E135: Wagner rebels, SCOTUS ends AA, AI M&A, startups gone bad, spacetime warps & more

By roughly 10 years from this July 1, 2023 episode (i.e., by July 1, 2033), Vladimir Putin will have lost power in Russia, with the primary precipitating cause understood in hindsight to be either (a) serious illness such as cancer or (b) negative consequences arising from his 2022 invasion of Ukraine.

“I think we're going to see in the next ten years Putin lose power and he's going to be out of power. And when we look back on it, it's going to be one of two causes. It's going to be either cancer... or through his invasion of Ukraine.” View on YouTube

If the 2023 Ukrainian counteroffensive achieves a meaningful level of success (e.g., significant territorial gains against Russian forces), then within a few subsequent months Russia will respond by moving toward or declaring full mobilization and escalating the war to a higher level of violence and mobilization than seen prior to that success.

“I think this is going to put more pressure on Putin to conduct the war in a more violent way… I think he’s going to do whatever it takes to win this war. And I think you could see now, over the next few months, a full mobilization in Russia. And I think that this could lead us to the next point of escalation in this war. That is, if this Ukrainian counteroffensive actually is successful on some level… if this counteroffensive succeeds, you will see the next level of escalation.” View on YouTube

Following the June 2023 Supreme Court decision striking down race-based affirmative action in college admissions, U.S. colleges and universities that receive federal funding will, over time, be forced—via subsequent lawsuits and legal pressure—to significantly curtail or eliminate both athletic-based and legacy-based admissions preferences.

“The next step is probably going to be around athletics based and legacy based admissions… his thought on this is that those things Will go away. Because if you can’t use race based admissions to kind of balance the scales, then it’ll become pretty quick where somebody launches a legacy based lawsuit or an athletic based bias lawsuit and wins that as well.” View on YouTube

In the wake of the June 2023 Supreme Court affirmative action ruling, major U.S. corporations (e.g., large public companies like Apple, Meta/Facebook, Exxon) that operate race-based hiring, recruiting, or advancement programs will face legal challenges that will force them to materially modify or end those explicitly race-based programs; as a result, some existing DEI- and ESG-related practices and metrics that rely on explicit racial preferences will become legally impermissible in the U.S.

“The really important question after that will be what happens to companies like Apple or Facebook or Exxon, who have race based programs to try to attract African American engineers or Hispanic chemists… Will those get challenged and will those companies have to change? And my friend’s thoughts on that were that, yes, that those would also change. And that’s going to have a really important impact on private enterprise and how they approach this stuff and how DEI stuff works and frankly, downstream, how ESG works, because all these ESG check boxes now, some of them will actually become illegal.” View on YouTube

The practical effects of the June 2023 Supreme Court decision ending race-based affirmative action in college admissions will initially manifest gradually in U.S. higher education institutions, and then accelerate, and in a subsequent phase similar legal and policy changes will spread from higher education into U.S. private enterprises’ policies and programs.

“The importance of this decision can’t be really understated. It’s going to the changes will be slow and then they’ll be fast. They’ll first touch higher ed, but then I think they’ll touch private enterprise.” View on YouTube

The industry-wide shortage of GPUs will persist for approximately 1–2 years from July 2023 (i.e., through at least mid‑2024 and possibly into mid‑2025), rather than materially easing sooner.

“GPUs are basically the scarce item right now, we have a GPU shortage, and it's probably not going to get better for a year or two if that.” View on YouTube

Following Databricks’ acquisition of MosaicML and Snowflake’s acquisition of Neeva, additional similar M&A deals (AI/tooling acquisitions by data infrastructure and enterprise software companies) will occur in the subsequent years after July 2023.

“there's other companies out there like them that are also going to need to strap on tools like this to make themselves competitive in this market scape, which means that there are more acquisitions still to come.” View on YouTube

In the period after July 2023, there will be additional M&A activity in AI tooling and infrastructure, with acquired companies generally commanding high valuations relative to their current revenues, driven by strategic value to large infrastructure providers assembling end‑to‑end AI toolchains.

“I think for sure there's going to be more M&A, and I think the valuations will be high, not because these companies have a lot of revenue yet, but because it's very strategic for these big infra companies to assemble the end to end Toolchain.” View on YouTube

The existing crunch in late‑stage startup financing as of mid‑2023 will worsen over the following 18 months, i.e., through the end of 2024 (less capital available and/or on tougher terms for late‑stage rounds).

“there's a huge crunch in late stage financing. It's only going to get worse over the next 18 months.” View on YouTube

For the cohort of roughly 1,400 "unicorn" startups existing as of mid‑2023, approximately 60% will ultimately go to zero; of the remaining 40%, about half will merely return invested capital, and about half of the remainder will return around 1.5x, yielding an overall blended multiple on invested capital for that unicorn cohort of roughly 1.1x once all outcomes are realized.

“I think 60% go to zero. Of the remaining 40%, half of them probably return money. And then of the remaining half, half of those maybe get one and a half x. And then you get a geometric distribution from there, which means the blended return of that entire stream of unicorns will be about 1.1 x, but it will be very massively distributed.” View on YouTube