Civitas Snippets - May 17, 2026
Issue 55
Dear friends of Civitas Growth Partners,
This issue arrives with the bond market repricing in real time: the US sold 30-year Treasuries at 5% for the first time since 2007, CPI ran 3.8% on Iran-war fuel, and Moody's downgrade language is no longer hypothetical. The AI capital cycle, meanwhile, has decoupled from anything resembling textbook capital allocation - Anthropic's secondary-market valuation flirted with $1.4 trillion and the company agreed to buy 100% of xAI's Colossus compute, all while OpenAI launched a $10 billion private-equity services arm. AI "bubble" watch of the week - Martha Stewart is running an AI home-repair startup… when one thinks about technology innovation, who does not think of Martha Stewart?
As Henry Adams once observed:
"Chaos was the law of nature; Order was the dream of man."
With that, let's get to it…
Politics and Current Events
Space.com reports NASA chief Jared Isaacman is publicly lobbying to reverse the 2006 IAU vote that demoted Pluto to a dwarf planet. The underlying science has not moved. Isaacman has correctly identified that you do not solve an astronomy problem with more taxonomy - you solve it with a popular mandate. The first SaaS-trained NASA administrator is treating classification as a marketing channel, and the dwarf planet now has a constituency.
Axios charts a generational and structural shift among Black voters moving right on cultural questions while remaining skeptical of Republicans on the economy. The cultural realignment was already visible by 2024; what is new is that economic policy has not closed the gap. Both parties are guessing at a coalition that no longer maps to its 1990s shape, and the first party to read it clearly inherits a structural advantage.
Doomberg's Information Grade essay opens with: "the goal of modern propaganda is no longer to transform opinion but to arouse an active and mythical participation in the adherents." His thesis is that the media stack has degraded faster than any institution it covers, and that a serious investor now spends more time triangulating sources than reading them. Read fewer outlets, read them more carefully (including this one).
Elon Musk posted "Bitches / Money / No Taxes / Party" on May 10, characteristically engineered for a reaction. The richest man on earth deploys his X account the way most people use a group text: low-information, high-signal, plausible deniability included. The market reaction is now nominal because the market stopped being surprised some time ago.
Axios writes that America's sin is now scaling: online gambling, legal weed, and deepfake content are all forming categories that Las Vegas spent decades building physically. Moral guardrails have been replaced with click-throughs, and the dollar volumes are real. Any product previously made expensive by social friction is now investable, and the distribution layer - apps, dispensaries, ad networks - is the platform.
The FT reports Americans are deeply anxious about AI's trajectory while Silicon Valley and the federal government continue to insist the technology is unstoppable. Polling shows majority concern about job losses, surveillance, and lack of oversight. The gap between the elite consensus and the public mood is now wide enough to be a political variable. Whoever campaigns on credible AI accountability first has a primary lane to themselves.
Axios AM has the economic alarm bells ringing: CPI at 3.8%, 30-year Treasuries at 5%, and a Moody's downgrade pending while President Trump flies to Saudi Arabia announcing trade deals. The Saudi headline is the kind of win the administration can pin to a press release; the bond market reads the actual numbers and prices accordingly. Two narratives, one economy - bond traders historically win these arguments.
The big news for our healthcare friends, CMS has frozen new Medicare enrollment for hospice and home-health agencies nationwide - a sweeping regulatory action with major implications for healthcare-services PE. Existing operators just got a structural moat; aspiring entrants got a much longer runway to break even. Roll up the licensed seats while the door is closed, then ride the demographic wave the agency just narrowed into. Regulatory capture in real time.
Forbes reports a new survey finds Americans prefer a nuclear plant in their neighborhood to an AI data center. That sentence would have been unpublishable five years ago. The objection set is entirely mundane - noise, water draw, power demand - and it still beat Three Mile Island in a neighborhood preference poll. Zoning boards are not impressed by compute.
Economy, Markets, and Business
Bloomberg has Fitch warning the US debt burden is now "far above" other AA-rated nations and the deficit trajectory is no longer comparable to peers. Moody's already moved earlier this year; Fitch's would be the second shoe to drop. Investors keep buying the dip on faith; the bond market is closer to fading that faith than the equity market is.
The NYT essay on the Age of Electricity is the most underrated piece of macro this issue: the US grid is aging, demand is surging from AI and EVs, and infrastructure investment has lagged for decades. Grid-adjacent industrials - turbines, transformers, substation equipment, copper - is one of the few unmet macro themes left where the bottleneck is real and the supply curve is inelastic.
James Lavish's latest walks through how a sovereign-debt crisis unfolds: the US already owes more than every American has ever earned, and the playbook - fiscal dominance, reserve status loss, capital controls - has run before in places we do not consider analogues. Lavish has been early on this trade since 2022; the bond market is now catching up. Worth reading even if you think he is wrong.
The WSJ found 67% of prediction-market participants lose money while a tiny cohort of "sharks" extracts the value. Polymarket and Kalshi are the cleanest live demonstration of why pari-mutuel markets are not zero-sum: the house and the sharks take their cut, and the recreational money supplies the liquidity. The structure is identical to retail options trading and slightly worse than a casino - at least a casino tells you the edge upfront.
Science published a study finding LLMs now outperform physicians on standardized clinical-reasoning tasks, not pattern matching but actual diagnostic logic. The AMA has been preparing testimony against this inflection for years. The residency pipeline now competes with a $20-per-month subscription, and the economic implications for outpatient medicine, second-opinion consults, and the payor-side utilization-management stack are larger than any single CMS rule. The training pyramid breaks first.
Axios Macro's Neil Irwin walks through what the 100% debt-to-GDP milestone actually means: the number is overstated, but the trajectory and interest-cost burden are alarming. Japan ran above 200% for two decades without a crisis - Japan has a closed savings pool and a captive central bank; the US has neither. Watch annual interest expense, which has now crossed defense spending, and the spread between marginal new debt and marginal GDP growth, which is negative.
Bloomberg notes Berkshire's cash pile hit a record $397.4 billion in Greg Abel's first quarter as CEO. The Berkshire cash pile has been a reliable late-cycle valuation gauge for decades - when Buffett would not buy, things eventually got cheaper. The market has yet to deliver Abel the same opportunity, and the longer the cash sits, the louder the implied prediction becomes.
The Kobeissi Letter highlights offshore USD deposits at a record $14.5 trillion as dollar dominance is supposedly under threat. BRICS alternatives keep getting announced; offshore dollar balances keep growing. The death of the dollar has been the most consistently predicted and persistently deferred event of the post-Bretton Woods era. Believe the de-dollarization trade when it shows up in the deposit data, not the press releases.
Forbes calculates SBF's early Anthropic and other venture positions would have been worth $100 billion had he stayed out of prison. The AI boom would have made him one of the richest people alive instead of one of the most famous prisoners. Being early on the right venture rounds is so lucrative that even competent fraud cannot match the opportunity cost. FTX is now a 12-figure mistake.
Axios Pro Rata reports major PE firms are forming JV structures with OpenAI and Anthropic, using PE capital to fund the compute buildout in exchange for preferred model and data access. The lines between PE and tech are blurring at the structure level, not just the rhetoric. The downside risk is that the JV becomes the new continuation vehicle: a place where capital goes to wait.
a16z argues the AI job apocalypse is a complete fantasy on the grounds that historical automation always created more jobs than it destroyed. The piece is reassuring, but it conflates aggregate outcomes with distributional ones. The Luddites lost the macro argument and won the local one; the displaced weavers were not retrained as software engineers. Both the historical record and the transition cost are real.
The Information reports Anthropic is buying 100% of compute from xAI's Colossus data center, an unexpected alliance between fierce AI competitors. Musk gets the revenue; Anthropic gets the capacity; both get to demonstrate that the AI cold war stops at the substation. The structure is a wholesale-toll-road deal - SpaceX builds, Anthropic rents, the model layer pays the bill. Coopetition is the dominant strategy when compute is the binding constraint.
The Kobeissi Letter reports President Trump's 10% Intel stake is now worth roughly $56 billion after the Apple chip-supply deal - gaining $47.6 billion since the purchase was announced during tariff negotiations. Industrial-policy capital plus a tariff overhang plus a single anchor customer equals a meme-stock-grade federal return - but only if you also run the relevant trade-policy apparatus. To our LP readers out there, feels like a Trump / SBF fund would be a good place to invest.
The WSJ reports a Greg Brockman diary from OpenAI's founding years has become central evidence in the Musk-vs-OpenAI trial. The diary details internal discussions about commercialization, safety, and control - the holy trinity of post-hoc AI litigation. Every internal message at a frontier lab is now a potential deposition exhibit. Also, how many grown men keep diaries? That feels like the strangest part, along with how did Elon's team know he was keeping one?
Cremieux Recueil posted a chart showing one Chinese shipbuilder produced more ships by tonnage in a single year than America did in the entire post-WWII era. The Jones Act protected a domestic industry that no longer exists at scale; the trade-policy debate is largely about whether to admit that or pretend a 1920 statute is still doing the work. The chart speaks for itself.
Axios Pro Rata reports Architect Capital invested $535M in OnlyFans at a $3.15B valuation. The unit economics are excellent - high gross margin, low capital intensity, sticky creator-side network effects. The headline risk is real and persistent. We respect a sponsor willing to put the LP base through that diligence call, and we suspect future fundraisings will feature a slightly more candid "how we make money" slide.
Axios Markets pitches sports franchises as the anti-AI trade: purely human content, no automation risk, natural scarcity, rising institutional demand. WNBA valuations are up 59%; average franchise sits north of $400M. It is a Veblen good with a salary cap, and league rules restrict supply more reliably than any zoning board. The thesis survives an AI bubble bust - and probably requires one.
The Economist finds US productivity growth is running at its fastest pace in decades, driven by AI adoption, post-COVID reinvention, and structural labor shifts. The miracle is real but concentrated: software up, hospitality up, healthcare flat, construction barely positive. If the productivity bump is sectoral rather than economy-wide, the macro narrative gets harder to defend. The dispersion is the actual story.
The WSJ documents the NACHO trade taking shape on Wall Street: Non-American, Commodities, Hard Assets, Oil - gold, oil, and ex-US equities outperforming. The acronym is silly; the underlying rebalancing is not. If US debt is a problem, then anything denominated in or correlated with US fiscal pressure is the trade to fade. We are not endorsing the call - we are just suckers for a good acronym.
CPI rose 3.8% in April, accelerating from prior months as the Iran-war fuel shock works through faster than expected. Energy is now feeding services inflation, not just goods - the pass-through curve is steeper than the Fed's models implied. The September rate-cut narrative gets harder to sustain, and the term-premium response is currently doing most of the work the policy rate should be doing.
Fortune reports Martha Stewart has co-founded Hint, an AI home-management startup that predicts and prevents home repairs, backed by Slow Ventures. The preventative-maintenance market has been waiting for the right consumer brand for a decade, and Martha Stewart is, arguably, the consumer brand. Whether the AI model beats an honest home inspector with a clipboard is the product question. We're watching the renewal rate.
Axios AM charts airfares climbing on the Iran-war fuel shock: domestic round-trip economy tickets rose significantly in April, direct passthrough of the oil price spike. Airlines are the most reliable pass-through engine in the consumer economy - ticket prices move within weeks of jet fuel moving. Summer demand is the test: at what price does the consumer finally trade down.
Society and Culture
The WSJ reports Sunday Mass at St. Joseph's in Greenwich Village is now packed with young professionals, a counter-cultural spiritual revival in the city that pioneered the secular brunch.
The WSJ profiles Americans going to extraordinary lengths - group orders, special importers, airport runs - for Alphonso and Kesar mangoes before the season ends. Supply is fixed at the orchard, demand is informational, and a single airport seizure cascades through a community Slack channel. The arbitrage is whoever knows the right importer. This category will eventually have its DTC founder.
The NYT Athletic documents the $100 caviar hot dog at the Miami Grand Prix, the Golden Glizzy at the Piper-Heidsieck Champagne Garden. F1 is the testing ground for what a sports-event consumer will pay for a designed luxury hour. The Glizzy is engineered for Instagram, not gastronomy, and that is the point - the food line keeps getting more absurd because the unit economics reward absurdity.
Ezra Klein argues the AI job apocalypse probably will not happen. A Quinnipiac poll has 40% of Americans expecting AI to cost them their job within a decade; Klein thinks this fear misreads how transitions unfold. The historical record is mostly with him; the lived experience of the displaced will not be. The recovery is always geographically and educationally selective - that is the political story of 2028.
Sportico publishes WNBA team values for 2026: the Golden State Valkyries lead at $850M, average team value jumped 59% YoY, and institutional capital is flooding the league. The new CBA is the structural unlock; league pricing is still catching up to the audience. We will note that several CGP partners are extremely happy the sport their daughters play has finally been repriced as an asset class.
The Guardian reports a man has produced sperm from testicular tissue frozen as a child before chemotherapy. The clinical implication is immediate for childhood-cancer survivors. Fertility preservation moves from a niche service into a default pediatric-oncology line item, pricing in a new SKU at every NICU-adjacent practice.
Axios reports some Gen Zers are ditching social media in pursuit of better mental health - the generation that grew up on TikTok is leading the retreat. The heaviest users now have the strongest aversion. The platforms designed maximum engagement for adolescents and are now reaping the first generation of adult-survivor refugees. The advertising model that funded the entire 2010s is downstream of teenagers, and the teenagers are voting with their attention.
The Economist documents city parenting as a financial flex: raising children in major cities was once stigmatized; now it is a status symbol for dual-income professionals with private-school budgets. The status arms race relocated from the suburbs to the zip code. Per-child urban parenting cost has overtaken private-college tuition on a present-value basis - the families paying it are signaling a specific kind of capital.
Axios reports violent crime fell sharply across the largest US cities in early 2026, with murder rates at multi-decade lows. Political claim-staking will be intense in both directions, but the structural drivers - aging population, post-pandemic correction, technology-enabled clearance rates - predate the current administration. Neither party has earned the credit cleanly. The historical stress test is a recession.
Axios AM finds America's pastor pipeline is collapsing: seminary enrollment is down sharply, average pastor age is rising, and the bench is gone. The institutional church has been losing the recruiting battle for a generation. The downstream effect on community formation - weddings, funerals, the social capital congregations build - is not easily picked up by alternative providers. The vacancy is real; the replacement is not.
Business Insider reports the hacky sack has returned - a '90s artifact associated with Phish fans and parking lots is now everywhere at Gen Z high schools. The thirty-year revival cycle is compressing to about a decade, which means 2032 is bringing something we are not ready for. Our biggest question: when do we start smelling patchouli and hearing Rusted Root again?
Jonathan Haidt's latest survey of US parents finds childhood has deteriorated across almost every metric over the past decade: mental health, independence, outdoor time, community. The diagnosis is comprehensive; the policy response is fragmented. Phone bans in schools are the visible intervention, but the harder questions - housing density, free play, time poverty - are downstream of decisions no school board can make. The will is there; the policy stack is not.
One Final Thing
In our last issue, we asked how Big Tech's combined AI capex in 2027 would compare to 2026's roughly $700 billion. The vote was decisive: 73% said higher than 2026, 18% said approximately flat, and 9% said lower. The hyperscaler capex curve is now consensus higher, which is itself interesting; the question that produces a 73% answer is usually no longer the right question to be asking.
This issue's macro spine is the bond market. The US sold 30-year Treasuries at 5% for the first time since 2007, CPI accelerated to 3.8% on the Iran-war fuel shock, and Fitch is warning the US debt burden is far above AA-rated peers. The administration's playbook depends on the curve cooperating, and the curve is currently in negotiation rather than agreement. Long rates are the variable that does the policy work no one in Washington wants to do publicly.
Our question for this issue: by year-end 2026, will the US 10-year Treasury yield (a) cross 5.25%, (b) hold between 4.50% and 5.25%, or (c) drop below 4.50%? Pick a side.
We're Looking for Exceptional, Overlooked Businesses
At Civitas Growth Partners, we invest our own capital in founder-led B2B businesses with $1M to $4M in recurring revenue, strong retention, and best-in-class products, but that fall outside the traditional VC or PE mold. These businesses are often bootstrapped, breakeven, and haven't yet invested meaningfully in sales and marketing. Growth is steady (0–25% annually), and the product quietly leads its niche.
This is a category of company that is often overlooked, but we see tremendous opportunity. Our model is to help these businesses scale to $10M or more in revenue by providing flexible capital (minority or control), supporting both organic and M&A growth, and deploying an operational playbook honed through years of work with capital-efficient founders. Because we invest our own capital — not institutional LP funds — we bring no artificial timelines or exit pressure. That gives founders room to build enduring value. Our partners have achieved more than 6x equity outcomes without compromising what makes their companies special.
We're actively looking for our next great partner. If you know a founder or operator building something exceptional — and overlooked — we'd be grateful for an introduction. We value the relationships that lead us to great companies and are always happy to have a conversation about referral arrangements. Reach out anytime at team@civitasgrowth.com.
Cheers,
The Civitas Team