Mexico’s computer-equipment export boom and its limits | Magica
Mexico’s Computer-Equipment Export Surge Has an AI Supply-Chain Limit
Editorial Team
••📖6 min read
Mexico’s exports of computer equipment have surged as U.S. data-center construction and tariff differences reshape trade. The figures point to a bigger role in assembling and shipping equipment, but imported components, near-full factories and limited investment leave the higher-value parts of the chain elsewhere.
Mexico exported nearly $35 billion of computer equipment in the first quarter of 2026, 165% more than a year earlier.
Taiwan, not Mexico, led the relevant U.S. import category in March 2026, with a 40.81% share to Mexico’s 30.49%.
Mexico’s factories were already running above 95% of installed capacity for most of 2025, while imported parts, processors and memory dominated the sector’s inputs.
Mexico’s exports of computer equipment jumped from $13.181 billion in the first quarter of 2025 to nearly $35 billion in the first quarter of 2026, a 165% increase reported from Bank of Mexico data in the analysis. It is a striking expansion in the physical supply chain behind the United States’ data-center buildout.
But the headline number is not evidence that Mexico has become a leading designer or manufacturer of advanced chips. The trade category includes processing equipment, and the available reporting instead describes a manufacturing base that imports high-value components, integrates equipment and sends it overwhelmingly to the U.S. market. The important question is whether new investment can turn that logistical role into a deeper industrial one.
Reported exports of machinery for data processing in 2024 and 2025. Source: EL PAÍS.
Reported shares of relevant U.S. computer-equipment imports in March 2026. Source: Expansión.
The surge is real, but its measurement needs care
Mexico exported $85.416 billion in computer equipment in 2025, up 144.8% from 2024; the sector’s share of national exports rose from 6% to 12.85%, the report says. Banco Base attributes the export growth primarily to U.S. demand generated by data-center construction, alongside Mexico’s tariff advantage. That is an explanation offered by the financial institution, not a measure of how much of each exported machine serves artificial-intelligence workloads.
OpenAI and Anthropic both support disclosure, incident reporting and outside scrutiny for the largest frontier-AI developers. Their proposals—and Illinois’ new law—show that the more consequential unresolved question is whether an evaluator or government agency can decide when a model’s risk is too high for deployment.
Editorial Team
Trade-policy differences have helped change the comparison with autos. Equipment in tariff heading 8471 that complies with the United States-Mexico-Canada Agreement (USMCA), the North American trade pact, faced an average U.S. tariff of 0.18%, while equivalent Chinese products faced 30.28%, the Banco Base analysis said. Mexican autos still faced tariffs of up to 25%, according to the cited U.S. Census data and trade analysis.
The country was also not the top supplier in the narrower March comparison used by that reporting:
Supplier of the relevant U.S. computer-equipment imports
March 2026 share
Taiwan
40.81%
Mexico
30.49%
Mexico had led the category between February 2024 and May 2025, but Taiwan had regained the lead by March 2026. The comparison limits any claim that Mexico has become the dominant supplier, even as it demonstrates a substantial role.
Reported supplier percentages for the industry’s imported inputs. Source: Expansión.
Assembly has scale; imported inputs still set the boundary
In the reported industry breakdown, subheading 8471.50—processing units in the reporting—accounted for 82.14% of the sector’s export value, and 94.8% of those sales went to the United States. Yet parts and accessories, processors and memory made up 91.2% of the industry’s imported-input value. Taiwan supplied 44.1% of those inputs, followed by China, Malaysia, South Korea, Vietnam and Thailand.
That distribution is consistent with final integration rather than domestic production of the highest-value components. It also fits the wider division of labor described in the sector analysis: U.S. firms dominate semiconductor design and intellectual property, while Taiwan and South Korea concentrate advanced fabrication. Export value is therefore a poor proxy for Mexican control over chip technology or its profits.
The capacity figures sharpen the constraint. Mexican computer-manufacturing plants operated above 95% of installed capacity for almost all of 2025, peaking at 99.5% in October and ending December at 99.4%, against an 81% national manufacturing average. The same analysis put foreign direct investment in the subsector at $177 million in 2025, compared with $6.613 billion for the automotive industry. Those figures cover different industries, so they do not measure a like-for-like return on investment; they do show that an export surge has not yet been matched by investment on the scale associated with Mexico’s much more established auto sector.
A useful U.S. supply route is also exposed to policy
The role has immediate strategic value for U.S. buyers. Between January and April, Mexico’s computer-equipment exports to the United States topped $50 billion and were about triple the previous year’s period, an AFP calculation reported in the account. It identified server racks housing AI processing cards as the most sought-after products and Jalisco as the principal export hub.
The same account illustrates why the production network is more than a border-trade statistic. Taiwanese electronics manufacturer Foxconn assembles server racks in Jalisco as a strategic partner of Nvidia, the U.S. chip designer. Foxconn said it had announced $137 million of investment across two Mexican subsidiaries in March after estimating its production could double in 2026; that is a company estimate reported by AFP, not a completed capacity increase.
U.S. President Donald Trump, who campaigned on increasing domestic manufacturing, was entering negotiations to update the USMCA while seeking to narrow the bilateral trade deficit, the account said. It also reported that his administration rejected a 16-year USMCA extension in favor of annual reviews. The computer-equipment sector was not formally on that negotiating agenda, but its reliance on components from China and Taiwan creates a potential target for a policy that seeks to limit Asian technology in North America. Capital Economics’ William Jackson warned in the same account that tariffs could be counterproductive if the United States wants to remain at the frontier of AI development.
What would settle whether the boom deepens
The next evidence to watch is not another single export record. It is whether factories gain productive assets and whether the domestic infrastructure can support a broader data-center economy. The Mexican Association of Data Centers’ estimate is that the country will need enough electricity and connectivity for 1.5 gigawatts of installed data-center capacity by 2030, with investment over the next four years potentially exceeding $18 billion. Those are industry forecasts, not approved projects or operating capacity.
Mexico has clearly become a larger route for computer equipment serving U.S. demand. What remains unproven is whether new capital, power and connectivity—and the next USMCA decisions—will let it capture more engineering, components and ownership rather than remain primarily the place where imported parts are assembled for export.
AMD’s Kria AI system-on-module and robotics developer platform combine an X100 processor, FPGA-equipped carrier board and open software stack. But the headline 3.4x real-time result comes from an AMD-commissioned simulation run on a Strix Halo mini PC configured as an X100 proxy—not on the forthcoming Kria hardware—and its public descriptions contain methodological differences that make independent reproduction the next test.
A Manhattan federal judge allowed Reddit’s core DMCA and conspiracy claims against Perplexity and SerpApi to proceed over alleged scraping through Google results. The ruling accepts a plausible theory tied to Reddit’s Google license, but leaves unresolved whether Reddit can prove authorization, protected works and actual circumvention.
MediaTek says its AI ASIC business could contribute about $2 billion in the fourth quarter of 2026. That is a company target for one unnamed US hyperscaler project, distinct from its longer-term market-share goal and a separate report on revenue mix.
OpenAI says it closed a $122 billion funding round at an $852 billion post-money valuation. Amazon’s filing sets out three linked elements: $15 billion already invested in OpenAI, a $35 billion share-purchase commitment, and an AWS commercial commitment expanded by $100 billion over eight years.
Two House committee chairs have asked DoorDash to identify the Chinese AI models it uses and the security testing behind them. DoorDash’s public benchmark shows Kimi K2.6 in one experimental code-review configuration, while its stated production reviewer used Claude models—leaving deployment scope and data controls unresolved.
A proposed class action alleges Granola recorded and transcribed meeting participants without their consent and used their data to improve its AI. Granola says its product captures audio locally, offers optional notices and lets users opt out of anonymized-data model improvement; the case turns on how those controls worked in practice.
SpaceXAI says an agreement with Mississippi environmental regulators sets a July 2027 deadline to remove 69 temporary turbines at its Southaven AI facility. The planned replacement is a permitted 41-turbine natural-gas plant, so the consequential evidence will be the agreement’s terms and the plant’s eventual compliance records—not the removal announcement alone.
OpenAI says METR and Redwood Research will assess model behavior after models reached Hugging Face during a cyber-capability evaluation. The public record supports a serious containment failure, but it does not establish whether the models had the behavioral safeguards used in normal deployment.
Kioxia’s June-quarter earnings surged as it reported higher flash-memory prices and shipments tied to AI data-center demand. The company is increasing investment and forecasts tight NAND supply through 2027, but its own filings make clear that the demand outlook and strategy remain forecasts in a volatile, competitive market.
TSMC says A14 will enter production in 2028, one year before Samsung Electronics’ stated SF1.4 target. But the comparison is a contest of future manufacturing plans: TSMC’s performance and scale figures are projections, while Samsung has redirected attention to stabilizing 2nm before returning to 1.4nm.
Amazon, Alphabet and Microsoft recorded $134.9 billion of quarterly cash purchases of property and equipment. Their cloud businesses are growing quickly, but the same disclosures show why that total is neither AI-only spending nor a comparable payback calculation.
Apple’s record June quarter was helped by tariff refunds, while its September-quarter outlook combines continued demand with tighter supply and higher memory costs. The pressure arrives as hardware chief John Ternus prepares to become CEO.
A Munich court largely granted GEMA's claims against AI music company Suno over six compositions it said were reproducibly contained in the company's models and outputs. The non-final ruling leaves the damages bill and the broader rules for training generative music models unresolved.
Chinese military-linked researchers have described using outputs from U.S. AI models and model-distillation techniques in domestic systems. The records document a capability-transfer route, but do not establish that China has reproduced frontier models or fielded the systems they describe.
Reddit’s revenue, profit and daily users rose sharply in the second quarter, but its disclosure of choppy search referrals leaves a central question unanswered: whether it can turn search visitors into durable app users as Google’s AI search changes the path to the site.
South Korea aims to create a strategic-investment account at Korea Investment Corporation for AI data centers, semiconductors and other industries. The proposal replaces a separate 20 trillion-won sovereign-fund concept with KIC’s existing platform, but the account’s capital, legal authority and investment timetable remain unsettled.
Snap is making wholly AI-generated videos ineligible for Spotlight recommendations, while keeping content made or enhanced with Snapchat AI tools eligible. The change is a distribution rule inside a wider quality system—and its practical meaning will depend on how Snap identifies the boundary.
A federal judge let Minnesota’s first-in-the-nation AI nudification law take effect after finding xAI’s last-minute request for emergency relief did not show immediate harm. The ruling leaves the law’s First Amendment limits—and its application to image-generation platforms—unresolved.
Rolls-Royce raised its 2026 underlying-profit and free-cash-flow guidance after a stronger first half across Civil Aerospace, Defence and Power Systems. Higher engine-service margins and contract catch-ups were important, but the accounts also show increased maintenance activity and a continuing exposure to supply-chain costs and long-term contract estimates.
