The Abyss Called Deep Tech — The Tectonic Shift Happening in 2026
The deep tech landscape has transformed into an entirely different terrain over the past 18 months. According to Boston Consulting Group's long-term tracking, deep tech's share of global venture funding has doubled from approximately 10% in 2014 to over 20% in 2024, with multiple institutions sharing estimates that the full year 2025 will reach a scale of $48 billion (approximately ¥7.2 trillion). Data from Q1 2026 simultaneously published by Crunchbase and KPMG Venture Pulse shows that global VC investment hit a record high of $300–$331 billion (approximately ¥45–49.65 trillion), yet an extreme concentration is occurring in parallel, with just four companies—OpenAI ($122 billion), Anthropic ($30 billion), xAI ($20 billion), and Waymo ($16 billion)—absorbing $188 billion (approximately ¥28.2 trillion), nearly 65% of the quarterly total.
Conversely, "money not riding the generative AI wave" is seeping into deeper physical, biological, and chemical layers. The 2026 Deep Tech Market Report by BCG and StartUs Insights puts forth "Hardware-Software Convergence (the fusion of the physical and computational)" as its central thesis, explicitly identifying 12–15 year fund lifespans, stage-gate capital deployment, and early co-creation with regulators, governments, and major industries as essential conditions. In a post-GTC 2026 report, Celesta Capital stated that "physics is an inescapable ceiling, and betting on it is the most effective hedge against the AI boom." In a field exceeding approximately ¥7 trillion in yen terms, the four domains covered in this article—fusion energy, quantum, synthetic biology, and advanced chemistry—occupy its very core.
Fusion — The Dawn of "Sun-as-a-Service"
The 6th Global Fusion Industry Report, published by the Fusion Industry Association (FIA) in July 2025, reported that private fusion companies worldwide raised a new $2.64 billion over the preceding 12 months, bringing cumulative funding to $9.7 billion. Responding companies numbered 53, employing 4,607 people directly and more than 9,300 in supply chain roles — more than double the 23 companies counted in 2021. The report also estimated that the total additional capital required by these companies to bring their first commercial plants online is approximately $77 billion, making clear the reality that "funding demand will expand to eight times its current level over the next five years."
At the center of this activity is Commonwealth Fusion Systems (CFS). On August 28, 2025, CFS raised $863 million in a Series B2 round, bringing its cumulative funding to approximately $2.9 billion. Notably, new investors included Nvidia and Morgan Stanley, as well as a Japanese consortium of 12 companies centered on Mitsui & Co. and Mitsubishi Corporation (including Tokyo Electric Power and Sumitomo Corporation). In September 2025, CFS signed a power purchase agreement (PPA) exceeding $1 billion with major Italian energy company Eni, and in July of the same year, Google signed what was described as "the largest direct corporate offtake contract in fusion history," purchasing 200 MW from the ARC pilot plant (in Chesterfield County, Virginia). CFS also agreed with Dominion Energy on joint site development. The MIT-spinout company aims to bring its demonstration device SPARC online within 2026 and to achieve net energy demonstration in the first quarter of 2027. In a collaboration with Google DeepMind, a formal agreement was signed in September 2025 to use the open-source plasma simulator TORAX to optimize SPARC's magnetic field geometry and plasma operating modes.
Helion Energy is backed by OpenAI CEO Sam Altman, Reid Hoffman, Mithril Capital, Capricorn, BlackRock, and KKR, among others. In a Series F in January 2025, the company raised $425 million, bringing cumulative funding to $1.03 billion. The company's prototype "Polaris" was physically completed at the end of 2024, and on February 13, 2026, the company announced it had achieved DT combustion at temperatures on the order of 150 million degrees Celsius (referenced in multiple sources including Wikipedia and the Helion newsroom). Helion has a PPA with Microsoft for 50 MW beginning in 2028 — recognized industry-wide as a standard benchmark as "the first PPA in history to promise commercial power generation in the 2030s."
Also noteworthy is Pacific Fusion's Series A of over $900 million, led by General Catalyst — the largest single funding round ever for a fusion company. The round is structured as tranche payments tied to milestone achievements. The company uses inertial confinement via coordinated electromagnetic pulses (pulsed-power driven), rather than lasers. In partnership with General Atomics, it is advancing toward production-scale pulser module testing, with plans to demonstrate net gain at the device level by 2030.
Germany's Marvel Fusion has raised a cumulative $162 million, securing €113 million in a Series B from EQT Ventures, Siemens Energy, and the European EIC Fund. In partnership with Colorado State University, it plans to establish a laser inertial confinement demonstration facility by 2027. Xcimer Energy has raised approximately $100 million from Hedosophia, Breakthrough Energy Ventures, Lowercarbon Capital, Emerson Collective, and Gigascale, and is designing a 10 MJ laser — five times the scale of the U.S. National Ignition Facility (NIF). Type One Energy was scooped by TechCrunch in January 2026 for an $87.5 million bridge round, positioning the company for a Series B of approximately $250 million. It plans to break ground in 2026 on its stellarator prototype "Infinity One" at the former Bull Run power plant site of the Tennessee Valley Authority (TVA), with its commercial machine "Infinity Two" (350 MWe) targeted for operation in the mid-2030s.
On the Japanese side, Kyoto Fusioneering announced in September 2025 a Series C extension of ¥9.3 billion combining equity and debt, bringing cumulative equity to ¥16.2 billion. An independent estimate by the Nikkei Shimbun put the company's valuation at ¥72.1 billion as of September 2024, placing it among unicorn candidates. In November 2025, the company completed a conceptual design report (CDR) for the FAST demonstration project — Japan's first privately led cross-sector fusion initiative — through a joint venture with Mitsubishi Heavy Industries, and plans to raise additional funding for the engineering design phase within 2026.
Finally, in December 2025, TAE Technologies announced an all-stock merger with Trump Media & Technology Group valued at approximately $6 billion, charting an unusual path for the fusion industry by accessing public market capital through a SPAC-like structure (terms and the final corporate form remain undetermined ahead of an expected close in 2026).
Quantum Computing — "The Year of the Public Market" and the Bet on Massive Hardware
Quantum sector full-year 2025 procurement reached $4.53 billion (approximately ¥679.5 billion, 32 deals), up 374% year-over-year, with the first nine months of 2025 alone totaling $3.77 billion (approximately ¥565.5 billion)—nearly three times the full-year 2024 total. MarketsandMarkets estimates the 2026 market size at $5.59 billion (approximately ¥838.5 billion), projected to expand to $25.63 billion (approximately ¥3.8445 trillion) by 2032 at a CAGR of 28.66%. McKinsey estimates up to $97 billion (approximately ¥14.55 trillion) for the entire "quantum computing technology stack" by 2035. For enterprise implementation, industry-shared figures show finance (risk modeling, portfolio optimization, fraud detection) at 46%, pharmaceuticals (molecular simulation) at 38%, and manufacturing (optimization) at 40%.
On the hardware front, 2026 is the "year of the public market." Quantinuum raised $600 million (approximately ¥90 billion) in September 2025 in a Honeywell-led round at a pre-money valuation of $1 billion, then confirmed plans to confidentially file an S-1 with the SEC dated May 8, 2026, and list on Nasdaq under the ticker "QNT." As noted by Reuters Breakingviews, HPCwire, and The Quantum Insider, the expected valuation is over $20 billion (approximately ¥3 trillion), with a pure IPO—not a SPAC—with pricing eyed for sometime in June. The company's flagship processor "Helios" commercially launched in November 2025, achieving 98 physical qubits and up to 94 logical qubits (error-detection basis; 48 logical on an error-correction basis) using trapped-ion technology. The company has established a "beyond break-even" state (error rate reduced by 1–2 orders of magnitude from physical to logical) and published an accelerated roadmap targeting universal, fully fault-tolerant operation by 2030.
PsiQuantum secured $1 billion (approximately ¥150 billion) in a Series E round in September 2025 led by BlackRock, Baillie Gifford, Temasek, and Nvidia's NVentures, with a post-money valuation of $7 billion (approximately ¥1.05 trillion). Its photonic quantum approach—an ambitious plan to realize a fault-tolerant device with approximately one million physical qubits by 2027—is moving forward, with formal groundbreaking scheduled for June 2026 at two major sites: Moreton Bay Central in Brisbane, Australia, and Chicago, Illinois. The combined site area is 500,000 square feet (approximately 46,000 m²), with Jacobs Solutions serving as EPC (engineering, procurement, and construction) contractor. The Australian federal government and Queensland state government jointly announced support totaling approximately AUD 940 million (approximately ¥94 billion). However, operations have been pushed back from the original end-of-2027 target to 2029, with Startup Daily and InnovationAus noting "schedule delays on what may be the world's largest cryptographic-asset-scale venture."
IonQ announced on February 25, 2026, that it had surpassed $100 million in single-year GAAP revenue for the first time in the quantum industry (full-year 2025: $130 million, approximately ¥19.5 billion, +202% year-over-year). Following its acquisition of Oxford Ionics in October 2025 ($1.075 billion, approximately ¥161.3 billion), the company completed acquisitions of optical communications startup Skyloom Global and Seed Innovations on January 28, 2026. IonQ is aggressively pursuing "full-stack" integration across quantum computing, quantum networking, quantum sensing, and quantum security through M&A, intensifying its competitive axis against Quantinuum in the trapped-ion space.
On the big tech side, Microsoft published a roadmap on its Azure Quantum blog to demonstrate a 10–20 logical qubit-class device within 2026, based on Majorana 1 (topological quantum), unveiled in February 2025. IBM has advanced to next-generation Quantum System Two following mass deployment of Condor (433 qubits), and Google demonstrated quantum advantage in portfolio optimization with Willow (up to 1,024 qubit-class). Atom Computing disclosed a 1,225-qubit neutral-atom system in 2025 with a stated goal of 5,000 qubits by 2027. At GTC in March 2026, Nvidia made CUDA-Q + NVQLink + cudaq-realtime generally available, mainstreaming a hybrid configuration that connects GB200 NVL72 with quantum hardware from Pasqal, Qblox, Classiq, IonQ, Quantinuum, and others in real time via the "DGX Quantum" architecture. At GTC 2026, Classiq showcased a real-world example of reducing the synthesis and execution time of a 31-qubit circuit on a single A100 GPU from 67 minutes to 2.5 minutes, signaling the shift of quantum-classical hybrid from "one-off computational demos" to "real production pipelines."
Synthetic Biology — The Full-Scale Commercialization of "Programmable Biology"
According to SynBioBeta's 2024 Investment Report, global VC investment in synthetic biology recovered to $12.2 billion in 2024, with over 12,000 investment deals and more than 900 companies analyzed. The fastest-growing areas are the intersections of "AI × protein design," "AI × RNA," and "AI × biofoundry."
The biggest news of 2026 is undoubtedly Isomorphic Labs. Spun out of Google DeepMind in 2021, the company raised $2.1 billion in a Series B on May 12, 2026, with Bloomberg, Reuters, and Endpoints News reporting simultaneously. The round was led by Thrive Capital, with participation from Alphabet, Abu Dhabi's MGX, Singapore's Temasek, and the UK Sovereign AI Fund — a landmark moment signaling the large-scale flow of sovereign wealth into AI biotech. Building on AlphaFold's structural prediction capabilities, the company has established drug discovery partnerships worth up to $3 billion with Eli Lilly and Novartis, and in January 2026 launched a cross-modality research program with Johnson & Johnson. CEO Demis Hassabis has repeatedly stated that "AI-designed drugs will enter clinical trials by the end of 2026."
Profluent, based in Emeryville in Silicon Valley, completed a $106 million Series B in November 2025, led by Altimeter Capital and Bezos Expeditions, bringing cumulative funding to $150 million. Founder Ali Madani was a pioneer who demonstrated in 2023 in Nature Biotechnology that LLMs can generate functional proteins. The company has built the world's first AI-designed CRISPR and the Protein Atlas — one of the industry's largest, housing 11.5 billion proteins — and has shown that scaling laws apply to protein design. Profluent is pursuing parallel strategies: commercializing its "Protein GPT" platform while releasing open models through "OpenCRISPR" and "OpenAntibodies."
Cradle Bio, based in Delft, Netherlands, has secured a cumulative $103 million (including $73 million in its Series B) from Index Ventures, IVP, and Kindred Capital. It has established direct contracts with Novo Nordisk, Johnson & Johnson, and Ginkgo Bioworks to advance wet lab automation. EvolutionaryScale, spun out of Meta AI Research's former protein team, secured $142 million in seed funding from Nat Friedman, Daniel Gross, Lux Capital, Amazon, and NVentures, anchored by ESM3 — a foundation model trained on 2.8 billion proteins with 98 billion parameters. Its de novo creation of GFP (green fluorescent protein), described as "500 million years of evolution compressed," was published in Science in January 2025 and dramatically raised public awareness of AI-native biology.
Jakob Uszkoreit, co-author of "Attention Is All You Need" during his time at Google Brain, co-founded Inceptive in 2021 with Stanford's Rhiju Das. The company has raised a total of $120 million across two rounds from a16z, Obvious Ventures, Section 32, and Nvidia, pursuing a strategy to deliver generative AI-designed mRNA medicines as "Biological Software." In 2025, the company announced that AI-designed RNA molecules showed expected activity in preclinical mouse models, making it one of the leading candidates in a potential second wave of mRNA therapies in the post-COVID era.
On the legacy side, Ginkgo Bioworks is struggling. Q1 2026 revenue (reported May 7) came in at $19 million, down 49% year-over-year, with an Adjusted EBITDA loss of $42 million and cash and equivalents of $373 million. The company sold its biosecurity business on April 3, 2026, and is concentrating resources on scaling its automated lab Nebula, while strengthening its lab-as-a-service channels through partnerships with OpenAI, AWS, Benchling, and Tamarind Bio. The company is in the midst of structural reform, including a $47 million contract with Pacific Northwest National Laboratory and a renegotiated reduction of $140 million in minimum contract value with Google Cloud.
Generate Biomedicines raised $400 million in an IPO on February 26, 2026 — the largest biotech IPO of the year. In May, Nvidia announced a strategic alliance with a $10.4 million investment. The company is advancing its AI-designed asthma antibody GB-0895 into Phase 3 trials, positioning it as a challenger to Tezspire (Amgen/AstraZeneca). Recursion completed a whole-genome CRISPR knockout neuron map for Roche/Genentech's neurodegeneration disease map, earning $30 million in milestone payments. Its $1.5 billion partnership with Bayer and simultaneous milestone achievements across five programs with Sanofi have been summarized as marking "techbio's transition from single-hit targeting to a platform economy of multiple simultaneous development tracks" (pharmaphorum).
Advanced Chemistry & New Materials — A Two-Front Strategy of Self-Driving Labs and Green Materials
The chemistry and materials sector is evolving along two axes: "green materials" and "AI × self-driving labs." In May 2026, Boston Metal secured a cumulative total of over $500 million (approx. ¥75 billion) through a $75 million (approx. ¥11.3 billion) round welcoming India's Tata Steel as a new strategic investor. The company's Molten Oxide Electrolysis (MOE) is a "carbon-free steelmaking" process that electrochemically converts iron ore directly into high-purity steel, and is advancing production-scale demonstrations of niobium, tantalum, and tin at its Brazilian subsidiary, Boston Metal do Brazil. Materially (formerly Materially USA) plans to expand into vanadium, nickel, and chromium by 2027.
Meanwhile, Sublime Systems (MIT-founded, low-carbon cement) saw an $87 million (approx. ¥13.1 billion) DOE grant it was set to receive canceled by the Trump administration in late 2025, leading to a two-thirds staff reduction in March 2026 and a temporary halt to construction of its Holyoke, Massachusetts facility. Solugen secured a conditional commitment for a $213.6 million (approx. ¥32 billion) DOE loan guarantee in October 2024, with production at Bioforge Marshall (southwest Minnesota) scheduled to begin in autumn 2025. The company uses ADM dextrose to produce low-carbon organic acids, avoiding 18,000 tonnes of CO2 per year. Aether uses CO2 captured from the atmosphere (in partnership with Switzerland's Climeworks) as feedstock to produce CVD lab-grown diamonds, and has raised a cumulative $18 million (approx. ¥2.7 billion) with a unique business model that guarantees 20 tonnes of CO2 removal per carat.
On the AI × self-driving lab front, Orbital Materials (founded by former DeepMind researcher Jonathan Godwin) raised a $16 million (approx. ¥2.4 billion) Series A from Radical Ventures and Toyota Ventures. Using a general-purpose foundation model called "Linus," the company conducts end-to-end exploration of battery materials, CO2 adsorbents, semiconductor materials, and organic molecules. A "closed-loop experiment" is operating at its own lab in New Jersey, cycling through benchmarks (synthesis, property measurement, and model training) in under 24 hours. MatterX integrates deterministic robotic labs with machine learning and has commercialized an autonomous experimentation platform for polymer design. Citrine Informatics, as a commercial SaaS for Generative AI for Materials, counts Shell, BASF, Toyota, and Panasonic among its customers and has demonstrated cases of reducing new product development lead times by 5–10x. Foundation models from Microsoft Research, DeepMind GNoME, Meta, and Google DeepMind (MatterGen, GNoME, Open Catalyst) are increasingly open-sourced, shifting the global materials discovery landscape from a battle of "physical exploration costs" to one of "simulation and robotics budgets."
Also notable is how four fields — HTS (high-temperature superconductors) for fusion reactor magnets, cryo-electronics for quantum computers, bioreactors for synthetic biology, and others — are rapidly deepening their "complementary goods relationships" at the materials and components layer. CFS's expanded manufacturing capacity for large HTS magnets, Nvidia's quantum-GPU interconnects, and the growing molecular datasets flowing through Ginkgo's Datapoints platform are all forming a structure that drives down technology costs across other fields.
Silicon Valley VC Perspective — The Intersection of "Physical AI" and National Security
As of May 2026, the VC landscape has clearly bifurcated into two poles. One pole is concentrated investment in mega generative AI rounds; the other is massive bets on deep tech in domains where AI alone cannot win. The leading example of the latter is Lux Capital, which closed its $150M Fund IX on January 7, 2026, bringing the firm's total AUM to $7 billion. Founder Josh Wolfe announced the launch of an internal R&D unit he calls "Lux Labs" in late 2025, targeting the cross-domain intersection of AI × physical systems × defense, with Eikon Therapeutics, Variant Bio, EvolutionaryScale, Hadrian, Anduril, Physical Intelligence, and Impulse Space as ignition points.
Eclipse Ventures closed a $1.3 billion fund on April 7, 2026, explicitly themed around "Physical AI." Structured with a $720M early-stage tranche plus a $591M growth reserve, the firm re-armed around manufacturing and mobility players including Wayve, Bedrock Robotics, Mind Robotics, Redwood Materials, and Arc Boats. Founding General Partner Lior Susan told TechCrunch: "Hardware has a ceiling — and that ceiling is precisely what creates barriers to entry."
DCVC closed Fund V at $725M and DCVC Bio III at $400M during 2025, reaching approximately $4B in AUM. Founders Fund closed a new $6B growth fund during 2025 and pursued an extreme concentration strategy — cutting a $1B solo check into Anduril and a $1.25B check into Anthropic, effectively deploying nearly the entire fund across just seven companies. Khosla Ventures raised a combined $3.5B. In early 2026, a16z continued building out its American Dynamism division alongside vehicles like 1789 Capital, while adding Erin Price-Wright as a dedicated energy partner, deepening its strategy of connecting deep tech to a "national agenda."
Vinod Khosla (Khosla Ventures) has maintained his conviction in fusion energy from CFS's Series A in 2019 through his most recent follow-on investments. His line — "entrepreneurs who aren't tackling what seems impossible are not candidates for my fund" — continues to be quoted repeatedly in media appearances since late 2025. Meanwhile, Bill Gates's Breakthrough Energy Ventures (AUM ~$4B) announced in February 2026 that it would halt new investments in its Catalyst Fund (a project-style fund focused on commercial deployment) and scale back, as reported by Bloomberg and GeekWire — a candid admission that the climate policy retreat under the Trump administration forced its hand. Core portfolio companies including CFS, Type One Energy, and KoBold Metals continue nonetheless. General Catalyst led Pacific Fusion's $900M Series A and has maintained ongoing investment in Helion as well.
Compressing the VC community's shared "2026 thesis" to its essentials yields three points: first, "Sovereign AI and Sovereign Energy have been unified into the same thesis cluster"; second, "Hardware-Software Convergence introduces verifiable financial metrics — hardware gross margins, manufacturing scale curves — making it a hedge against the valuation expansion seen in pure-AI rounds"; and third, "12–15 year fund lifespans with staged, milestone-based capital calls have become the new standard, premised on an ecosystem in which governments, regulators, large corporations, and universities serve as co-investors."
Media Coverage and Analyst Readings — The Temperature Gap Between "The Boiling Point Is Still Ahead" and "The IPO Window Is a Few Months Away"
The tone of coverage varies clearly by media outlet. Bloomberg and the Financial Times treat fusion, quantum, and synthetic biology as the "top post-generative AI candidates," while Breakingviews journalist Anita Ramaswamy noted that the Quantinuum IPO would be "the largest quantum event to hit the public markets within months." Reuters, by contrast, takes a more neutral stance, noting that "the eve of IPO pricing has arrived, but markets remain wary of a quantum bubble."
TechCrunch's April 10, 2026 piece "Every fusion startup that has raised over $100M" catalogued fusion companies, listing more than ten including CFS, Helion, Pacific Fusion, TAE, Marvel Fusion, Xcimer, Type One, Tokamak Energy, Realta Fusion, and Zap Energy. In a March feature, the publication forecast that "physical validation of fusion will be concentrated in 2026–2027, and once that threshold is crossed, the cost of capital will fall dramatically." The Wall Street Journal's May 2026 special "The Year of Real Quantum Revenue" positioned IonQ's $130 million per year as "a mainframe-like break-even point for quantum computing," while pointedly noting that "hardware costs and energy consumption remain unresolved."
McKinsey's 2026 update to "The Rise of Quantum Computing" places the 2035 market forecast for the entire quantum technology stack at up to $97 billion, with quantum molecular simulation in pharmaceuticals projected to become "commercially viable in 5–10 years, though near-term use will be limited to resolving individual bottlenecks in drug discovery pipelines." BCG's revised 2026 edition of "From Tech to Deep Tech"—re-cited by MIT Technology Review—projects that European deep tech has the potential to generate $1 trillion in economic value and 1 million jobs by 2030, and is cited as evidence for the European EIC committing €1.4 billion in its 2026 program.
The Nikkei Shimbun's September 2025 feature independently estimated Kyoto Fusioneering's enterprise value at ¥72.1 billion, and Nikkei Asia, in April 2026, coined the term "Pacific Rim Deep Tech Window" to describe the simultaneous public market access of Quantinuum's IPO and Kyoto Fusioneering, pointing to the correlated capital inflows into fusion and quantum sectors in the US and Japan. MIT Technology Review's "10 Breakthrough Technologies 2026" excluded both fusion and quantum computing from direct selection, deeming it "unlikely that commercial breakthroughs will be achieved within 2026," and instead chose three items from the biotech space: "base-edited babies," "de-extinction," and "embryo scores." In an editorial note by the MIT Technology Review editor-in-chief, an analyst consensus was shared: "SPARC fusion next year, Quantinuum's IPO this year, Isomorphic Labs' first clinical readout between December 2026 and Q1 2027."
That said, not all coverage is optimistic. The Financial Times Lex column and Bloomberg Opinion columnists have repeatedly sounded alarms that "commercial PPAs for fusion will generate no meaningful cash flows until the late 2030s" and that "quantum IPO valuations carry an abnormally high market cap per dollar of GAAP revenue." As the case of Sublime Systems illustrates, the withdrawal of government subsidies, US-China export controls, and the competition for talent all remain deep risks capable of changing the fate of individual companies overnight.
Key Milestones and Uncharted Territory for the Next 12–18 Months
The period from the second half of 2026 through Q1 2027 represents a singular timeframe in which multiple deep tech sectors simultaneously reach "physical and financial validation points." In fusion, CFS's SPARC is scheduled to begin operations within 2026, targeting a net energy demonstration in Q1 2027. Helion has announced completion of a power generation demonstration using Polaris within 2026, on the path to delivering 50 MW to Microsoft beginning in 2028. Pacific Fusion plans to demonstrate production-scale pulsar modules jointly with General Atomics within 2026; Type One Energy plans to break ground on Infinity One and close a $250M Series B within the same year. Marvel Fusion is on track to complete a demonstration facility at Colorado State University by 2027. Japan's FAST plans to complete engineering design fundraising within 2026.
In quantum, several milestones converge in 2026–2027: Quantinuum's IPO is priced for June 2026; PsiQuantum's Brisbane site breaks ground in June 2026 with operations delayed to 2029; Microsoft targets a demonstration of 10–20 topological logical qubits within 2026; IBM Quantum System Two roadmap updates; and Atom Computing's path toward ~3,000 qubits. As Nvidia's NVQLink real-time API increases the frequency of quarterly hybrid benchmark disclosures (from Schrödinger, Vertex, JPMorgan, and others), there is growing expectation of convergence toward verifiable, demonstrable metrics that prevent the term "quantum advantage" from becoming hollow.
In synthetic biology, Isomorphic Labs' AI-designed drug Phase I clinical trials are expected to begin in Q4 2026–Q1 2027, making the world's first human administration of an AI-generated pharmaceutical a realistic near-term event. Milestone achievements from Recursion–Sanofi and Recursion–Bayer, the first interim analysis of Generate Biomedicines' GB-0895 Phase 3 trial, and Inceptive's first clinical candidate nomination may all occur in rapid succession. Profluent's OpenCRISPR and Cradle Bio's expanding new partnerships are also noteworthy developments.
In chemistry and materials, Boston Metal's Brazilian plant moving to full operation, Solugen's Bioforge Marshall launch, Orbital Materials' Linus public model updates, and the commercial rollout of MatterX/Citrine's Generative Materials products are all unfolding in sequence. Meanwhile, multiple industry publications have flagged the possibility that rescue financing or acquisition negotiations for Sublime Systems could materialize by end of 2026.
On the VC side, the second half of 2026 is expected to see closings of a second wave of deep tech–focused funds (DCVC VI, Eclipse VII, Lux X, a16z American Dynamism II) and the completion of a new Khosla fund. Sovereign AI Funds (UK, UAE, Singapore, Saudi Arabia, South Korea) are expected to make additional commitments, and a deepening of overseas deep tech investment by Japanese LPs through the Japanese Fusion Energy Council (J-Fusion) — primed in part by the Japanese consortium's participation in CFS — is anticipated to gain steady momentum. With generative AI valuations at historically elevated levels, the accelerating flow of capital into companies that command the physical world can be seen as the inevitable outcome of Silicon Valley VCs' diversification strategies.