Book Summary
The Coming Wave starts from a simple observation with vertiginous implications: a wave of transformative technology — artificial intelligence first, synthetic biology close behind — is about to reshape economies, states, and everyday life on a timescale of years, not decades. Mustafa Suleyman’s central claim is that containing this wave — capturing its benefits while bounding its harms — is the defining challenge of the twenty-first century. And his warning, drawn from the history of every powerful technology that came before, is that containment almost always fails.
The wave metaphor does heavy lifting, and it earns it. Suleyman treats transformative technologies as forces with their own momentum: fire, agriculture, the printing press, the steam engine, electricity — each arrived as a wave that societies could ride or drown in, but never simply opt out of. This wave differs in three ways that matter: it is fast, it is cheap, and it proliferates. The cost of creating powerful capabilities keeps collapsing, which means capability spreads outward — from states to companies to small groups to individuals — faster than our institutions can adapt.
The book is unsparing about both sides of the ledger. On the promise side: scientific discovery accelerated by orders of magnitude, cures for intractable diseases, energy abundance, material plenty — a plausible path to a world of radical abundance. On the peril side: engineered pandemics, cyberattacks executed at machine speed, autonomous weapons, disinformation at industrial scale, and the longer-run alignment problem of systems pursuing goals we did not choose. Suleyman’s point is that the promise and the peril are the same technology. You cannot have one wave without the other.
For the investor reader, this is a framework disguised as a policy book. Waves concentrate returns in their enabling layers before diffusing them everywhere — which tells you where to look first. The containment problem is the regulatory overhang hanging over every big-tech valuation: the multiples assume the current regime persists, and Suleyman’s history suggests regimes built on uncontained powerful technology do not persist. And the proliferation dynamic structurally favors offense over defense, which has direct implications for where security spending — public and private — has to go.
Who is Mustafa Suleyman?
Mustafa Suleyman co-founded DeepMind in 2010 with Demis Hassabis and Shane Legg — the London lab behind AlphaGo, the program whose 2016 victory over Lee Sedol announced the deep-learning era to the world. Google acquired DeepMind in 2014. Suleyman later co-founded Inflection AI in 2022, which raised over a billion dollars to build personal AI assistants, and in 2024 he joined Microsoft as CEO of Microsoft AI. He is that rare thing: an insider to the technology writing critically about its trajectory.
He has been a consistent voice on AI governance for years, arguing — well before it was fashionable — that the industry’s capabilities were outrunning its safety work and its oversight. The Coming Wave, published in 2023, is his fullest statement: part history of technology, part risk assessment, part containment agenda. It was written from inside the industry’s inner circle, which is both its authority and its complication.
The complication deserves a sentence. Suleyman builds the systems he warns about, and now runs AI for one of the world’s largest technology companies. Readers should hold that in mind when weighing the prescriptions: this is a builder arguing for restraint on building, and the tension is never fully resolved. It does not invalidate the diagnosis — insiders often see the risks first — but it is context the book itself does not dwell on.
Lessons From The Coming Wave
Containment is the trade. Suleyman’s history of powerful technologies is essentially a catalog of containment failures and near-misses — controls imposed, then eroded, then overwhelmed by proliferation and incentives. For investors, the translation is direct: price the regulatory overhang. AI leaders trade at multiples that assume today’s operating environment persists indefinitely. The book’s argument is that uncontained transformative technology never leaves the operating environment unchanged — and the adjustment, when it comes, will not be priced in advance.
Bet the enabling layer. Every technological wave rewards its picks and shovels first: the wave’s infrastructure captures the early rents before the benefits diffuse. In this wave, that means compute, models, and tooling — which is why semiconductors have become the central geopolitical asset of the decade, the subject of Chip War. Follow the bottleneck: whoever controls the scarce input to the wave collects the toll.
Falling costs favor the small and the hostile. When powerful capability gets cheap, it flows downhill — to startups, to criminal groups, to rogue states — and offense gains on defense across every domain the book surveys. The investment implication is structural, not cyclical: demand for cybersecurity, verification, and defense technology grows with the wave itself, not with any particular conflict. In a proliferating-capability world, security spending is the closest thing to a permanent bid.
Plan for the bad branch. Suleyman practices what he calls pessimism of the intellect: assume proliferation happens, assume controls erode, and ask what survives. Most portfolios are a pure long bet on the wave — long the beneficiaries, with no hedge against the disruption the book describes. Investors call the missing piece tail-risk hedging, and almost nobody holds it. The book will not tell you how to hedge; it will convince you that unhedged exposure to a transformative wave is a choice, and an expensive-looking one.
Race dynamics are multipolar traps. US-China competition accelerates deployment while degrading the cooperation that containment requires — each side’s safety restraint looks like unilateral disarmament. Markets should therefore discount rosy “responsible AI” narratives from any single lab or company: the incentives of the race punish restraint and reward speed. When evaluating AI leaders, weight the competitive dynamics heavier than the corporate messaging.
Criticisms of the Book
Diagnosis over prescription. The book’s containment agenda — international treaties, technical controls on models and compute, corporate restraint, public oversight — is sketched rather than engineered. Suleyman is far stronger on why containment historically fails than on what would make it work this time, and the gap matters: a reader convinced by the first 250 pages will find the final “what to do” section the least persuasive part of the book. Naming the problem is not solving it.
The insider’s blind spot. For a book about power, there is strikingly little examination of whether “containment” is compatible with the commercial incentives of the labs racing to deploy — including the ones Suleyman has run. The argument asks companies to restrain the very capabilities their valuations depend on, and never fully confronts that contradiction. Skeptics will read the prescriptions as reputation management: concern as a moat, safety as a brand.
It underweights the state. The containment mechanisms proposed — treaties, verification regimes, international norms — require a level of great-power trust that does not currently exist, and the book gives the US-China rivalry less weight than it deserves. Technology control regimes historically work between allies and fail between adversaries; this wave’s decisive competition is between adversaries. That is the hardest page of the problem, and the book turns it a little too quickly.
Who is This Book For?
For the tech investor who wants the risk case to balance the bull case. If your portfolio is long the AI wave — and at this point, most equity portfolios are — this book is the clearest statement of what could go wrong and why the current regulatory calm may not last. Read it alongside Chip War for the full picture: one book on the technology’s trajectory, one on the physical bottleneck it runs through.
For the AI-curious reader who wants the policy debate in one serious volume. Suleyman writes more carefully than the alarmists and more urgently than the boosters, and the history-of-technology chapters stand on their own. Readers of Homo Deus will find a kindred argument here — Harari’s dataism as philosophy, Suleyman’s containment as policy — and the two books argue productively with each other.
Not for readers seeking technical depth on how the models actually work, or a guide to investing in AI directly. This is a book about power and governance, not a stock-picking manual — it will change what you worry about owning, not tell you what to buy. And not for committed techno-optimists: if you believe every technology problem has a technology solution, the book’s central metaphor will feel like fatalism.
Final Thoughts
The Coming Wave is the most serious policy-minded AI book yet written by a genuine industry insider, and the wave framing is its durable contribution. Whether or not containment proves possible, thinking in waves — momentum, proliferation, diffusion, and the institutions that always arrive late — is now the right way to think about technology’s collision with markets and states.
The book’s value for the money reader is as a risk framework, not a forecast. It will not tell you which lab wins or what the models can do in 2030. It will tell you that transformative technology redistributes power faster than institutions adapt, that the adjustment is never priced in advance, and that the current calm around AI leaders is a bet on containment holding. Name the bet explicitly, and size it accordingly.
Read it, argue with the prescriptions, keep the diagnosis. The promise half of the ledger is real — the abundance Suleyman describes would be the largest wealth-creation event in history — but so is the peril half, and markets are currently priced as if only the first half exists. The Coming Wave is the corrective: the full ledger, from someone who helped write it.
