Author: Claude, TechFlow

TechFlow Introduction: The Q1 2026 earnings season for tech giants reveals a new phenomenon: while AI helps companies freeze hiring and cut jobs, its own token consumption and GPU depreciation are eating into gross margins in reverse. Shopify's subscription business gross margin is being suppressed by LLM costs, roughly a quarter of Roblox's full-year margin guidance reduction is directly attributed to incremental AI investment, and Amazon, Meta, Microsoft, and Google combined are expected to spend $725 billion on AI capital expenditures in 2026, up 77% year-over-year. The two sides of the AI dividend—labor savings and computing power consumption—are being reconciled on the same earnings report for the first time, with the latter clearly outweighing the former.

The first-quarter earnings season is applying a corrective patch to the simple narrative of "AI replacing labor."

A cohort of tech companies, while delivering results of frozen hiring and accelerated product iteration, are now forced to explain a thornier issue to investors: soaring AI chip depreciation and unpredictable token consumption are eating away at the money saved from layoffs.

Shopify President Harley Finkelstein stated during the May 5, 2026 earnings call that AI now handles over 50% of the company's code writing and helps Shopify deliver more than 300 products and features while keeping headcount stable. However, during the same call, management acknowledged that the gross margin of subscription solutions is being partially offset by large language model (LLM) costs, a trend expected to continue.

Shopify: The LLM Cost Black Hole Behind 80% Gross Margin

Shopify's Q1 subscription solutions gross margin stood at 80%, flat year-over-year, but the cost of maintaining that figure is changing.

According to Shopify's 10-Q filing with the SEC, subscription solutions costs in Q1 2026 rose 20% year-over-year to $148 million, up from $123 million in the same period last year. Cloud and infrastructure costs (including AI-related usage) alone increased by $22 million, the main driver of cost expansion. Shopify CFO Jeff Hoffmeister said on the earnings call that "scale benefits and support efficiency improvements were partially offset by higher LLM costs, primarily driven by merchant usage of Sidekick, and we expect this dynamic to continue."

Sidekick is Shopify's embedded AI assistant, with weekly active stores up 385% year-over-year this quarter. During the quarter, merchants created over 12,000 custom apps using Sidekick, up more than 200% quarter-over-quarter, and nearly half of Shopify Flows were AI-generated. AI-driven store traffic grew 8x year-over-year, and orders from AI search increased nearly 13x year-over-year.

But this surge in usage means exponential growth in AI inference calls. Every interaction between a merchant and Sidekick, every proactive suggestion generated by the Pulse feature, corresponds to a token bill paid to upstream model providers.

Shopify breaks down its "internal AI" and "external AI" books separately for investors: internal AI for writing code and reducing personnel costs is a "cost game" victory, while external AI products offered to merchants represent a strategic choice to "tie infrastructure costs to merchant usage depth." Finkelstein summed up this logic on the earnings call as "AI is a structural advantage, not just a cost."

Roblox: A Quarter of the Margin Reduction Directly from AI

Roblox CFO Naveen Chopra explicitly disclosed on the Q1 2026 earnings call on April 30 that about a quarter of the full-year margin reduction relative to prior guidance came from incremental AI investment and adjustments to DevEx (developer exchange) for U.S. users aged 18 and older.

Roblox currently runs over 400 AI models on its own and cloud GPUs, processing 1.5 million inference calls per second, covering scenarios like discovery recommendations, communication security, marketplace recommendations, and 3D generation.

The company is trying to cut inference costs through business model adjustments. Roblox co-founder and CEO David Baszucki said on the call that the upcoming "Roblox Reality" project—a technology capable of running 2K real-time photorealistic video models at 60Hz—will not be offered for free. "This will use cloud computing resources. There will be some form of subscription or payment mechanism, so we believe we can offset the cost of real-time inference on that side," Baszucki explained.

Chopra added that the company's 2026 capital expenditure guidance remains unchanged, mainly relying on deploying GPUs in its own data centers to meet inference demand this year, while some training tasks will still use the cloud. Roblox previously disclosed that by migrating some AI inference workloads from third-party clouds to its own data centers by the end of 2025, it achieved a 10x efficiency improvement in specific workloads like safety moderation and content discovery.

But Roblox's full-year guidance for this quarter includes multiple pressures such as the aforementioned incremental AI investment, fixed-cost deleveraging from lower-than-expected booking scale, and increasing the DevEx rate for 18+ adult content creators to 37.8%, ultimately triggering a market repricing of its full-year margins.

Industry Ledger: $725 Billion in CapEx vs. $2.7 Billion in Salary Savings

The micro cases of Shopify and Roblox sit within a larger macro structural imbalance.

According to data cited by 24/7 Wall St., Amazon, Meta, Microsoft, and Google combined will spend $725 billion on AI capital expenditures in 2026, up 77% year-over-year. Meta's full-year capEx guidance ranges between $125 billion and $145 billion, meaning it spends up to $370 million per day on data center construction; Microsoft's calendar 2026 CapEx is $190 billion, with Amazon committing $200 billion.

The ratio between this expense and labor costs is quite stark. Meta's total employee compensation—all salaries, benefits, and stock incentives—amounts to about $27 billion. Even if Meta fired all employees tomorrow, the cost savings would be less than a fifth of its 2026 infrastructure spending.

Wedbush Securities analyst Dan Ives estimated in an April 25 research note that Meta's upcoming layoffs of 8,000 people could free up about $2.4 billion in annual operating expenses, offsetting only about 12% of the incremental depreciation drag in 2026. In other words, the financial pressure of every dollar spent on AI computing power would require nearly ten dollars in labor cost savings to fully hedge.

Meta CFO Susan Li said on the Q4 2025 earnings call that Meta's headcount reduction is aimed at "creating a leaner operating model to help offset the large investments we are making." This statement clearly frames layoffs as a financial tool for AI CapEx, not a byproduct of productivity gains.

The Triumph of Model Providers, the Dilemma of Application Layers

The biggest beneficiaries of this ledger game are the underlying model and computing power providers. Microsoft Cloud's gross margin remains at 69% under the pressure of AI infrastructure expansion, OpenAI's gross margin is estimated externally at about 50%, and Anthropic's at around 60%. Nvidia continues to record roughly 70% gross margin in fiscal 2026.

Application-layer companies, especially SaaS players that both consume AI and package it into subscription products for sale, face a new financial structure: revenue is highly correlated with AI usage intensity, but the cost curve is dictated by upstream model providers' pricing, and every model upgrade can bring new token consumption.

In his analysis of AI gross margins, Tanay Jaipuria notes that although inference costs for a single model are declining at 80%-90% annually, prices for frontier models remain stable or even rise. If application-layer companies insist on calling the strongest model for every request, their COGS is effectively driven by the model providers' price list.

Shopify's response is to position AI products as strategic entry points deeply tied to merchant traffic and loyalty, turning the growth in inference costs into a proxy metric for "platform embedding depth." Roblox's approach is to strip high-end AI experiences from the free tier, forcing users to pay for inference costs. Behind both paths lies a common consensus: relying purely on layoffs to cover AI computing bills is mathematically unsustainable.