Unlocking AGI: How Life Changes for Everyone w/ Jack Hidary, Salim Ismail & Dave Blundin | EP #213
Summary
The panel frames AI and data-center expansion as constrained by energy hardware, but expects scarcity to flip toward abundance within roughly a decade. Diamandis contrasts China’s 429 GW of new power in 2024 with 51 GW in the US; Hidary says gas-turbine lead times have reached 4.5 years, even as SMRs, new turbine factories, and solar-material breakthroughs could change the curve in seven to nine years. “We’re in a very odd limited period of scarcity.”
If compute can travel to energy, Saudi Arabia’s reportedly three-times-cheaper solar and underused hydrocarbon resources could attract data centers whose sunk capital makes them sticky. Blundin’s reversal is that energy historically moved to populations, whereas “AI can come to the energy.” Once a massive data center is installed, he argues, it will not relocate merely because fusion later becomes available.
Abundant energy would reprice water, healthcare, transport, food, and oil-dependent sovereign economies—not merely electricity. Hidary says desalination could become far cheaper and make fresh water available; Ismail says 50% of African hospital beds are occupied because of bad-water-related infections and diseases. Hidary contrasts roughly $50-a-barrel marginal extraction in Canada with $7 in the Gulf and offers a thesis that Russia’s oil-and-gas exports could lose much of their value. “Energy tips everything.”
Humanoids look investable first in factories, logistics, and hospitals, while the household remains the contested end market. Diamandis says five robot-building-robot factories—two in the US and three in China—are under construction, with completion expected over six to 18 months; he also says about 20% of US hospitals use basic robotic helpers. Ismail’s reality check: “Can we please get the Roomba working” before forecasting robots in every room?
China’s robotics race is simultaneously an export strategy and a response to its shrinking labor pool. Diamandis cites Chinese robot-import growth of 1,700% in Poland, 275% in Mexico, 135% in Russia, and 114% in Vietnam, while advanced markets import less. Hidary predicts Chinese robots will ultimately serve even Japan’s aging population and could repeat BYD’s disruption of European automakers.
Hidary’s 2030 quantum forecast carries two opposing catalysts: commercially useful simulation and a cybersecurity rupture. Quantum could model drugs, materials, and fusion plasma at the subatomic level, but it could also break current public-key protections used by phones, WhatsApp, and blockchains. His warning is blunt: “I thought I was encrypted. You were encrypted—just now we can read it.”
Government stakes may lift strategic-technology valuations while creating a precedent the panel distrusts. With Intel shares rising after a reported 10% US government position and quantum stocks gaining 15–20% on similar speculation, Blundin calls the specific investments “fantastic” but the governing precedent “terrible.” Hidary prefers a national quantum-compute reserve; separately, he sees a path from Aramco’s $1.8 trillion market cap to a $5 trillion energy-and-compute company, while Ismail calls the valuation question irrelevant under true abundance.
Deep dive
1. Today’s energy shortage precedes a step-change in supply
Diamandis frames the bottleneck with China adding 429 GW of power in 2024 versus 51 GW in the US, alongside 37 VC-backed fusion companies, corporate nuclear purchases, SMRs, Gen 4 fission, and still-inadequate supply.
Hidary’s evidence is the gas-turbine market: gas provides about 40% of US and 35% of European electricity, yet turbines from GE, Mitsubishi Heavy, and Siemens carry roughly a 4.5-year wait. Manufacturers remember being burned by earlier expansions and remain reluctant to add capacity.
Diamandis describes a quantum-data-center builder who could not obtain the desired 5-GW generators, so bought every remaining 3-GW unit and one million valves for its liquid-cooled system. The logic was defensive procurement: “Everything was sold out. I’m not taking that chance anymore.”
Hidary expects the shortage to reverse in seven to nine years as SMRs, other nuclear capacity, turbine factories, solar, and batteries arrive. Solar conversion, currently around 27%, could move higher through materials-science and quantum breakthroughs: abundance is delayed, in his view, rather than doubtful.
2. Abundant power relocates compute and geopolitical leverage
Hidary begins with desalination: abundant energy would make fresh water much cheaper, potentially removing half of the world’s infectious diseases; he says food would also become easier and the ripple effects would propagate through society.
Ismail adds the healthcare mechanism: 50% of hospital beds on the African continent, he says, are occupied because of bad water, including infections, bacteria, viruses, water-related mosquitoes, and standing water. He calls the downstream effect multitrillion-dollar, not merely multibillion-dollar, and says cheaper energy would also reduce transport costs and change global economics.
The oil adjustment would be uneven. Hidary says oil represents 40% of Canadian exports and costs roughly $50 a barrel at the margin there versus $7 in the Gulf. His more speculative thesis is that Russia invaded Ukraine partly because, in a few years, 70% of its oil-and-gas exports may not be worth much. He also expects abundant energy to reduce the need for oil-linked aggression, while acknowledging that humans may find other reasons to fight.
Blundin sees Saudi Arabia “playing the cards absolutely perfectly”: natural gas is often wasted and oil can be expensive to export, making a local data center attractive. The discussion cites Saudi solar as three times cheaper than in the US, and Blundin notes that ACWA Power, a major solar and wind deployer, is Saudi.
The crucial infrastructure reversal is that data centers can move to energy instead of transporting energy to users. Once installed, their enormous capital base creates lock-in: “The data center is not going to move,” even if fusion subsequently changes global power economics.
3. Humanoids monetize controlled environments before homes
Diamandis sets the bull case with Elon Musk’s forecast of 10 billion humanoids by 2040 at $20,000–$30,000 each. He also cites 1X offering its NEO Gamma robot at $20,000 or $499 monthly—eventually implying labor near $10 a day or 40 cents an hour, running “GPT-5, GPT-6, or GPT-7” or Grok-level intelligence.
Diamandis says the first five factories where robots build robots—two American and three Chinese—have begun construction and could be ready in six, 12, or 18 months. He argues that robot-made robots, combined with LLMs that can see and act without task-by-task coding, point toward “iPhone production scales,” though he does not endorse Musk’s exact volume.
Near-term deployment is less glamorous: warehouses, logistics, and hospital delivery of linens, food, towels, and medicine. Diamandis says roughly 20% of US hospitals already have some robotic helper and predicts numerous robots at the same conference within two years, perhaps even presenting onstage.
Ismail accepts industrial demand for “dull, dirty, dangerous jobs” but doubts supply chains can support a billion robots soon. Homes contain children, pets, and far more variables than factories; Diamandis says 1X may want that diversity as training data, while Ismail preserves the risk with the dark joke: “They may injure some kids on the way.”
4. China’s robot push is necessity as much as industrial ambition
Diamandis estimates about 30 US humanoid-robot companies, says China has at least that many, and says he is tracking on the order of 100 humanoid robots, many from China. Chinese imports are surging most in less technologically advanced buyers—Poland up 1,700%, Mexico 275%, Russia 135%, and Vietnam 114%—rather than South Korea, Germany, or the US.
Hidary explains the demographic and industrial pressure: China’s decades of 10%-plus growth relied heavily on a 1.4-billion-person manufacturing labor pool that is now contracting. China must automate to preserve its advantage; the US needs robotics for reshoring because it lacks that labor pool. Blundin agrees with the diagnosis.
Hidary uses Japan as the warning against arriving too early. Japan had a head start with Honda’s ASIMO and expected robots to help care for an aging population as its population shrinks from 145 million to 90 million. His prediction is that “they’re going to be Chinese robots, not Japanese robots.” He expects some coexistence in the US, where more expensive Tesla, Figure AI, and other robots may serve sensitive supply-chain and defense applications.
5. Quantum’s commercial and security shocks converge around 2030
Peter says Google’s Willow chip reached what Google calls quantum advantage—a marketing term—and announced a 13,000-fold increase over classical computing. Hidary predicts a “usable quantum computer” by 2030—not the ultimate machine, but one capable of meaningful business and societal impact. It could simulate cancer and Alzheimer’s drugs or new carbon composites at the subatomic level, and he expects several winners because “this is not a winner-takes-all game.”
Quantum sensing is already here today in Hidary’s telling, reading the heart’s magnetic field more effectively than an ECG or MRI. Because humans cannot natively interpret quantum outputs, AI is essential; Sandbox AQ’s name encodes that thesis—“A for AI, Q for quantum”—as the “twin engines” for modeling and changing the physical world.
Blundin sees the biggest upside in “nonintuitive, nonhuman domains”: programming quantum machines or controlling fusion’s complicated magnetic fields. This is additive rather than substitutive AI—“It’s not replacing humans”—because it makes previously inaccessible work possible.
Hidary expects two ChatGPT moments. First, hackers could expose government and corporate secrets once scalable, cloud-accessible quantum systems defeat today’s encryption; people who failed to protect themselves will discover, “I thought I was encrypted. You were encrypted—just now we can read it.” Second, quantum simulation could model fusion plasma and the laser-driven reaction at the National Ignition Facility well enough to accelerate fusion by 10–20 years. That closes the episode’s loop: quantum helps create the energy abundance that, in turn, powers more compute.
The ownership debate remains unresolved. After Peter cites a 10% US government position in Intel and a 15–20% rise in quantum-computing stocks on possible government positions, Blundin calls the investments brilliant but the precedent terrible, warning that defense implications could lead toward nationalizing the sector and restricting innovation. Hidary agrees that government attempts to pick winners have a horrific history, but proposes a strategic quantum-compute reserve without government ownership. His $5 trillion candidate is an Aramco- or Exxon-like company that expands from energy and chemicals into compute—prompting Ismail’s dissent: under genuine abundance, “the whole economy becomes meaningless.”