You are 55, sitting on a portfolio of maybe fifty angel investments spread across SaaS, marketplaces, fintech, consumer AI, and things that looked promising in 2020. You have built a good life on the back of the IT career you made through the dot-com era: the IPO in 2000, a couple of trade sales, some failures that taught you what not to do again, and four years of watching founders and their teams build companies while you learned what made some of them work. You know the playbook. You are good at spotting founders with the right instincts. You have built a network in Stockholm's startup scene that delivers deal flow without you pushing for it. By 2026, it is reasonable to assume you have another fifteen, maybe twenty active years ahead of you, and that angel investing would gradually wind down toward something more like mentorship, golf, and whatever comes after ambition. I should have told you this would not happen. Not because you got any of those assumptions wrong, but because the playbook was about to be rewritten by people who had not read the 2010s version.
1. What I know now that I did not know then
The First Autonomous Fund crossed ten billion euros in assets under management in Q4 2026, and it was not a curiosity. It was the opening shot. By 2028, three of the top ten performing venture funds in continental Europe operated with AI-native infrastructure and fewer than twenty human employees. The business of picking winners had begun to change. Not because human judgment had become worthless - it had not - but because the decision-making infrastructure, the data synthesis, the pattern recognition, the speed of iteration: these were now being done by teams where AI was not a tool but a core actor. The cognitive work that had differentiated experienced investors from naive ones was becoming something that machines could do faster, with less emotional attachment, and with the ability to process more data than any human investor could synthesize. Your edge - the pattern recognition, the founder intuition, the systems thinking - was still valuable. But the context in which it was valuable had shifted. You would need to develop a new context to make the old edges matter.
The portfolio you held in January 2026 was skewed toward human-labour-leveraged SaaS companies, and those companies got repriced sharply downward between Q3 2027 and 2029. Your core holdings were businesses built on the assumption that software wrapped around humans with sales teams, customer success teams, professional-services revenue - those were the defensible, high-margin models of the 2010s. A company with twenty people could serve five hundred customers if each customer had a dedicated success manager. A consulting firm could charge premium rates because the knowledge workers were expensive and scarce. A SaaS platform with a sales team could go upmarket and extract higher contract values. These were solid assumptions for an economy where good people were expensive and automation was narrower and slower. By 2027, those assumptions were upside down. The companies in your portfolio that depended most heavily on human labour - the ones you had picked precisely because the playbook said that was where margins lived - became the ones most exposed to disruption. The repricing was harsh, but not because the founders were bad operators. They were being penalised for building companies that fit the 2010s economy, and that economy was stopping.
The First Fully Autonomous Firm reached five hundred million euros in revenue in Q4 2027, and it reset the benchmark for what a company needed in terms of human headcount. A company with one billion euros in revenue in 2020 had, typically, three hundred to five hundred employees. A company with five hundred million euros in revenue in 2027 might have five. Not five hundred. Five. The unit economics had inverted. The cost structure had flipped. Your portfolio companies with fifty to one hundred and fifty people, burning cash through market development and customer success, looked suddenly embarrassing next to competitors with five people, better margins, and faster iteration cycles. You saw this happen. I saw you see it happen. What I should have told you was that this was not a sign that you had picked badly. It was a sign that the benchmark had moved, and that moving the benchmark was the story of the next five years.
The Swedish Compact in Q3 2026 created public compute infrastructure and AI-augmented retraining pathways, and it redirected smart capital toward the edges where policy met venture. Sweden's decision to build this infrastructure as a public utility, and to create the governance frameworks and retraining pathways to go with it, was the largest non-military public investment in a generation. And it created, around itself, a set of opportunities that were nowhere on your existing map. Workforce transition platforms, AI governance tooling, compliance infrastructure for AI-augmented systems, regulatory technology, edge applications where public compute met private needs. This is where the second wave of real returns happened in 2028-2029. You were not there. Not because you were incompetent, but because the playbook you had learned did not teach you to look for returns at the policy-capital boundary. SaaS returns came from product-market fit and operator discipline. The Swedish Compact returns came from understanding where the public sector was genuinely going to spend money, what it genuinely needed to buy, and being positioned between what the government had built and what the market needed to do with it.
The Longevity Threshold hit you at fifty-eight, in Q3 2029, and it changed your time horizon and your risk tolerance in a way that made the advice I would have given you in 2026 premature by decades. The proof-of-concept, finally credible, that ageing was modifiable. Not a marginal improvement. Not a five or ten percent extension. A reversal of decline in cognitive and physical function, and the data suggesting that another forty or even sixty years of active life was something you would need to plan for rather than something that would happen to you passively. At fifty-five, your mental model was: I have twenty years, maybe twenty-five, before I wind down. At fifty-eight, with the longevity data in front of you, the mental model became: I have forty, possibly sixty years. That is not a tweak to the same strategy. That is a complete reformation of what strategy makes sense. The angel investments you had made assuming you would exit most of them within a decade, or at least be less active in ten years, suddenly looked short-duration when your own duration had just doubled.
Nordic Tech Club, where you spent many evenings and where much of your network was built, went through its own transition between 2026 and 2028. The founders who showed up in 2027 and 2028 were different in kind: younger, AI-native, entirely uninterested in referencing the 2010s playbooks that the senior angels in the room kept bringing up. They had not come through the SaaS scaling phase. They were building from first principles in an economy with agents, not just software. They had different questions, wanted different kinds of advice, and - in many cases - preferred investors who had not built their fortunes on the previous paradigm. You remained the room's senior voice, its resident IPO-era credibility, its connection to the older generation of Swedish tech. But the room itself was becoming somewhere else. The network that had delivered deal flow without effort in 2024-2025 was delivering deal flow to you that your instincts were still learning how to evaluate.
2. What I got wrong about your situation
I assumed your network alone would keep you in deal flow and that network would keep you relevant. It did both of those things, for a while. But the most valuable deals of 2028-2030 were not happening through Nordic Tech Club discussions and the founder networks you had built over decades. They were happening in AI-native founder circles, policy-capital forums, and emerging-technology communities that you were not really in. You had institutional credibility. You did not have immediate access to the people who were building the next layer of infrastructure. The friction of getting up to speed on agent architecture, on public compute infrastructure, on AI governance - this mattered more than you would have guessed.
I assumed that your IPO credit from 2000 and your thirty years of startup history gave you signalling weight with new founders. It did, and it still does. But by 2028, many of the best founders wanted investors who had built or invested in companies that understood agent infrastructure, not investors whose most credible track record was building SaaS companies in a labour-arbitrage era. You were not less credible. You were credible for a different world. The transition period - the five years where the old world and the new world overlap - is a painful place for people whose credibility is tied to the old one.
I underestimated the personal-identity reshaping that came with the Longevity Threshold. At fifty-five, your plan made coherent sense: you would gradually move out of active angel investing in your mid-sixties, take on more mentorship, slow down. At fifty-eight, after the longevity data became credible, that plan looked premature by decades. But the shift was not just logistical. It was emotional and identity-based. You had begun, at fifty-five, to think of yourself as entering a wind-down phase. The longevity data forced you to think of yourself as having another major act to play. That is a larger identity shift than I accounted for, and I should have.
3. What I should have told you to do, in order
In 2026, starting from Q1
Audit your portfolio ruthlessly. For each company, sit down and answer one question: does this business still work if AI agents do sixty percent of the tasks that humans currently do here? If the answer was no, or even "I am not sure," cut exposure. Do not wait for the repricing to confirm what you already suspected. The repricing comes too late. You would have fewer total companies in your portfolio, but the ones that remained would have been companies designed to work in a world where labour was less important.
Rotate at least thirty percent of your remaining capital into AI-native plays. Not the obvious consumer AI bets that everyone else was making. Agent infrastructure, AI governance tooling, AI safety, data and infrastructure for agents, compute provisioning. The best returns of the 2028-2030 period came from companies that were building the tooling and infrastructure that autonomous systems needed, not from companies that were trying to sell autonomous systems directly to consumers or to large enterprises buying them as black boxes.
Get close to the Swedish Compact work immediately. Do not do this for a payday. Do this as a way to stay close to where the policy-capital edges were forming. Offer to advise one of the Compact pilot programmes for one day a week. For free, if necessary. You needed to understand the government's genuine priorities, the infrastructure being built, and the private-sector businesses that would thrive in this new public-private landscape. That knowledge was worth more than the money you would have made from an additional series B investment in a company you half-understood.
Start building relationships in AI governance and safety circles. You had a strong network in the startup founder community. That was valuable, but it was a relatively narrow community. The people thinking hard about how autonomous systems should be governed, how to ensure they did not create uncontrollable risks, how policy should interact with technology - these were different people. Some of them were academics. Some were in the public sector. Some were founders too, but they were focused on policy, not product. Getting to know that community would have been harder than deploying capital with founders you already knew. It would also have been more valuable.
In 2027-2028, during the deep transition
Let your SaaS portfolio run off or exit at reduced valuations. This is the hard part. You are not going to get fair value. You are going to watch companies you believed in get repriced downward. It will feel like failure. It is not failure. It is recognition. The capital you would have tied up in a three to five-year wind-down of a labour-intensive business could have been deployed elsewhere. The optionality mattered more than the vanity of not taking a loss.
Make five to ten small cheques into early AI-agent companies. Fifty to two hundred thousand euros each. Most of them would fail. One or two might not. Those one or two would be the best returns of your entire portfolio. The reason to do this was not because you had a special ability to pick AI-agent winners - you did not. It was because you needed to be in the game, to develop pattern recognition, to build relationships with the founders who were building first-layer infrastructure. You needed to lose money on a few bets to develop the instincts to make good bets on the others.
Partner with AI-native founders, not just fund them. Your thirty years of systems thinking, your ability to ask hard questions, your experience surviving multiple transitions in technology - these were genuinely useful skills. But they needed to be translated. A piece of advice that made sense for a SaaS operator in 2015 might not make sense for an agent-infrastructure company in 2027. The work was to do the translation. The founders who found this translation valuable were the ones who would remember you, who would come back to you for future rounds, who would refer other founders to you. The capital deployed was less important than the relationship built.
In 2029-2030, after the Longevity Threshold
Take the longevity interventions seriously. Not as a vanity play, not as life extension for its own sake. Take them as a career recalibration. If you were going to be active and cognitively sharp until eighty-five or ninety, your investment horizon had just doubled. The companies you invested in now should be companies that would still be relevant in 2040 or 2045. That changed what you should invest in. Short-cycle exits became less important. Building real long-duration assets became more important.
Move up the cap table. Lead small seed rounds rather than taking half a percent in someone else's Series B. You would make smaller total returns this way on a capital deployed basis, but you would have more control, more information, and more equity concentration. The world of 2030-2040 was not going to deliver you fragmented portfolio returns. It was going to deliver concentrated returns to people who had put real capital and real time into the companies that mattered.
Start redirecting some capital toward long-duration plays that would compound over your extended horizon. Biotech, energy transition, AI safety, things that would take fifteen or twenty years to resolve but would reshape the world when they did. At fifty-five, those time horizons are too long. At sixty-five, with forty more years of life ahead, they become reasonable. You would need capital that was willing to sit for that long, and patience that matched that timeline.
4. What I should have told you to stop believing
That the 2010s SaaS playbook transferred. It did not. The playbook said: hire great people, build a good product, create a sales organisation, scale customer success, build switching costs and network effects, exit at a premium multiple. That playbook worked because the scarce resource was human talent and human effort. A business that could leverage human talent more efficiently than the competition had an edge. The playbook does not work the same way in a world where the scarce resource is compute, capital, and the people who understand how to build with agents. The economics are different. The value creation is different. A piece of advice that served you well in 2005 might not serve you well in 2027.
That your network is your edge. Your network is valuable. It was your network. It was not necessarily the right network any more. The most important deals of 2028-2030 were not happening in the rooms where you spent your time. They were happening in other rooms, with other people, discussing other priorities. Being well-connected in a shrinking community is less valuable than being connected to an expanding one, even if that connection is thinner and less comfortable.
That you had fifteen to twenty active years left. The Longevity Threshold gave you thirty to forty, maybe more. That is not a marginal change to the same strategy. That is a complete recalibration of what makes sense. The companies you were comfortable with at fifty-five, on a fifteen-year timeline, were not the companies you should have been comfortable with at sixty, on a forty-year timeline. Your time horizon, your risk tolerance, your exit expectations - all of these needed to change.
That angel investing was about picking winners. In the Inhuman Economy that was forming, it was about being in the rooms where the new rules were being written. It was about having the relationships and the understanding to see where capital was going to be needed, where value was going to accumulate, and where the old playbooks did not apply. The best returns came not to the people who best predicted the future, but to the people who helped shape it.
5. What I am telling you now, looking forward from 2031
Stockholm is a better place to be investing from in 2031 than almost anywhere in the West. The Swedish Compact created a policy-capital-talent triangle that most countries are trying to replicate and not succeeding. The government invested in real infrastructure. Founders and investors followed. The ecosystem is younger, more AI-native, more policy-literate than most of its peers. You are positioned there. That is an advantage.
Your best asset turned out not to be the IPO credit from 2000 or the track record of picks in the 2010s. It was something simpler and more durable: the willingness to sit with uncomfortable founders, to ask hard questions, to tell them something they needed to hear instead of something they wanted to hear. AI-native founders, as it turned out, were short of that. They had access to capital. They had access to technical advisors. They were often short of people who would slow them down, who would push on their assumptions, who would point out the risks they were moving too fast to see. Mentorship at capital scaled. The best founders came back, not because they had made money yet, but because the relationship was valuable.
The Longevity Threshold has given you a genuine second act. Not a wind-down. Not a graceful exit from the game. A second build. The capital you have, the relationships you have, the clarity you have developed about what works and what does not - these are not the diminishing assets of someone approaching retirement. They are the accumulated resources of someone with twenty years of active work ahead. The people who understood this early had a kind of freedom that those who stayed locked in the old time horizons did not have. You have that freedom now.
Siri Southwind
Written 31 December 2030