retention packages. In competitive AI talent markets, a team of five strong ML engineers might generate $5–10 million in total team-side consideration through this mechanism.

IP sale price. If the acquirer wants the codebase, training data, or model weights, there may be a meaningful asset sale price on top of the team packages. This consideration typically goes through the company and is distributed according to the liquidation waterfall — meaning preferred shareholders are paid first.

Investor recovery. Depending on the consideration and the investor structure, preferred shareholders may recover a portion of invested capital. In a 1x non-participating liquidation preference structure, investors recover their investment before founders see anything — but this is better than zero, which is the alternative in a shutdown.

Founder equity. If there is consideration left after satisfying the liquidation preference, founders participate according to their equity ownership. In many soft landings, particularly those with heavy investor liquidation preferences, founder equity recovery is modest. Manage expectations accordingly.

What You Cannot Expect

Full valuation. You are not going to achieve the valuation from your last round, and definitely not the valuation you projected at that round. Acquirers know the circumstances. The discount is the price of the option.

A competitive process. Unless you are genuinely proactive and early, you are unlikely to have multiple serious acquirers bidding simultaneously. Soft landings usually come down to one or two conversations run quickly.

Time. Every soft landing process has an implicit deadline set by your runway. This creates pressure that erodes negotiating leverage. Managing this pressure — including being honest with acquirers about your timeline — is part of the process.

How to Approach Potential Acquirers

The framing matters. If you approach a potential acquirer with “we are running out of money and need help,” you will get acquirer pricing. If you approach with “we have built a valuable team and technology, we have evaluated our path forward, and we believe the right next chapter involves partnering with a company that has the distribution and resources to take this further,” you will get a materially better conversation.

Both statements can be true simultaneously. Lead with the asset, not the distress.

Who to call first: Start with acquirers who already know you. Existing investors who have portfolio companies that might benefit. Partners you have built integrations with. Companies in adjacent categories where your capability fills a gap. Warm introductions produce better outcomes than cold outreach because they skip the trust-building phase that eats up time you do not have.

What to lead with: The team profile and what the team can do. The technology and why it is non-trivial to replicate. The specific capability or data asset that the acquirer cannot easily build themselves. Not the growth metrics you wish you had.

Honest framing works. Acquirers doing soft landing acquisitions know what they are. They have seen the pattern. Pretending you are not in a soft landing situation rarely fools anyone and wastes the founder’s limited time on positioning instead of negotiation.

Your Obligations to Investors and Employees

During a soft landing process, you have active obligations to both groups.

Investors: Your board members and lead investors should be informed early — not necessarily on day one of exploration, but certainly before you get to the point of a term sheet. Most institutional investors have fiduciary obligations of their own and need time to understand the situation and respond appropriately. Surprises at the term sheet stage create legal risk and delay transactions. Preferred shareholders will need to approve any transaction that does not fully satisfy their liquidation preferences.

Employees: You are not obligated to disclose a potential transaction until you have a signed term sheet, and generally should not. But you do have legal obligations around notice periods, final pay, and — depending on jurisdiction — redundancy payments. In an acqui-hire, the acquirer typically handles employment directly, but the timing and communication need to be coordinated carefully to prevent key people from leaving before the deal closes.

The rule of thumb: treat people with more dignity than the legal minimum requires. The AI community is small. How you handle a soft landing — how you treat the team, how you communicate with investors — will follow you.

When to Engage an Advisor

Earlier than you think.

If you are 12 months from a potential wind-down scenario, that is the right time to have a conversation with an M&A advisor who has experience with distressed or sub-scale transactions. Not because you have committed to a sale, but because understanding your options and having the groundwork laid — prepared materials, a target list, warm relationships with likely acquirers — means you can move quickly when the decision is made.

Advisors working on soft landings typically operate on a reduced retainer with a success fee structure that aligns their interest with getting the transaction closed. Even a modest soft landing — $2–3 million in total consideration — justifies advisor engagement when the alternative is zero.

Amafi Advisory works with AI startup founders across Asia Pacific on sell-side transactions at all stages, including soft landings and distressed situations. If you are evaluating your options, the initial conversation is free. See also: how to prepare your AI company for acquisition and raise or sell — a founder’s framework.