2024 in Review: Deep Tech Startups & Venture Capital - Chapter 2
From AI Overconcentration to Nuclear Resurgence
✨ HI There! This is a Free Edition of The Scenarionist. To enjoy the full experience, become a Premium Member!
The Scenarionist Premium is designed to make you a better Deep Tech Founder, Investor, and Operator. Premium members gain exclusive access to unique insights, analysis, and masterclasses with the wisdom of the world’s leading Deep Tech thought leaders. Invest in yourself, and upgrade today!
Welcome to Chapter 2 of 2024 in Review: Deep Tech Startups & Venture Capital.
If Chapter 1 distilled the year into its most striking data points, this chapter takes us beneath the surface. The real story of 2024 wasn’t just in the headlines or the numbers—it was in the themes that threaded their way across industries. These aren’t isolated events. They’re forces. Movements. Gravitational pulls that reshaped capital, strategy, and innovation.
As we laid out in Chapter 1, this series isn’t a tidy chronology of the year’s biggest moments. Instead, we’re following the cross-cutting themes that defined 2024. These themes aren’t just stories; they’re patterns that influenced who got funded, which sectors were prioritized, and whose ideas got the green light. Think of them as the undercurrents pulling at the fabric of deep tech, often unseen but impossible to ignore.
In total, this report covers 16 key themes, spread across four chapters. Today, we focus on the five foundational forces that set the tone for the year. These are the themes that dictated how the first dominoes fell, rippling through the rest of the ecosystem.
Here’s what’s ahead:
The Five Themes of Chapter 2
Theme 1: The AI Overconcentration
Theme 2: China-Free Supply Chains
Theme 3: Energy Megaprojects & Uncertainties
Theme 4: Industrial Automation’s Acceleration
Theme 5: Nuclear Resurgence—Struggling for Small Modular Reactors
What to Expect in Chapter 2
This chapter unpacks how these forces collided and coalesced, shaping deep tech in 2024. We’ll look at how AI mania pulled resources away from other industries, how geopolitics forced startups to rethink their supply chains, and how energy innovation grappled with the challenge of scaling moonshot ideas.
For industrial automation and nuclear, the stories are different—less about hype, and more about grinding through the realities of execution. These themes show us where deep tech excelled, where it faltered, and where the next set of opportunities might lie.
Chapter 2 isn’t just a recap; it’s a map—a way to understand the forces that defined the year and set the stage for what comes next. Let’s dive in.
✨ Theme 1: The AI Overconcentration
If 2023 was the year generative AI popped into mainstream culture, 2024 was when that excitement metastasized into a near-obsession. By Q1, the sheer scale of capital chasing AI deals began to eclipse entire industries. At first, it felt like a mere extension of last year’s frenzy—until we saw the numbers come in.
OpenAI’s $6.6B round, pushing its valuation north of $157B, was the single biggest private funding event across all deep tech categories in 2024. People in the industry described it like a black hole at the center of the venture galaxy: everything else struggled to get recognized or financed.
By Q2, the velocity of AI deals was almost comedic—eight-figure rounds felt small, overshadowed by the perpetual question, “Could you be the next GPT-level success story?”
Yet while OpenAI hogged the limelight, Databricks quietly demonstrated that big enterprise AI can come with big revenue. Their $10B raise pinned a new valuation at around $62B, but the more impressive figure was nearly $3B in annual revenue—real money from real clients who rely on Databricks for day-to-day analytics.
The company’s near-profitable status validated the notion that not all AI capital is hype. Where OpenAI was building futuristic models, Databricks was monetizing the “infrastructure picks and shovels” behind the AI gold rush.
Then came SoftBank’s vow to invest up to $100B in U.S. AI projects. That figure was not a single round, obviously—more a multi-year pledge—but its symbolic weight was immense. Founders and VCs joked that if you wanted to raise capital in 2024, you could do worse than throw in an “AI-driven” or “foundation model” somewhere in your pitch deck. Ironically, as the year progressed, smaller advanced-energy or manufacturing startups lamented that they’d wasted months talking to investors who suddenly pivoted to pure AI hunts.
Some began rewriting their own narratives: a microreactor startup might highlight how nuclear can feed AI data centers 24/7, while an Agritech venture might emphasize AI analytics in vertical farming. It was, to many, the only way to command attention in the shadow of these multi-billion-dollar AI headlines.
The Impact on Everyone Else
This “AI black hole” phenomenon wasn’t just about overshadowing headlines; it shaped supply chain dynamics too. The HPC (high-performance computing) hardware pipeline started to buckle under the new demand. GPU supply became a premium conversation topic, with even well-funded AI ventures struggling to secure enough chips. For segments like quantum computing or advanced robotics that rely on specialized HPC resources, it grew tougher to navigate lead times. Some formed alliances with AMD or smaller chipmakers out of necessity, while big names like Nvidia appeared to hold all the cards. If you were building an industrial automation system that also needed HPC cycles, you found yourself queuing behind more purely AI-driven operations.
Meanwhile, those who questioned the sustainability of these mega AI valuations pointed to analogies with the dot-com era’s inflated expectations. But for 2024, the capital felt undeniably real. Databricks, with its strong revenue fundamentals, served as a counter-example to pure hype. Sure, there was worry about a bubble, but that worry didn’t stop big checks from closing. Practically every top 10 venture round this year was AI-related, crowding out the potential for equally large deals in nuclear or energy storage.
If 2023 gave us a taste, 2024 codified it: the epicenter of capital formation in deep tech was undeniably AI, and those not in the gravitational well had to scramble, pivot, or resign themselves to smaller raises. It was a bonanza for HPC data center buildouts and a trial by fire for emerging battery lines, industrial robotics, or space startups trying to secure series B or C expansions. “Unless you show how your solution is fueling an LLM or enabling HPC,” one founder quipped, “the conversation is 90% shorter.”
Where Does That Leave Us? By the end of the year, AI leaps forward in HPC expansions, language model training, and real enterprise traction. But it’s also left entire sub-industries underfunded or overshadowed, waiting to see if 2025 might bring a rebalancing or if AI mania only grows stronger.
✨ Theme 2: China-Free Supply Chains
While AI soared, another storyline bubbled up from a very different corner: the race to extricate advanced technologies from Chinese suppliers. It wasn’t an overnight phenomenon—tensions had been building for years over intellectual property disputes and trade imbalances—but 2024 saw that slow burn transform into a full-on mania for “China-free.” Governments, especially in the U.S. and Europe, not only signaled they’d pay a premium for domestic or allied sources, they began embedding those requirements into procurement standards, particularly for defense and critical infrastructure.
A New Golden Age for Local Manufacturing?
One of the more illustrative examples is Orqa, a Croatian startup that netted €5.8 million to manufacture drones with strictly non-Chinese parts. If you’d told most VCs five years ago that there’d be a robust market for higher-cost, 100%-EU-manufactured drones, they’d likely shrug. But the geopolitics changed.
Defense agencies across NATO countries no longer trust Chinese UAV components, seeing them as potential backdoors for espionage. Orqa’s selling proposition became straightforward: yes, you’ll pay more, but you get guaranteed supply chain security and zero risk of questionable hardware. Founders at Orqa recounted how quickly their pipeline of potential defense clients grew once the brand “China-free” took hold. The question is whether these cost premiums become sustainable if the political winds shift or if budgets tighten.
Rare Earths, Battery Lines, and the Northvolt Echo
Beyond UAVs, the push to dethrone Chinese supply lines reverberated through advanced materials and battery manufacturing. Tidal Metals, for instance, pitched an $8.5 million seed effort to extract magnesium from seawater rather than rely on heavily polluting land mines often controlled by China.
Meanwhile, recycling outfits like Cyclic Materials set up local processes to reclaim critical metals from e-waste, again marketing themselves as “less reliant on Chinese refining.” There’s a sense of empowerment here—maybe the West can wean itself off a behemoth supply chain and forge a greener, more autonomous future in the process.
Yet the meltdown of Northvolt hangs over these efforts like a cautionary tale. The Swedish gigafactory once symbolized Europe’s bid to go head-to-head with Chinese battery giants, only to fold under cost overruns and the brutal reality that Chinese packs dipped below $50 per kWh.
Founders in the newly minted local supply chain wave speak with a mixture of pride and humility, aware that while “China-free” can garner big initial checks, delivering cost parity is an entirely different game.
Defense-Driven Demand vs. Commercial Realities
In 2024, the biggest impetus for “China-free” almost always came from defense or government channels. If your drones or electronics were slated for a Department of Defense contract, you’d likely find guaranteed demand. But consumer or commercial markets, especially in cost-sensitive verticals, remained more price-conscious, and slower to adopt local but pricier supply solutions.
Founders across Europe and the U.S. found themselves leaning on public grants, strategic funds, or direct defense deals to offset that 20–30% cost markup. This, in turn, created a sense of moral hazard: if the cost premium is always cushioned by the public sector, do these supply lines ever become truly competitive on open-market terms? No one quite had the answer. The vibe at year’s end was that “China-free” mania would continue at least as long as Sino-Western relations remain tense, and as defense budgets keep swelling.
✨ Theme 3: Energy Megaprojects & Uncertainties
If AI mania hogged the headlines and “China-free” dominated policy chatter, the energy sector tried its best to remain in the spotlight. On any normal day, projects involving multi-billion-dollar nuclear expansions or drilling into magma would be the year’s biggest stories. But in 2024, they had to shout to be heard.
Magma Dreams and Geothermal Gains
At one extreme, Magma Power unveiled a $20 billion plan to exploit magma’s heat for approximately 10 gigawatts of baseload power, targeting the insatiable energy appetite of AI data centers. Supporters claimed that tapping the near-limitless thermal energy beneath Earth’s crust could provide a stable, carbon-free power supply far surpassing most renewables. However, skeptics highlighted the colossal engineering unknowns: how do you drill into magma flows without frying equipment or risking volcanic instability?
While the company insisted it had secured pilot funds and was developing specialized drilling rigs, even its proponents acknowledged commercialization was likely years or decades away.
More grounded progress came from modest geothermal expansions led by players like Fervo Energy and Gradient Geothermal. Their deals with tech giants such as Meta and Google signaled a strong demand for round-the-clock renewable power. While the geothermal sub-sector raised around $700 million since 2020—a fraction of AI’s billions—it proved to be one of the more stable and reliable green energy solutions. The contrast was clear: companies like Fervo and Gradient targeted well-understood geothermal reservoirs, not the volcanic extremes pursued by Magma Power.
Battery Lines: Grand Plans, Grand Failures
The push to localize battery supply chains away from China brought both ambition and caution. Lyten teased a potential $650 million investment from the U.S. Export-Import Bank to commercialize lithium-sulfur designs. If successful, lithium-sulfur could surpass traditional Li-ion batteries with better energy density and safer thermal profiles. But that’s a big “if.” While lab results look promising, scaling these designs to gigafactory output remains unproven.
On the darker side, the meltdown of Northvolt loomed large. Once the centerpiece of Europe’s battery self-sufficiency ambitions, Northvolt succumbed to cost overruns, competition from $45-per-kWh Chinese battery packs, and perhaps naive optimism. This collapse served as a stark warning: local battery manufacturing may look good on a policy deck, but without competitive cost structures, projects risk financial disaster. Founders of smaller battery startups now reference “the Northvolt effect” to describe the heightened investor scrutiny that followed.
Fusion and Microreactors
Fusion energy, often labeled as the perennial “energy of the future,” made incremental progress in 2024. Companies like Focused Energy and HB11 (working on hydrogen-boron fusion) secured modest funding rounds in the tens of millions but remained overshadowed by the AI frenzy. Despite the buzz, commercial fusion remains at least a decade away.
In the meantime, microreactors within the nuclear sector showed more immediate promise. Oklo announced a 20-year offtake MOU with Switch Inc., designed to power data centers, while NuScale opened a training center at The Ohio State University to address the nuclear workforce gap.
However, the thorny licensing process continues to stall widespread deployment. Observers wonder if enthusiasm for microreactors will persist given the five-to-seven-year delays often caused by regulatory inertia. With AI continuing to absorb capital and attention, nuclear energy faces an uphill battle in securing the funding needed for these capital-intensive projects.
The Mood
As 2024 came to a close, energy watchers expressed mixed feelings: a mix of excitement over the scale of new proposals—magma drilling, microreactors, and advanced batteries—and sobriety over cautionary tales like Northvolt’s collapse and other failed ventures. Many noted that these ambitious projects were often overshadowed by the relentless spotlight on AI. In a typical year, energy megaprojects might dominate the conversation; in 2024, they often felt like second-tier stories.
Looking ahead, investors are torn: will 2025 bring a more balanced approach that renews focus on energy innovations, or will AI mania and the challenges of localized supply chains continue to hamper the scale-up of advanced energy solutions?
✨ Theme 4: Industrial Automation’s Acceleration
While advanced nuclear or fusion soared in “visionary” territory, a quieter corner of deep tech found a surprisingly stable path: industrial automation. For many mid-sized factories, distribution centers, or supply chain nodes, 2024 was about adopting new robotic solutions not for hype, but for real ROI.
It wasn’t a domain that attracted multi-billion-dollar raises, but the sums it did attract often translated directly into cost savings for paying clients—a dynamic that stands in stark contrast to the more intangible promises of AI or fusion.
Slip Robotics and the Truck-Loading Revolution
One success story retold at every logistics conference was Slip Robotics. The company’s “SlipBots” effectively reduce trailer loading times from an hour to a mere five minutes, opening up an immediate throughput advantage. They raised a robust, if not earth-shattering, $28 million in a Series B. But the interesting part is that their customers—Nissan, John Deere, among others—raved about productivity gains. This wasn’t future speculation; it was a daily operational improvement.
The founders recounted how, at every new deployment site, the data is “unambiguous.” Instead of guesswork, they show managers the real-time shaved off every loading cycle, which translates directly to the bottom line.
T-robotics and GPT-Fueled Simplicity
While Slip attacked logistics, T-robotics aimed at the programming friction that left many industrial arms idle, imagine a mid-sized factory that bought a fancy $100,000 robot arm but rarely uses it beyond a single repetitive task because reprogramming is too complex or expensive.
T-robotics introduced “ActGPT,” a natural-language-based system that interprets everyday prompts—like “move this part from Station A to Station B, then rotate it 90 degrees”—and translates that into motion scripts. The entire concept garnered a seed round of $5.4 million.
Not a mind-blowing sum, but enough to scale up pilots. In an environment starved for capital outside AI, T-robotics managed to stand out by focusing on real, near-term ROI for an underserved manufacturing base.
The Pragmatic Allure of Industry 4.0
Though overshadowed in the press by generative AI mania, the domain of industrial automation found a comfortable niche. Factories always love cost reductions, and advanced robotics have become cheaper and more user-friendly. The synergy with HPC or AI arises too—some factory solutions harness data analytics, minor machine learning, or HPC scheduling, though rarely in the generative AI sense.
The important part is the revenue: these solutions do not rely on vague “potential.” They invoice clients who see payback in months. That level of tangibility made it the bright spot for certain VCs wary of purely hype-driven subfields. The question is whether large-scale adoption can outpace the overshadow from AI. So far, 2024 ended with a quiet sense that industrial automation might soon become a more mainstream darling, especially if hype in other corners eventually cools.
✨ Theme 5: Nuclear Resurgence—Struggling for Small Modular Reactors
Mention nuclear, and you conjure images of fission plants from decades past or futuristic talk of fusion. Yet 2024 gave us the subtheme of nuclear microreactors or SMRs stepping forward with fresh deals and expansions. Policymakers talked about “resurgence.” Startups hammered out pilot contracts. Universities opened training centers. On paper, the momentum looked unstoppable.
The reality, though, is that scaling nuclear—especially next-gen formats—remains a labyrinth of licensing, cost, and local acceptance. If the theme of energy megaprojects painted a broad story of ambition vs. friction, nuclear microreactors deserved a dedicated lens because they exemplify that tension so perfectly.
NuScale, Oklo, and the Dreams of 24/7 Carbon-Free Power
Two names kept surfacing in 2024’s microreactor conversation: NuScale and Oklo. NuScale garnered attention with the Energy Exploration (E2) center at The Ohio State University, an initiative pitched as a workforce pipeline and demonstration hub for its 12-module SMR design. The idea is a standardized, factory-assembled nuclear reactor that can be slotted into local grids to replace retiring coal plants or power new data centers. For a public eager for zero-carbon electricity, SMRs sound like a fix. But these designs still face multi-year regulatory reviews from the NRC, and each site or design tweak can trigger fresh scrutiny.
Oklo, in parallel, boasted a 20-year offtake arrangement with Switch Inc., promising microreactor-powered data centers. The synergy is obvious: AI expansions demand crazy amounts of power; microreactors promise an always-on, carbon-free supply.
The catch? The contract is “non-binding,” the timeline for deployment remains fuzzy, and Oklo still has to finalize crucial licensing steps. That didn’t stop the hype machine from churning. Some see microreactors as a near-future reality that can scale faster than conventional nuclear plants, while others fear that if licensing or cost issues drag on, the industry’s momentum might fade.
The Workforce and Licensing Choke Points
NuScale’s E2 center signals more than just a new building on a university campus. It’s an admission that the nuclear industry has a knowledge gap. The number of skilled nuclear engineers, technicians, and operators has waned over the years. If you plan to deploy microreactors in dozens of sites, you need a robust pipeline of trained professionals. That means forging partnerships with universities, awarding scholarships, and building simulators. Meanwhile, every time a microreactor design changes something in its cooling system or tries a new fuel approach, the regulatory process resets in part. The friction is enormous.
Public Perception: A Low-Key Struggle
Beyond workforce and licensing, microreactor advocates face the intangible but powerful force of public opinion. Even meltdown-resistant reactors, using advanced fuel forms like TRISO, can scare local communities.
The meltdown fiascos at big-scale plants decades ago left a lingering distrust that new nuclear has to surmount. SMR boosters argue their smaller footprints are safer and can’t melt down in the old sense.
But explaining that to local governments or activists is an ongoing challenge. It’s no shock that some microreactor firms spend as much time on PR and community outreach as they do on engineering.
Why This Resurgence Matters
The broader climate conversation provides the backdrop. Renewables like wind and solar are intermittent, grid-scale batteries remain costly, and data center expansions demand a stable baseload. If microreactors can deliver that stable power, you solve a chunk of decarbonization in one swoop.
That’s why governments keep funneling R&D grants, and big companies entertain off-take MOUs. But from a 2024 vantage, we remain in a place of partial momentum. Enough capital to keep the dream alive, enough red tape to keep it from blossoming too fast.
💡 Conclusion: A Year in Overdrive
By weaving through these five themes—AI Overconcentration, China-Free Supply Chains, Energy Megaprojects & Uncertainties, Industrial Automation’s Acceleration, and Nuclear Resurgence—we’ve tapped into the central forces that defined deep tech in 2024.
Each thread intersects with the others. AI’s capital overshadowed alternative subfields but also demanded stable energy. China-free mania reshaped battery lines and drone supply chains, while industrial automation solutions quietly delivered near-term ROI. Nuclear innovation aimed to feed AI’s insatiable appetite for clean baseload power, yet wrestled with licensing delays. It’s all connected.
If there’s a single word that captures the vibe of 2024, it might be “overdrive.” Funding for AI soared to previously unimaginable levels. Ambitions in nuclear, geothermal, and advanced manufacturing climbed equally high. Geopolitical tension fueled calls to disentangle from Chinese supply chains. And yet, alongside this explosive growth, there were fiascos: Northvolt’s collapse, an ill-fated “underwater data center” pilot in San Francisco Bay blocked by environmental regulators, partial eVTOL bankruptcies, and more. It was a year of extremes—massive checks that dwarfed entire industries and cautionary tales that underscored how precarious scaling can be.
In Chapter 3, we’ll continue with the next suite of themes: from the tug-of-war between Innovation and Regulation, to defense-tech expansions driven by evolving security threats, to breakthroughs in space commercialization and orbital security.
Stay tuned. There’s more to cover on how these threads unfold, how themes like advanced regulation and defense synergy took root, and how the sector collectively grapples with the question: are we on the cusp of a deeper transformation, or riding a wave that might recede if the capital pivot shifts again?
We’ll pick that up in Chapter 3! :)