Amazon Web Services Chief Executive Matt Garman said investor anxiety over artificial intelligence undermining traditional software companies has likely gone too far, even as the sector posts one of its steepest pullbacks in years.

“Look, my own opinion is that much of the fear is overblown,” Garman told CNBC’s Jon Fortt on Thursday, addressing concerns that generative AI platforms could erode the dominance of large software-as-a-service providers.

Technology shares, particularly enterprise software names, have fallen sharply in 2026 following the rapid commercialization of AI tools built on models from companies such as OpenAI and Anthropic. The selloff reflects mounting concern that AI-native applications could commoditize existing SaaS offerings or reduce enterprise spending on legacy systems.

The iShares Expanded Tech-Software Sector ETF is down 24% this year, putting it on track for its worst performance since 2022, when inflation and rising interest rates forced companies to trim technology budgets after a surge in pandemic-driven digital investment.

Market analysts have described the correction as a “SaaS apocalypse,” citing slowing growth rates, compressed valuation multiples, and uncertainty about how AI will reshape software consumption patterns.

Cloud Strength and AI Infrastructure Demand

The turbulence in software contrasts with AWS’s own results. Parent company Amazon reported that fourth-quarter revenue at its cloud infrastructure division rose approximately 24% year over year to $35.6 billion, exceeding analyst expectations. AWS posted a 35% operating margin, slightly higher than the previous quarter, underscoring sustained profitability in its core business.

The divergence highlights a structural distinction in the AI value chain. While application-layer software companies face questions about displacement, hyperscale cloud providers supply the compute, storage, and networking infrastructure required to build and deploy AI systems at scale.

“There’s a huge disruption,” Garman said. “AI is absolutely a disruptive force that’s going to change how software is consumed and how it’s built. And I would argue that the systems of record, as you call them, the SaaS providers and the large players of today have an inside track to winning that business.”

Systems of record — enterprise platforms managing financial data, human resources, compliance, and customer relationships — are deeply embedded within corporate workflows. Replacing them entails operational risk, data migration complexity, and integration challenges, which can slow wholesale displacement.

AWS generates revenue from established vendors, including Adobe, Intuit, and Zillow, while also benefiting from AI model developers expanding compute usage. In November, AWS disclosed a $38 billion spending commitment from OpenAI, reflecting the scale of infrastructure required to train and run large language models.

“Our perspective is that our customers are going to consume more compute technology and more infrastructure than they ever have,” Garman said, arguing that whether companies build AI internally or buy AI-enabled SaaS, overall infrastructure demand should rise.

Slowing SaaS Growth and Broader Spillover

Even as infrastructure spending climbs, growth among major SaaS firms has moderated. ServiceNow, an AWS customer, recently reported fourth-quarter revenue growth of 20.7% year over year, down from nearly 26% two years earlier. While still strong relative to many sectors, the deceleration has weighed on valuations that were priced for sustained hypergrowth.

Investor concern extends beyond enterprise software. Florida-based Algorhythm Holdings said Thursday that an AI-powered product enabled logistics clients to quadruple freight volumes without increasing headcount. Shares of C.H. Robinson Worldwide fell about 23% in midday trading, reflecting fears that AI-driven efficiency gains could pressure revenue models tied to transaction volume or labor-intensive processes.

The underlying debate centers on whether generative AI will cannibalize traditional software categories or expand total addressable markets by unlocking new use cases. Historically, major computing transitions — from on-premises infrastructure to cloud, and from desktop software to web applications — have produced both displacement and expansion. Companies that adapted architecture and pricing models often retained leadership; those that failed to evolve lost relevance.

The current correction represents a repricing of growth expectations and risk premiums for investors. For AWS, the calculus differs. As long as AI development requires large-scale compute, hyperscale cloud providers remain positioned to capture incremental spending, regardless of which application-layer companies ultimately prevail.

Garman’s remarks suggest confidence that AI will alter the shape of enterprise software without necessarily shrinking its economic footprint. The market’s volatility indicates that investors are still determining where value will accrue in an ecosystem being rapidly rewritten by artificial intelligence.

Anthropic’s $30 billion Series G — lifting its valuation to $380 billion — underscores how capital continues to flood into artificial intelligence despite persistent concerns about an emerging bubble.

Anthropic announced Thursday that it has closed a $30 billion Series G financing round, nearly doubling its valuation to $380 billion from $183 billion in its previous Series F.

The deal ranks among the largest private funding rounds in technology history and reinforces the scale of investor conviction in frontier AI companies.

The round was led by Singapore’s sovereign wealth fund GIC and investment firm Coatue Management. Co-leads included D. E. Shaw Ventures, Founders Fund, and Abu Dhabi-backed MGX. Additional investors included Accel, General Catalyst, Jane Street, and the Qatar Investment Authority.

The diversity of backers — spanning sovereign wealth funds, quantitative trading firms, and venture capital investors — signals that AI is now viewed not simply as a venture theme, but as a strategic asset class with geopolitical and macroeconomic implications.

A Capital Arms Race With OpenAI

Anthropic’s funding comes as it competes closely with OpenAI, which has indicated it is seeking as much as $100 billion in additional funding. If secured, that round would reportedly value OpenAI at approximately $830 billion.

The escalating numbers illustrate the extraordinary capital intensity of frontier AI development. Training large-scale models requires cutting-edge semiconductors, vast data center infrastructure, and sustained research investment. Compute costs alone run into billions of dollars annually for top-tier labs.

This funding environment has created what some investors describe as a capital arms race. Firms that control the most advanced models can attract enterprise customers, secure cloud partnerships, and shape standards across industries. Access to capital becomes both a competitive moat and a prerequisite for staying at the technological frontier.

Anthropic’s valuation leap, more than doubling in a single funding cycle, reflects market expectations that a small cluster of companies will capture disproportionate value as AI systems become embedded across software, enterprise operations, and consumer applications.

Anthropic has positioned its Claude model family as enterprise-oriented, emphasizing reliability, security, and structured workflows.

“Whether it is entrepreneurs, startups, or the world’s largest enterprises, the message from our customers is the same: Claude is increasingly becoming more critical to how businesses work,” said Krishna Rao, the company’s chief financial officer. “This fundraising reflects the incredible demand we are seeing from these customers, and we will use this investment to continue building the enterprise-grade products and models they have come to depend on.”

The company has previously indicated that more than 80% of its revenue comes from enterprise customers. That revenue mix contrasts with AI platforms that rely more heavily on consumer subscriptions or advertising models.

Enterprise contracts typically provide larger and more stable revenue streams, but they require extensive compliance capabilities, security assurances, and infrastructure reliability. The fresh capital is likely to be directed toward model development, data center capacity, and enterprise integrations.

Persistent Bubble Concerns — and Why Money Keeps Flowing

The scale of AI funding has fueled comparisons to prior technology manias, including the dot-com era. Valuations measured in hundreds of billions of dollars for private companies, combined with multi-trillion-dollar infrastructure spending projections, have prompted debate about sustainability.

Yet capital continues to flow.

Several structural factors explain the momentum:

First, AI is widely viewed as a general-purpose technology with economy-wide impact potential. Investors are betting not merely on incremental software gains, but on productivity shifts comparable to previous industrial transformations.

Second, sovereign wealth funds and state-backed investors see AI leadership as strategically significant. Participation by entities such as GIC and the Qatar Investment Authority signals geopolitical as well as financial motivations.

Third, hyperscale cloud providers and institutional customers are already embedding AI tools into workflows, creating tangible demand even if long-term monetization curves remain uncertain.

Finally, investors may view concentration risk as acceptable. Rather than funding hundreds of speculative startups, capital is increasingly concentrated in a handful of perceived category leaders. This concentration can drive valuations higher even amid broader caution.

Still, the cost structure of training next-generation models continues to rise. Competition for specialized chips is intense. Regulatory scrutiny is increasing across jurisdictions. And enterprise adoption timelines may not always align with investor projections.

What a $380 Billion Valuation Implies

At $380 billion, Anthropic joins a small group of private technology firms with valuations comparable to established public companies. Such pricing embeds expectations of significant future revenue growth, sustained technological leadership, and durable competitive advantages.

The valuation also reflects the belief that foundation model providers could become core infrastructure providers in the next computing cycle — analogous to operating systems or cloud platforms in earlier eras.

Analysts note that fully realizing expectations will depend on multiple variables, including model performance improvements, regulatory clarity, customer retention, and cost discipline.

However, the fundraising demonstrates that even amid bubble discussions, AI remains one of the most powerful capital magnets in global markets. Investors are wagering that artificial intelligence will redefine productivity, reshape software economics, and create enduring platform companies.

Anthropic’s latest round suggests that, at least for leading players, confidence in that thesis remains exceptionally strong.

The Dangote Petroleum Refinery says it has reached its full designed capacity of 650,000 barrels of crude oil per day (bpd), a benchmark the company describes as historic and unprecedented for a single-train refinery of that scale.

In a statement issued Wednesday, the refinery said the milestone followed optimization of its Crude Distillation Unit (CDU) and Motor Spirit (MS) production block, stabilizing steady-state operations at what is Africa’s largest oil refining complex.

The company has begun an intensive 72-hour performance test programme in collaboration with its technology licensor, UOP, to validate efficiency and confirm that operational parameters meet global benchmarks.

Chief Executive Officer David Bird said the integration and stabilization of the CDU and MS Block demonstrate the refinery’s engineering strength and operational resilience.

“Our teams have demonstrated exceptional precision and expertise in stabilizing both the CDU and MS Block, and we are pleased to see them functioning at optimal efficiency. This performance testing phase enables us to validate the entire plant under real operating conditions. We are confident that the refinery remains firmly on track to deliver consistent, world-class output,” Bird said.

He added that the CDU and MS Block — comprising the naphtha hydrotreater, isomerization unit, and reformer — are operating steadily at the full nameplate capacity of 650,000 bpd. Phase 2 testing of the remaining processing units is scheduled to commence next week.

Performance test runs of this nature are standard in large-scale refining projects. They involve stress-testing throughput rates, monitoring product yields, assessing energy efficiency ratios, and verifying emissions and safety parameters under continuous load.

Achieving sustained operation at nameplate capacity is distinct from intermittent peak runs. It signals that feedstock flows, heat integration, pressure systems, and product recovery units are synchronized under full operational strain.

The refinery said it supplied between 45 million and 50 million liters of Premium Motor Spirit (PMS) daily during the recent festive season. With the CDU and MS Block fully optimized, it now has the capacity to deliver up to 75 million liters of PMS daily to the domestic market if required.

Nigeria has historically imported more than 80 percent of its refined petroleum products due to limited and underperforming domestic refining capacity. The new throughput levels significantly alter that supply equation.

At full operation, the refinery’s output extends beyond PMS to include diesel, aviation fuel, LPG, and petrochemical feedstocks. The facility’s configuration is designed to maximize high-value light products while minimizing low-margin residual fuel oil.

Macroeconomic Significance

Analysts estimate that full and sustained operation of the 650,000 bpd facility could save Nigeria up to $10 billion annually in foreign exchange previously spent on fuel imports. Reduced import dependency may ease pressure on the naira, improve balance-of-payments stability, and moderate inflationary pressures linked to fuel pricing.

Fuel importation has been a major structural drain on Nigeria’s foreign reserves. By refining crude domestically, the country captures greater value across the hydrocarbon chain, from crude extraction to finished product distribution.

The refinery’s operations are also expected to generate thousands of direct and indirect jobs across logistics, engineering, distribution, and petrochemical manufacturing.

From a regional standpoint, surplus output could position Nigeria as a net exporter of refined petroleum products within West and Central Africa, strengthening energy security across the sub-region.

Industrial and Downstream Impact

Beyond fuels, the refinery complex is designed to support petrochemical integration. Expansion plans include production of linear alkylbenzene, base oils, and increased polypropylene capacity.

In October 2025, industrialist Aliko Dangote announced plans to scale the facility from 650,000 bpd to 1.4 million bpd. If completed, that expansion would surpass the 1.36 million bpd capacity of the Jamnagar Refinery in India, currently regarded as the largest refinery complex globally.

The proposed expansion would also raise annual polypropylene output from one million metric tons to 1.5 million metric tons, supporting plastics manufacturing and industrial production.

Such vertical integration aligns with Nigeria’s broader industrialization strategy, which seeks to reduce reliance on imported refined products and manufactured petrochemical derivatives.

The refinery sits on a 6,180-acre (2,500-hectare) site within the Lekki Free Trade Zone in Lagos. It is supplied by an approximately 1,100-kilometre subsea pipeline network designed to ensure steady crude feedstock delivery.

The facility incorporates advanced residue upgrading units, Sulphur recovery systems, and emissions control technologies aimed at meeting Euro V fuel standards.

Its single-train configuration — as opposed to multiple parallel refining trains — makes the achievement of full nameplate capacity technically significant. Scaling a single integrated unit to 650,000 bpd requires highly coordinated engineering design and process stability.

If operations remain stable and expansion plans proceed as outlined, the Dangote Petroleum Refinery could fundamentally reshape Nigeria’s downstream sector, shift regional trade flows, and anchor broader industrial growth tied to hydrocarbons.

Reaching full nameplate throughput positions the refinery at the center of Nigeria’s most consequential energy transition in decades. The next phase will test whether full-capacity operations can be maintained over extended periods, how efficiently products are distributed nationwide, and how pricing dynamics evolve in a market long shaped by subsidy regimes and import dependency.