AI Infrastructure Startup Nexlyr Raises $340M Series C
A Landmark Round in a Crowded Market
San Francisco-based AI infrastructure startup Nexlyr closed a $340 million Series C funding round on Tuesday, pushing its valuation to $2.8 billion and cementing its position as one of the most closely watched companies in enterprise AI tooling. The round was led by Andreessen Horowitz's Growth Fund, with participation from Coatue Management, Spark Capital, and strategic investor Salesforce Ventures. For a market that saw late-stage valuations compress sharply through 2024 and early 2025, the deal signals a genuine thaw in investor appetite for infrastructure plays with demonstrable revenue traction.
Nexlyr, founded in 2022 by former Google DeepMind researchers Priya Anand and Marcus Teller, builds orchestration and observability tooling for enterprises deploying large language models at scale. The company reported $67 million in annualized recurring revenue heading into the raise, representing 3.1x year-over-year growth. Those numbers, sources close to the deal say, gave investors confidence the valuation multiple — roughly 42x ARR — was defensible in the current environment.
Why Infrastructure Is Attracting Capital Again
The investment climate for AI has bifurcated sharply in 2026. Consumer-facing generative AI applications have struggled to convert hype into durable retention metrics, while the infrastructure layer — the picks-and-shovels companies helping enterprises actually run and monitor AI systems — has emerged as a far more predictable revenue story. "Every large organization that committed to an AI strategy in 2023 and 2024 is now drowning in model sprawl, compliance requirements, and escalating inference costs," said Meredith Cho, a partner at Coatue who will join Nexlyr's board. "Nexlyr sits directly in the path of that problem."
The company's flagship product, NexOps, integrates with major cloud providers including AWS, Azure, and Google Cloud to give engineering teams granular visibility into model performance, cost attribution, and regulatory audit trails. It competes with players like Weights & Biases, Arize AI, and enterprise offerings from Datadog, but has differentiated itself through native support for multi-model pipelines — a critical capability as organizations increasingly chain together specialized models rather than relying on single general-purpose systems.
The Valuation Math and What It Reflects
At $2.8 billion, Nexlyr's valuation is a deliberate recalibration from the speculative peaks of 2021. Investors and analysts note that the company had reportedly fielded term sheets at valuations as high as $3.4 billion but chose the more conservative figure to avoid the down-round risk that has plagued several high-profile AI startups over the past 18 months. "There's real discipline here," said Jordan Kwame, a tech analyst at Redburn Atlantic. "They left money on the table to protect their cap table health. That's a management team thinking two rounds ahead."
The round also included a secondary component of approximately $45 million, allowing early employees and seed investors to achieve partial liquidity — a structural feature that has become increasingly standard in large private rounds as the IPO window remains narrower than founders would prefer. Nexlyr has not committed to a public offering timeline, though co-CEO Priya Anand told Verodate the company is "building toward the optionality" of a listing in 2027 or 2028.
Deployment Plans and Hiring Agenda
Nexlyr intends to deploy the new capital across three priority areas: expanding its European go-to-market presence, deepening integrations with sovereign AI deployments in regulated industries like finance and healthcare, and accelerating R&D on real-time cost optimization features that the company previewed at its annual user conference in March. The European push is particularly timely, given that the EU AI Act's compliance provisions are now in active enforcement, creating urgency among enterprises to instrument their AI stacks before audit obligations escalate.
On the hiring front, the company plans to grow its 310-person headcount by roughly 40 percent over the next 12 months, with the heaviest concentration in sales engineering and enterprise customer success — roles that signal Nexlyr is chasing larger, stickier contracts rather than high-velocity SMB growth. Current customers include three of the top ten US banks, two major European insurers, and a cluster of healthcare systems navigating HIPAA-compliant AI deployment.
What Comes Next for Enterprise AI Tooling
Nexlyr's raise arrives as the broader MLOps and AI observability market is projected to reach $8.2 billion by 2028, according to a February 2026 report from IDC. Consolidation in the space is accelerating, with Cisco's acquisition of Robust Intelligence last year and IBM's expanded investment in its Watson orchestration stack squeezing the middle tier of the market. For Nexlyr, the strategic imperative is clear: scale fast enough to become a default platform before hyperscalers and legacy vendors close the gap. With $340 million in fresh capital and a revenue engine already firing, the runway to find out looks longer than most of its competitors can claim.
Supply Chain Attack Prevention: What Works in 2026
The Threat That Keeps Escalating
Three years after the SolarWinds and XZ Utils incidents reshaped how the security industry thinks about software dependencies, supply chain attacks have only grown more sophisticated. In the first quarter of 2026 alone, Sonatype's State of the Software Supply Chain report logged a 178% year-over-year increase in malicious packages uploaded to public repositories like npm, PyPI, and Maven Central. The message is unambiguous: organizations that treat third-party software as inherently trustworthy are operating on borrowed time.
The attack vector is deceptively simple. Adversaries compromise a vendor, an open-source maintainer, or a build pipeline — then use that foothold to distribute malware to every downstream customer simultaneously. One poisoned update can detonate across thousands of organizations before a single alert fires. "The economics favor the attacker completely," says Tanya Janca, founder of We Hack Purple and a frequent advisor to Fortune 500 security teams. "One successful upstream compromise multiplies indefinitely. Defenders have to be right every time; attackers only need to be right once."
Software Bills of Materials Are Now Table Stakes
The U.S. Cybersecurity and Infrastructure Security Agency's 2025 directive mandating SBOMs — software bills of materials — for any vendor selling to federal agencies has accelerated adoption across the private sector. An SBOM is essentially an ingredient list for software: every library, framework, and dependency catalogued with version numbers and known vulnerability data. As of early 2026, roughly 61% of enterprise organizations now require SBOMs from critical suppliers, up from 23% in 2023, according to Gartner.
But generating an SBOM is only half the battle. The real value comes from continuous monitoring. Platforms like Chainguard, Anchore, and Snyk have built automated pipelines that ingest SBOM data and cross-reference it against vulnerability databases in near real time. When a critical CVE drops — as happened in February 2026 with a severe flaw in a widely used JSON parsing library — teams using active SBOM monitoring were able to identify affected systems and begin patching within hours rather than days.
Signed Artifacts and the SLSA Framework
Code signing has existed for decades, but the Supply Chain Levels for Software Artifacts framework — SLSA, pronounced "salsa" — brings structured rigor to the process. Developed collaboratively by Google and the Open Source Security Foundation, SLSA defines four trust levels based on how verifiably a software artifact was built. At Level 3 and above, builds must occur in isolated, tamper-resistant environments, and every step must produce cryptographically signed provenance records.
Kubernetes hit SLSA Level 3 compliance in late 2025, a milestone that sent a clear signal to the broader ecosystem. Major cloud providers are now incentivizing customers to prefer SLSA-compliant dependencies. "Provenance is the new password," says Dan Lorenc, CEO of Chainguard. "If you can't prove where your software came from and how it was built, you're essentially running code on faith." GitHub's Artifact Attestations feature, which became generally available in mid-2025, has made generating and verifying provenance accessible even for small development teams without dedicated security staff.
Zero Trust Principles Applied to the Pipeline
Network-level zero trust — verifying every user and device before granting access — is well understood. Applying the same philosophy to CI/CD pipelines is newer territory, and arguably more urgent. Attackers who compromise a build system can inject malicious code during compilation, after all the human review has already happened. Stopping them requires treating every stage of the pipeline as potentially hostile.
Practical implementation involves short-lived, scoped credentials for pipeline jobs rather than long-lived API tokens; mandatory code review gates that cannot be bypassed programmatically; and runtime integrity checks that verify deployed artifacts match their signed hashes. CrowdStrike's 2026 Global Threat Report highlighted that 34% of software supply chain intrusions last year exploited overprivileged CI/CD service accounts — a finding that has driven renewed urgency around secrets management tools like HashiCorp Vault and AWS Secrets Manager.
The Human Layer Remains Critical
Technology alone cannot close the gap. The XZ Utils attack succeeded partly because a determined social engineer spent almost two years cultivating trust with an overworked open-source maintainer before inserting a backdoor. OpenSSF's Secure Open Source Rewards program, which pays maintainers for completing security audits, now funds over 1,200 projects — an acknowledgment that burned-out volunteers cannot be the last line of defense for infrastructure used by billions of devices.
Enterprise security teams are also embedding "supplier security scorecards" into vendor procurement processes, evaluating potential partners on patch cadence, penetration test results, and SBOM maturity before signing contracts. The paradigm shift is clear: software supply chain security is no longer a niche concern for DevSecOps specialists. It has become a boardroom-level risk management issue, and the organizations treating it as such are measurably harder to compromise.
Nuclear Fusion Hits 100-Second Milestone in 2026 Breakthrough
The Plasma That Changed Everything
South Korea's KSTAR tokamak reactor held a superheated plasma at 100 million degrees Celsius for a sustained 102 seconds in February 2026, shattering its own previous record of 48 seconds and sending shockwaves through the global energy research community. The achievement, confirmed by the Korea Institute of Fusion Energy and independently verified by the International Atomic Energy Agency, represents the most significant milestone in controlled nuclear fusion since the National Ignition Facility's historic ignition moment in December 2022. Scientists are now speaking openly about something they rarely allowed themselves to voice before: a realistic pathway to grid-scale fusion power before 2045.
"This isn't incremental progress — this is a phase transition in what we believe is physically achievable," said Dr. Siyeon Lim, a plasma physicist at Seoul National University who consulted on the KSTAR experiments. "Sustaining plasma at these temperatures for over 100 seconds means we've moved from proving a concept to engineering a technology."
Private Capital Is Accelerating the Timeline
The KSTAR result arrives against a backdrop of unprecedented private investment flooding the fusion sector. Commonwealth Fusion Systems, backed by $1.8 billion in funding including a major 2025 Series B round led by Google and Breakthrough Energy Ventures, is on track to switch on its SPARC demonstration reactor in Cambridge, Massachusetts, by late 2027. The company's high-temperature superconducting magnets — operating at 20 tesla, roughly 40 times stronger than a hospital MRI machine — have already passed stress testing that the team described as "beyond spec." Meanwhile, TAE Technologies reported in January 2026 that its Norman plasma device achieved hydrogen-boron fusion reactions at commercially relevant energy ratios for the first time, a pathway that would produce virtually no radioactive waste.
Helion Energy, backed by a $2.2 billion commitment from OpenAI's Sam Altman and a historic power purchase agreement with Microsoft, is targeting first electricity generation in 2028. The company's seventh-generation device, Polaris, completed its first full operational cycle in Everett, Washington, in March 2026, with internal metrics the company says are "ahead of our internal 18-month forecast." If even one of these parallel private tracks delivers, the economics of global power generation shift fundamentally.
The Engineering Gaps That Still Exist
Optimism, however, must be measured against the formidable engineering challenges that remain unsolved at scale. Tritium supply is among the most pressing: the isotope of hydrogen used as fusion fuel is extraordinarily scarce, with global reserves estimated at roughly 25 kilograms — enough to power a small demonstration reactor but nowhere near sufficient for fleet-scale deployment. Breeding tritium inside a lithium blanket surrounding the reactor, the proposed solution, has never been demonstrated at commercial scale. The ITER project in southern France, the international megaproject now targeting first plasma in 2027 after years of delays and a budget exceeding $22 billion, is designed partly to validate this blanket concept.
Materials science presents an equally stubborn barrier. The neutron bombardment inside a fusion reactor degrades structural components far faster than in conventional fission plants, and no material has yet been demonstrated to withstand a full operational lifetime at commercial duty cycles. The EU's EUROfusion consortium announced a dedicated €340 million materials research program in January 2026 specifically targeting this bottleneck, with results expected to inform DEMO — Europe's planned demonstration power plant — by 2032.
Geopolitics and the Race for Fusion Leadership
The stakes extend well beyond energy markets. China's Experimental Advanced Superconducting Tokamak, known as EAST, has been quietly logging operational hours that rival KSTAR, and Beijing's 14th Five-Year Plan allocated $1.5 billion to domestic fusion research programs. Chinese officials have stated publicly that achieving fusion energy leadership is a national strategic priority, framing it explicitly in terms of energy independence and geopolitical influence. The United States Congress responded in late 2025 by passing the Fusion Energy Act, creating a new Office of Fusion Energy Sciences with a $400 million annual budget and a mandate to reach a domestic demonstration plant by 2035.
"We are in the Sputnik moment of fusion," said Dr. Carolyn Cochran, former deputy director of the DOE's Office of Science and now a senior fellow at the Brookings Institution. "The question is no longer whether fusion works. The question is who builds the first commercial plant and what that means for the next century of power."
What a Fusion Economy Actually Looks Like
Analysts at BloombergNEF modeled three fusion deployment scenarios in a February 2026 report and found that even the most conservative — commercial plants operating at scale by 2050 — would reduce projected global carbon emissions by 8 gigatons annually by 2060, more than current total U.S. emissions. The aggressive scenario, with grid-connected fusion by 2038, would fundamentally reshape energy pricing, potentially driving industrial electricity costs below $20 per megawatt-hour in developed markets. Uranium markets have already begun pricing in fusion risk, with spot prices softening 12 percent since mid-2025 despite tight near-term supply. The 102-second plasma in Seoul didn't just break a record. It started a clock.
How Enterprises Are Finally Making Blockchain Work in 2026
The Quiet Revolution in Corporate Ledgers
For years, blockchain technology occupied an uncomfortable middle ground in enterprise strategy — too promising to ignore, too complex to implement at scale. That calculus is shifting decisively. According to a February 2026 report from Gartner, 41% of Fortune 500 companies now run at least one production-grade blockchain application, up from just 18% in 2023. The driver isn't speculative crypto markets this time. It's operational efficiency, supply chain transparency, and a maturing infrastructure stack that has quietly made distributed ledgers genuinely deployable.
The shift mirrors what happened with cloud computing around 2012 — a period when early skepticism gave way to boardroom mandates. Enterprise blockchain is entering that same inflection point, and the companies moving fastest are those treating it as infrastructure rather than innovation theater.
Stablecoins Are Now Treasury Tools, Not Novelties
Corporate treasury departments have emerged as an unexpected battleground for blockchain adoption. Following the U.S. Digital Asset Clarity Act signed in late 2025, which established clear regulatory frameworks for stablecoin issuance and corporate crypto holdings, companies including Siemens, Maersk, and several mid-cap pharmaceutical firms have begun settling cross-border invoices in USDC and JPMorgan's proprietary JPM Coin. The operational appeal is straightforward: transactions that previously took three to five business days through correspondent banking networks now clear in under 90 seconds at a fraction of the cost.
"We're not making a philosophical bet on decentralization," said Clara Hoffmann, CFO of a Munich-based logistics firm that processed €340 million in stablecoin payments last quarter. "We're cutting wire transfer fees and FX conversion costs by roughly 60%. That's a spreadsheet decision, not a manifesto." Citigroup's 2026 Digital Money report estimated that enterprise stablecoin transaction volume reached $2.4 trillion globally in 2025, a figure expected to double before year-end.
Supply Chain Transparency Gets a Ledger
Retail giants and pharmaceutical companies are deploying blockchain not for payments but for provenance tracking — and regulators are accelerating the timeline. The EU's expanded Digital Product Passport regulation, which took full effect in January 2026, requires manufacturers in electronics, textiles, and batteries to maintain immutable records of component origins and recycling data. Blockchain infrastructure providers including Morpheus Network and the IBM-spinout TrustChain have seen enterprise contract signings surge 78% year-over-year as a direct result.
Walmart Canada's blockchain-based freight reconciliation system, now in its third year of operation, reduced invoice disputes with carriers by 97% and saved an estimated $12 million annually in administrative overhead. The model is being replicated across retail. Target confirmed in March 2026 that it would extend similar infrastructure to its top 200 private-label suppliers by Q3, covering food safety traceability from farm to shelf.
Layer-2 Networks Make the Cost Math Work
One persistent barrier to enterprise blockchain adoption — prohibitive transaction costs and throughput limitations on base-layer networks — has been substantially addressed by the maturation of Layer-2 scaling solutions. Ethereum's rollup ecosystem, including Arbitrum and Optimism, now processes a combined average of 180 transactions per second at costs below $0.002 per transaction, making high-volume business applications economically viable for the first time. Polygon's enterprise division reported signing 34 new Fortune 1000 clients in Q1 2026 alone.
"The 2021 era conversation was about Ethereum gas fees making enterprise use impossible," said Raj Patel, head of distributed ledger strategy at Deloitte's technology consulting practice. "That objection is largely obsolete. The infrastructure has caught up to the ambition." Solana's enterprise push has also gained traction, particularly among fintech startups building high-frequency settlement applications that require sub-second confirmation times at scale.
The Talent and Integration Gap Remains Real
Despite accelerating adoption, meaningful friction points persist. A March 2026 survey by the Enterprise Ethereum Alliance found that 63% of enterprise blockchain projects cite integration with legacy ERP systems as their primary implementation challenge, ahead of regulatory uncertainty and developer talent shortages. SAP and Oracle have both released dedicated blockchain middleware modules in the past six months, which analysts expect will compress integration timelines significantly for their existing customer bases.
The talent pipeline is tightening in a different direction. Bootcamp enrollment for Solidity and Rust smart contract development rose 120% in 2025 according to Course Report data, suggesting the developer shortage will ease within two to three years. What's less certain is whether established enterprises will move fast enough to capture efficiency gains before nimbler, blockchain-native competitors make legacy models structurally uncompetitive. The window for orderly adoption, most practitioners agree, is open — but not indefinitely.
IBM's 10,000-Qubit Processor Rewrites Quantum Computing Rules
A Threshold Moment for Quantum Science
In a packed auditorium at IBM Research's Yorktown Heights facility last Tuesday, company scientists unveiled what many in the field are already calling the most significant hardware achievement in quantum computing's short but turbulent history. The Condor Pro 2 processor — a 10,247-qubit device operating at temperatures near absolute zero — successfully executed a fault-tolerant algorithm at a scale previously confined to theoretical papers and optimistic grant proposals. For a research community that has spent two decades navigating a relentless cycle of promise and setback, the moment landed with unusual weight.
"We've been chasing this number for years, but raw qubit count was never really the point," said Dr. Priya Mehta, IBM's VP of Quantum Hardware, in a press briefing following the announcement. "What matters is what those qubits can do coherently, under real computational loads, with errors corrected in real time. That's what we demonstrated today."
What the Numbers Actually Mean
The Condor Pro 2 achieved a two-qubit gate error rate of 0.08 percent — a dramatic improvement over the 0.3 percent threshold that physicists have long cited as the minimum requirement for practical quantum advantage. More critically, the system maintained quantum coherence for 1.4 milliseconds across its full qubit array, nearly triple the benchmark set by IBM's previous generation processor in 2024. Google's competing Willow architecture, which made headlines in late 2024 for solving a narrow benchmark problem, operated with a coherence window roughly a fifth of that duration at comparable qubit counts.
Independent verification came swiftly. Researchers at MIT's Center for Quantum Engineering confirmed the results using blind testing protocols — submitting problems to the system without prior knowledge of IBM's configuration — and reported that the processor solved a class of molecular simulation problems relevant to nitrogen fixation chemistry in 47 minutes. Classical supercomputers, including Frontier at Oak Ridge National Laboratory, have no known path to solving the same problem class in under 10,000 years.
The Error Correction Breakthrough Underneath the Headlines
Behind the qubit count, the more technically significant story is IBM's implementation of the surface code error correction protocol at operational scale. Quantum systems are notoriously fragile; environmental noise collapses quantum states before computations complete, which is why most current quantum processors are described as "noisy intermediate-scale" — useful for research but not production workloads. IBM's team encoded 1,024 logical qubits from the physical qubit array using a 10-to-1 overhead ratio, meaning ten physical qubits protect each logical one from decoherence.
"Getting surface code to work at this scale without the overhead eating your advantage is genuinely hard," said Professor Sankar Das Sarma, a condensed matter physicist at the University of Maryland who was not involved in the research. "I've been skeptical of timeline claims for a decade. This particular result makes me revise that skepticism." Das Sarma noted in a post on his institutional profile that the nitrogen fixation result, if reproducible, could have direct implications for developing energy-efficient fertilizer synthesis — an agricultural process currently responsible for approximately 1.4 percent of global CO2 emissions.
Industry and Government React Quickly
Within 24 hours of the announcement, shares in IonQ rose 18 percent as investors recalibrated expectations for the broader quantum sector. Microsoft, which has pursued a rival topological qubit architecture, issued a measured statement acknowledging IBM's milestone while emphasizing that topological approaches would offer "superior long-term stability characteristics." DARPA confirmed it is accelerating review of a $900 million quantum computing initiative originally slated for 2027, with a spokesperson telling Verodate that the IBM result "changes the calculus on near-term deployment timelines."
The pharmaceutical industry is watching closely. Pfizer's computational chemistry division has been among the 180 organizations enrolled in IBM's Quantum Network. A senior scientist there, speaking on background, said the company has already identified three protein-folding simulation tasks that would be immediately queued for the Condor Pro 2 architecture once API access opens in Q3 2026.
What Comes Next — and What Doesn't
Skeptics are quick to temper the enthusiasm. The nitrogen fixation benchmark, while striking, represents a narrow problem class. General-purpose quantum advantage — the ability to outperform classical computers across a broad range of commercially relevant tasks — remains years away by most serious estimates. Qubit connectivity constraints mean the processor still cannot tackle unstructured optimization problems at enterprise scale without significant reformulation overhead.
IBM has committed to publishing the full methodology in a peer-reviewed submission to Nature Physics by September, which will subject the results to the scrutiny they deserve. But the directional signal is difficult to dismiss: quantum computing crossed a threshold last Tuesday that most researchers privately doubted would arrive before 2030. The next milestone, whatever form it takes, just got a great deal closer.
